2012-08-13 16:52:54 +02:00
|
|
|
|
/* aarch64-tbl.h -- AArch64 opcode description table and instruction
|
|
|
|
|
operand description table.
|
2019-01-01 11:31:27 +01:00
|
|
|
|
Copyright (C) 2012-2019 Free Software Foundation, Inc.
|
2012-08-13 16:52:54 +02:00
|
|
|
|
|
|
|
|
|
This file is part of the GNU opcodes library.
|
|
|
|
|
|
|
|
|
|
This library is free software; you can redistribute it and/or modify
|
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
|
the Free Software Foundation; either version 3, or (at your option)
|
|
|
|
|
any later version.
|
|
|
|
|
|
|
|
|
|
It is distributed in the hope that it will be useful, but WITHOUT
|
|
|
|
|
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
|
|
|
|
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public
|
|
|
|
|
License for more details.
|
|
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
|
along with this file; see the file COPYING. If not, write to the
|
|
|
|
|
Free Software Foundation, 51 Franklin Street - Fifth Floor, Boston,
|
|
|
|
|
MA 02110-1301, USA. */
|
|
|
|
|
|
|
|
|
|
#include "aarch64-opc.h"
|
|
|
|
|
|
2016-05-03 12:48:56 +02:00
|
|
|
|
#ifndef VERIFIER
|
|
|
|
|
#error VERIFIER must be defined.
|
|
|
|
|
#endif
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Operand type. */
|
|
|
|
|
|
|
|
|
|
#define OPND(x) AARCH64_OPND_##x
|
|
|
|
|
#define OP0() {}
|
|
|
|
|
#define OP1(a) {OPND(a)}
|
|
|
|
|
#define OP2(a,b) {OPND(a), OPND(b)}
|
|
|
|
|
#define OP3(a,b,c) {OPND(a), OPND(b), OPND(c)}
|
|
|
|
|
#define OP4(a,b,c,d) {OPND(a), OPND(b), OPND(c), OPND(d)}
|
|
|
|
|
#define OP5(a,b,c,d,e) {OPND(a), OPND(b), OPND(c), OPND(d), OPND(e)}
|
|
|
|
|
|
|
|
|
|
#define QLF(x) AARCH64_OPND_QLF_##x
|
|
|
|
|
#define QLF1(a) {QLF(a)}
|
|
|
|
|
#define QLF2(a,b) {QLF(a), QLF(b)}
|
|
|
|
|
#define QLF3(a,b,c) {QLF(a), QLF(b), QLF(c)}
|
|
|
|
|
#define QLF4(a,b,c,d) {QLF(a), QLF(b), QLF(c), QLF(d)}
|
|
|
|
|
#define QLF5(a,b,c,d,e) {QLF(a), QLF(b), QLF(c), QLF(d), QLF(e)}
|
|
|
|
|
|
|
|
|
|
/* Qualifiers list. */
|
|
|
|
|
|
|
|
|
|
/* e.g. MSR <systemreg>, <Xt>. */
|
|
|
|
|
#define QL_SRC_X \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(NIL,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MRS <Xt>, <systemreg>. */
|
|
|
|
|
#define QL_DST_X \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SYS #<op1>, <Cn>, <Cm>, #<op2>{, <Xt>}. */
|
|
|
|
|
#define QL_SYS \
|
|
|
|
|
{ \
|
2016-12-13 13:37:18 +01:00
|
|
|
|
QLF5(NIL,CR,CR,NIL,X), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SYSL <Xt>, #<op1>, <Cn>, <Cm>, #<op2>. */
|
|
|
|
|
#define QL_SYSL \
|
|
|
|
|
{ \
|
2016-12-13 13:37:18 +01:00
|
|
|
|
QLF5(X,NIL,CR,CR,NIL), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ADRP <Xd>, <label>. */
|
|
|
|
|
#define QL_ADRP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. B.<cond> <label>. */
|
|
|
|
|
#define QL_PCREL_NIL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. TBZ <Xt>, #<imm>, <label>. */
|
|
|
|
|
#define QL_PCREL_14 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,imm_0_63,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. BL <label>. */
|
|
|
|
|
#define QL_PCREL_26 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDRSW <Xt>, <label>. */
|
|
|
|
|
#define QL_X_PCREL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDR <Wt>, <label>. */
|
|
|
|
|
#define QL_R_PCREL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,NIL), \
|
|
|
|
|
QLF2(X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDR <Dt>, <label>. */
|
|
|
|
|
#define QL_FP_PCREL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S,NIL), \
|
|
|
|
|
QLF2(S_D,NIL), \
|
|
|
|
|
QLF2(S_Q,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PRFM <prfop>, <label>. */
|
|
|
|
|
#define QL_PRFM_PCREL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. BR <Xn>. */
|
|
|
|
|
#define QL_I1X \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(X), \
|
|
|
|
|
}
|
|
|
|
|
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
/* e.g. STG [<Xn|SP>, #<imm9>]. */
|
|
|
|
|
#define QL_LDST_AT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(imm_tag), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. RBIT <Wd>, <Wn>. */
|
|
|
|
|
#define QL_I2SAME \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
QLF2(X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CMN <Wn|WSP>, <Wm>{, <extend> {#<amount>}}. */
|
|
|
|
|
#define QL_I2_EXT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
QLF2(X,W), \
|
|
|
|
|
QLF2(X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MOV <Wd|WSP>, <Wn|WSP>, at least one SP. */
|
|
|
|
|
#define QL_I2SP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(WSP,W), \
|
|
|
|
|
QLF2(W,WSP), \
|
|
|
|
|
QLF2(SP,X), \
|
|
|
|
|
QLF2(X,SP), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. REV <Wd>, <Wn>. */
|
|
|
|
|
#define QL_I2SAMEW \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. REV32 <Xd>, <Xn>. */
|
|
|
|
|
#define QL_I2SAMEX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define QL_I2SAMER \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
QLF2(X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
2013-02-28 20:18:40 +01:00
|
|
|
|
/* e.g. CRC32B <Wd>, <Wn>, <Wm>. */
|
|
|
|
|
#define QL_I3SAMEW \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,W,W), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SMULH <Xd>, <Xn>, <Xm>. */
|
|
|
|
|
#define QL_I3SAMEX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
2013-02-28 20:18:40 +01:00
|
|
|
|
/* e.g. CRC32X <Wd>, <Wn>, <Xm>. */
|
|
|
|
|
#define QL_I3WWX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,W,X), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. UDIV <Xd>, <Xn>, <Xm>. */
|
|
|
|
|
#define QL_I3SAMER \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,W,W), \
|
|
|
|
|
QLF3(X,X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ADDS <Xd>, <Xn|SP>, <R><m>{, <extend> {#<amount>}}. */
|
|
|
|
|
#define QL_I3_EXT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,W,W), \
|
|
|
|
|
QLF3(X,X,W), \
|
|
|
|
|
QLF3(X,X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MADD <Xd>, <Xn>, <Xm>, <Xa>. */
|
|
|
|
|
#define QL_I4SAMER \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W,W,W,W), \
|
|
|
|
|
QLF4(X,X,X,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SMADDL <Xd>, <Wn>, <Wm>, <Xa>. */
|
|
|
|
|
#define QL_I3SAMEL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,W,W), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SMADDL <Xd>, <Wn>, <Wm>, <Xa>. */
|
|
|
|
|
#define QL_I4SAMEL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(X,W,W,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CSINC <Xd>, <Xn>, <Xm>, <cond>. */
|
|
|
|
|
#define QL_CSEL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W, W, W, NIL), \
|
|
|
|
|
QLF4(X, X, X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CSET <Wd>, <cond>. */
|
|
|
|
|
#define QL_DST_R \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, NIL), \
|
|
|
|
|
QLF2(X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. BFM <Wd>, <Wn>, #<immr>, #<imms>. */
|
|
|
|
|
#define QL_BF \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W,W,imm_0_31,imm_0_31), \
|
|
|
|
|
QLF4(X,X,imm_0_63,imm_0_63), \
|
|
|
|
|
}
|
|
|
|
|
|
[BINUTILS, AARCH64, 2/8] Add Tag generation instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag generation instructions from
MTE. These are the following instructions added in this patch:
- IRG <Xd|SP>, <Xn|SP>{, Xm}
- ADDG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- SUBG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- GMI <Xd>, <Xn|SP>, <Xm>
where
<Xd|SP> : Is the 64-bit destination GPR or Stack pointer.
<Xn|SP> : Is the 64-bit source GPR or Stack pointer.
<uimm6> : Is the unsigned immediate, a multiple of 16
in the range 0 to 1008.
<uimm4> : Is the unsigned immediate, in the range 0 to 15.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10 as new enums.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (aarch64_field_kind): New FLD_imm4_3.
(OPD_F_SHIFT_BY_4, operand_need_shift_by_four): New.
* aarch64-opc.c (fields): Add entry for imm4_3.
(operand_general_constraint_met_p): Add cases for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_ADDG): New.
(aarch64_opcode_table): Add addg, subg, irg and gmi.
(AARCH64_OPERANDS): Define UIMM4_ADDG and UIMM10.
* aarch64-asm.c (aarch64_ins_imm): Add case for
operand_need_shift_by_four.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: New.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
* testsuite/gas/aarch64/illegal-memtag.d: Likewise.
2018-11-12 13:52:55 +01:00
|
|
|
|
/* e.g. ADDG <Xd>, <Xn>, #<uimm10>, #<uimm4>. */
|
|
|
|
|
#define QL_ADDG \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(X,X,NIL,imm_0_15), \
|
|
|
|
|
} \
|
|
|
|
|
|
2015-11-27 16:25:08 +01:00
|
|
|
|
/* e.g. BFC <Wd>, #<immr>, #<imms>. */
|
|
|
|
|
#define QL_BF1 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (W, imm_0_31, imm_1_32), \
|
|
|
|
|
QLF3 (X, imm_0_63, imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. UBFIZ <Wd>, <Wn>, #<lsb>, #<width>. */
|
|
|
|
|
#define QL_BF2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W,W,imm_0_31,imm_1_32), \
|
|
|
|
|
QLF4(X,X,imm_0_63,imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SCVTF <Sd>, <Xn>, #<fbits>. */
|
|
|
|
|
#define QL_FIX2FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,W,imm_1_32), \
|
|
|
|
|
QLF3(S_S,W,imm_1_32), \
|
|
|
|
|
QLF3(S_D,X,imm_1_64), \
|
|
|
|
|
QLF3(S_S,X,imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. SCVTF <Hd>, <Xn>, #<fbits>. */
|
|
|
|
|
#define QL_FIX2FP_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (S_H, W, imm_1_32), \
|
|
|
|
|
QLF3 (S_H, X, imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. FCVTZS <Wd>, <Dn>, #<fbits>. */
|
|
|
|
|
#define QL_FP2FIX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,S_D,imm_1_32), \
|
|
|
|
|
QLF3(W,S_S,imm_1_32), \
|
|
|
|
|
QLF3(X,S_D,imm_1_64), \
|
|
|
|
|
QLF3(X,S_S,imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. FCVTZS <Wd>, <Hn>, #<fbits>. */
|
|
|
|
|
#define QL_FP2FIX_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (W, S_H, imm_1_32), \
|
|
|
|
|
QLF3 (X, S_H, imm_1_64), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SCVTF <Dd>, <Wn>. */
|
|
|
|
|
#define QL_INT2FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D,W), \
|
|
|
|
|
QLF2(S_S,W), \
|
|
|
|
|
QLF2(S_D,X), \
|
|
|
|
|
QLF2(S_S,X), \
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-08 15:02:52 +02:00
|
|
|
|
/* e.g. FMOV <Dd>, <Xn>. */
|
|
|
|
|
#define QL_INT2FP_FMOV \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S,W), \
|
|
|
|
|
QLF2(S_D,X), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. SCVTF <Hd>, <Wn>. */
|
|
|
|
|
#define QL_INT2FP_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, W), \
|
|
|
|
|
QLF2 (S_H, X), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. FCVTNS <Xd>, <Dn>. */
|
|
|
|
|
#define QL_FP2INT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,S_D), \
|
|
|
|
|
QLF2(W,S_S), \
|
|
|
|
|
QLF2(X,S_D), \
|
|
|
|
|
QLF2(X,S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-08 15:02:52 +02:00
|
|
|
|
/* e.g. FMOV <Xd>, <Dn>. */
|
|
|
|
|
#define QL_FP2INT_FMOV \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,S_S), \
|
|
|
|
|
QLF2(X,S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. FCVTNS <Hd>, <Wn>. */
|
|
|
|
|
#define QL_FP2INT_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (W, S_H), \
|
|
|
|
|
QLF2 (X, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
2016-11-18 10:53:45 +01:00
|
|
|
|
/* e.g. FJCVTZS <Wd>, <Dn>. */
|
|
|
|
|
#define QL_FP2INT_W_D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (W, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. FMOV <Xd>, <Vn>.D[1]. */
|
|
|
|
|
#define QL_XVD1 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMOV <Vd>.D[1], <Xn>. */
|
|
|
|
|
#define QL_VD1X \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D,X), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. EXTR <Xd>, <Xn>, <Xm>, #<lsb>. */
|
|
|
|
|
#define QL_EXTR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W,W,W,imm_0_31), \
|
|
|
|
|
QLF4(X,X,X,imm_0_63), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LSL <Wd>, <Wn>, #<uimm>. */
|
|
|
|
|
#define QL_SHIFT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,W,imm_0_31), \
|
|
|
|
|
QLF3(X,X,imm_0_63), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. UXTH <Xd>, <Wn>. */
|
|
|
|
|
#define QL_EXT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
QLF2(X,W), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. UXTW <Xd>, <Wn>. */
|
|
|
|
|
#define QL_EXT_W \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,W), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SQSHL <V><d>, <V><n>, #<shift>. */
|
|
|
|
|
#define QL_SSHIFT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B , S_B , S_B ), \
|
|
|
|
|
QLF3(S_H , S_H , S_H ), \
|
|
|
|
|
QLF3(S_S , S_S , S_S ), \
|
|
|
|
|
QLF3(S_D , S_D , S_D ) \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SSHR <V><d>, <V><n>, #<shift>. */
|
|
|
|
|
#define QL_SSHIFT_D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D , S_D , S_D ) \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. UCVTF <Vd>.<T>, <Vn>.<T>, #<fbits>. */
|
|
|
|
|
#define QL_SSHIFT_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S , S_S , S_S ), \
|
|
|
|
|
QLF3(S_D , S_D , S_D ) \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:46:21 +01:00
|
|
|
|
/* e.g. UCVTF <Vd>.<T>, <Vn>.<T>, #<fbits>. */
|
|
|
|
|
#define QL_SSHIFT_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (S_H, S_H, S_H) \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SQSHRUN <Vb><d>, <Va><n>, #<shift>. */
|
|
|
|
|
#define QL_SSHIFTN \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B , S_H , S_B ), \
|
|
|
|
|
QLF3(S_H , S_S , S_H ), \
|
|
|
|
|
QLF3(S_S , S_D , S_S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SSHR <Vd>.<T>, <Vn>.<T>, #<shift>.
|
|
|
|
|
The register operand variant qualifiers are deliberately used for the
|
|
|
|
|
immediate operand to ease the operand encoding/decoding and qualifier
|
|
|
|
|
sequence matching. */
|
|
|
|
|
#define QL_VSHIFT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_16B, V_16B, V_16B), \
|
|
|
|
|
QLF3(V_4H , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_8H , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_2D ) \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SCVTF <Vd>.<T>, <Vn>.<T>, #<fbits>. */
|
|
|
|
|
#define QL_VSHIFT_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_2D ) \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:40:03 +01:00
|
|
|
|
/* e.g. SCVTF <Vd>.<T>, <Vn>.<T>, #<fbits>. */
|
|
|
|
|
#define QL_VSHIFT_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (V_4H, V_4H, V_4H), \
|
|
|
|
|
QLF3 (V_8H, V_8H, V_8H) \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SHRN<Q> <Vd>.<Tb>, <Vn>.<Ta>, #<shift>. */
|
|
|
|
|
#define QL_VSHIFTN \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8H , V_8B ), \
|
|
|
|
|
QLF3(V_4H , V_4S , V_4H ), \
|
|
|
|
|
QLF3(V_2S , V_2D , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHRN<Q> <Vd>.<Tb>, <Vn>.<Ta>, #<shift>. */
|
|
|
|
|
#define QL_VSHIFTN2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_16B, V_8H, V_16B), \
|
|
|
|
|
QLF3(V_8H , V_4S , V_8H ), \
|
|
|
|
|
QLF3(V_4S , V_2D , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SSHLL<Q> <Vd>.<Ta>, <Vn>.<Tb>, #<shift>.
|
|
|
|
|
the 3rd qualifier is used to help the encoding. */
|
|
|
|
|
#define QL_VSHIFTL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_4S , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_2D , V_2S , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SSHLL<Q> <Vd>.<Ta>, <Vn>.<Tb>, #<shift>. */
|
|
|
|
|
#define QL_VSHIFTL2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_16B, V_16B), \
|
|
|
|
|
QLF3(V_4S , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2D , V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. TBL. */
|
|
|
|
|
#define QL_TABLE \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_16B, V_8B ), \
|
|
|
|
|
QLF3(V_16B, V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA1H. */
|
|
|
|
|
#define QL_2SAMES \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ABS <V><d>, <V><n>. */
|
|
|
|
|
#define QL_2SAMED \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CMGT <V><d>, <V><n>, #0. */
|
|
|
|
|
#define QL_SISD_CMP_0 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D, S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCMEQ <V><d>, <V><n>, #0. */
|
|
|
|
|
#define QL_SISD_FCMP_0 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S, S_S, NIL), \
|
|
|
|
|
QLF3(S_D, S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 5/14] Support FP16 Scalar Two Register Misc. instructions.
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the floating point and Adv.SIMD support. This patch adds FP16
instructions to the group Scalar Two Register Misc, making them
available when +simd+fp16 is enabled.
The instructions added are: FCMGT, FCMGE, FCMEQ, FCMLE, FCMLT, FCVTNS,
FCVTNU, FCVTPS, FCVTPU, FCVTMS, FCVTMU, FCVTZS, FCVTZU, FCVTAS, FCVTAU,
SCVTF, UCVTF, FRECPE, FRSQRTE and FRECPX.
The general form for these instructions is
<OP> <Hd>, <Hs>
or
<OP> <Hd>, <Hs>, #0.0
Tested the series for aarch64-none-linux-gnu with cross-compiled
check-binutils and check-gas.
gas/testsuite/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/advsimd-fp16.d: Update expected output.
* gas/aarch64/advsimd-fp16.s: Add tests for scalar two register misc.
instructions.
opcodes/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_SISD_FCMP_H_0): new.
(QL_S_2SAMEH): New.
(aarch64_opcode_table): Add fp16 versions of fcvtns, fcvtms,
fcvtas, scvtf, fcmgt, fcmeq, fcmlt, fcvtps, fcvtzs, frecpe,
frecpx, fcvtnu, fcvtmu, fcvtau, ucvtf, fcmge, fcmle, fcvtpu,
fcvtzu and frsqrte to the scalar two register misc. group.
Change-Id: I19b25baae33027ce1bade68cc8dc47a4321d045c
2015-12-14 17:57:04 +01:00
|
|
|
|
/* e.g. FCMEQ <V><d>, <V><n>, #0. */
|
|
|
|
|
#define QL_SISD_FCMP_H_0 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (S_H, S_H, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. FMAXNMP <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_SISD_PAIR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, V_2S), \
|
|
|
|
|
QLF2(S_D, V_2D), \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:35:47 +01:00
|
|
|
|
/* e.g. FMAXNMP <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_SISD_PAIR_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, V_2H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. ADDP <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_SISD_PAIR_D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D, V_2D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. DUP <V><d>, <Vn>.<T>[<index>]. */
|
|
|
|
|
#define QL_S_2SAME \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B, S_B), \
|
|
|
|
|
QLF2(S_H, S_H), \
|
|
|
|
|
QLF2(S_S, S_S), \
|
|
|
|
|
QLF2(S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTNS <V><d>, <V><n>. */
|
|
|
|
|
#define QL_S_2SAMESD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, S_S), \
|
|
|
|
|
QLF2(S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 5/14] Support FP16 Scalar Two Register Misc. instructions.
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the floating point and Adv.SIMD support. This patch adds FP16
instructions to the group Scalar Two Register Misc, making them
available when +simd+fp16 is enabled.
The instructions added are: FCMGT, FCMGE, FCMEQ, FCMLE, FCMLT, FCVTNS,
FCVTNU, FCVTPS, FCVTPU, FCVTMS, FCVTMU, FCVTZS, FCVTZU, FCVTAS, FCVTAU,
SCVTF, UCVTF, FRECPE, FRSQRTE and FRECPX.
The general form for these instructions is
<OP> <Hd>, <Hs>
or
<OP> <Hd>, <Hs>, #0.0
Tested the series for aarch64-none-linux-gnu with cross-compiled
check-binutils and check-gas.
gas/testsuite/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/advsimd-fp16.d: Update expected output.
* gas/aarch64/advsimd-fp16.s: Add tests for scalar two register misc.
instructions.
opcodes/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_SISD_FCMP_H_0): new.
(QL_S_2SAMEH): New.
(aarch64_opcode_table): Add fp16 versions of fcvtns, fcvtms,
fcvtas, scvtf, fcmgt, fcmeq, fcmlt, fcvtps, fcvtzs, frecpe,
frecpx, fcvtnu, fcvtmu, fcvtau, ucvtf, fcmge, fcmle, fcvtpu,
fcvtzu and frsqrte to the scalar two register misc. group.
Change-Id: I19b25baae33027ce1bade68cc8dc47a4321d045c
2015-12-14 17:57:04 +01:00
|
|
|
|
/* e.g. FCVTNS <V><d>, <V><n>. */
|
|
|
|
|
#define QL_S_2SAMEH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SQXTN <Vb><d>, <Va><n>. */
|
|
|
|
|
#define QL_SISD_NARROW \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B, S_H), \
|
|
|
|
|
QLF2(S_H, S_S), \
|
|
|
|
|
QLF2(S_S, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTXN <Vb><d>, <Va><n>. */
|
|
|
|
|
#define QL_SISD_NARROW_S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVT. */
|
|
|
|
|
#define QL_FCVT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, S_H), \
|
|
|
|
|
QLF2(S_S, S_D), \
|
|
|
|
|
QLF2(S_D, S_H), \
|
|
|
|
|
QLF2(S_D, S_S), \
|
|
|
|
|
QLF2(S_H, S_S), \
|
|
|
|
|
QLF2(S_H, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* FMOV <Dd>, <Dn>. */
|
|
|
|
|
#define QL_FP2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, S_S), \
|
|
|
|
|
QLF2(S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* FMOV <Hd>, <Hn>. */
|
|
|
|
|
#define QL_FP2_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SQADD <V><d>, <V><n>, <V><m>. */
|
|
|
|
|
#define QL_S_3SAME \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B, S_B, S_B), \
|
|
|
|
|
QLF3(S_H, S_H, S_H), \
|
|
|
|
|
QLF3(S_S, S_S, S_S), \
|
|
|
|
|
QLF3(S_D, S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CMGE <V><d>, <V><n>, <V><m>. */
|
|
|
|
|
#define QL_S_3SAMED \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D, S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SQDMULH <V><d>, <V><n>, <V><m>. */
|
|
|
|
|
#define QL_SISD_HS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H, S_H, S_H), \
|
|
|
|
|
QLF3(S_S, S_S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SQDMLAL <Va><d>, <Vb><n>, <Vb><m>. */
|
|
|
|
|
#define QL_SISDL_HS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S, S_H, S_H), \
|
|
|
|
|
QLF3(S_D, S_S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* FMUL <Sd>, <Sn>, <Sm>. */
|
|
|
|
|
#define QL_FP3 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S, S_S, S_S), \
|
|
|
|
|
QLF3(S_D, S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* FMUL <Hd>, <Hn>, <Hm>. */
|
|
|
|
|
#define QL_FP3_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (S_H, S_H, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* FMADD <Dd>, <Dn>, <Dm>, <Da>. */
|
|
|
|
|
#define QL_FP4 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_S, S_S, S_S, S_S), \
|
|
|
|
|
QLF4(S_D, S_D, S_D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* FMADD <Hd>, <Hn>, <Hm>, <Ha>. */
|
|
|
|
|
#define QL_FP4_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4 (S_H, S_H, S_H, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. FCMP <Dn>, #0.0. */
|
|
|
|
|
#define QL_DST_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, NIL), \
|
|
|
|
|
QLF2(S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. FCMP <Hn>, #0.0. */
|
|
|
|
|
#define QL_DST_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* FCSEL <Sd>, <Sn>, <Sm>, <cond>. */
|
|
|
|
|
#define QL_FP_COND \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_S, S_S, S_S, NIL), \
|
|
|
|
|
QLF4(S_D, S_D, S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* FCSEL <Hd>, <Hn>, <Hm>, <cond>. */
|
|
|
|
|
#define QL_FP_COND_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4 (S_H, S_H, S_H, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. CCMN <Xn>, <Xm>, #<nzcv>, <cond>. */
|
|
|
|
|
#define QL_CCMP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W, W, NIL, NIL), \
|
|
|
|
|
QLF4(X, X, NIL, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. CCMN <Xn>, #<imm>, #<nzcv>, <cond>, */
|
|
|
|
|
#define QL_CCMP_IMM \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W, NIL, NIL, NIL), \
|
|
|
|
|
QLF4(X, NIL, NIL, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCCMP <Sn>, <Sm>, #<nzcv>, <cond>. */
|
|
|
|
|
#define QL_FCCMP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_S, S_S, NIL, NIL), \
|
|
|
|
|
QLF4(S_D, S_D, NIL, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 3/3] Add floating-point FP16 instructions
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the ARMv8 FP support. This patch adds the new FP16 instructions,
making them available when the architecture extension +fp+fp16 is
specified.
The instructions added are:
- Comparisons and conditionals: FCMP, FCCMPE, FCMP, FCMPE and FCSEL.
- Arithmetic: FABS, FNEG, FSQRT, FMUL, FDIV, FADD, FSUB, FMADD, FMSUB,
FNMADD and FNMSUB.
- Rounding: FRINTN, FRINTP, FRINTM, FRINTZ, FRINTA, FRINTX and FRINTI.
- Conversions: SCVTF (fixed-point), SCVTF (integer), UCVTF (fixed-point)
UCVTF (integer), FCVTZS (fixed-point), FCVTZS (integer), FCVTZU
(fixed-point), FCVTZU (integer), FCVTNS, FCVTNU, FCVTAS, FCVTAU,
FCVTPS, FCVTPU, FCVTMS and FCVTMU.
- Scalar FMOV: immediate, general and register
gas/testsuite/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/float-fp16.d: New.
* gas/aarch64/float-fp16.s: New.
opcodes/
2015-11-27 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_FIX2FP_H, QL_FP2FIX_H): New.
(QL_INT2FP_H, QL_FP2INT_H): New.
(QL_FP2_H, QL_FP3_H, QL_FP4_H): New
(QL_DST_H): New.
(QL_FCCMP_H): New.
(aarch64_opcode_table): Add 16-bit variants of scvt, ucvtf,
fcvtzs, fcvtzu, fcvtns, fcvtnu, scvtf, ucvtf, fcvtas, fcvtau,
fmov, fcvtpos, fcvtpu, fcvtms, fcvtmu, fcvtzs, fcvtzu, fccmp,
fccmpe, fcmp, fcmpe, fabs, fneg, fsqrt, frintn, frintp, frintm,
frintz, frinta, frintx, frinti, fmul, fdiv, fadd, fsub, fmax,
fmin, fmaxnm, fminnm, fnmul, fmadd, fmsub, fnmadd, fnmsub and
fcsel.
Change-Id: Ie6d40bd1b215a9bc024e12ba75e52afbe1675eb7
2015-11-27 17:32:21 +01:00
|
|
|
|
/* e.g. FCCMP <Sn>, <Sm>, #<nzcv>, <cond>. */
|
|
|
|
|
#define QL_FCCMP_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4 (S_H, S_H, NIL, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. DUP <Vd>.<T>, <Vn>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_DUP_VX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , S_B ), \
|
|
|
|
|
QLF2(V_16B, S_B ), \
|
|
|
|
|
QLF2(V_4H , S_H ), \
|
|
|
|
|
QLF2(V_8H , S_H ), \
|
|
|
|
|
QLF2(V_2S , S_S ), \
|
|
|
|
|
QLF2(V_4S , S_S ), \
|
|
|
|
|
QLF2(V_2D , S_D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. DUP <Vd>.<T>, <Wn>. */
|
|
|
|
|
#define QL_DUP_VR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , W ), \
|
|
|
|
|
QLF2(V_16B, W ), \
|
|
|
|
|
QLF2(V_4H , W ), \
|
|
|
|
|
QLF2(V_8H , W ), \
|
|
|
|
|
QLF2(V_2S , W ), \
|
|
|
|
|
QLF2(V_4S , W ), \
|
|
|
|
|
QLF2(V_2D , X ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. INS <Vd>.<Ts>[<index>], <Wn>. */
|
|
|
|
|
#define QL_INS_XR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H , W ), \
|
|
|
|
|
QLF2(S_S , W ), \
|
|
|
|
|
QLF2(S_D , X ), \
|
|
|
|
|
QLF2(S_B , W ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SMOV <Wd>, <Vn>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_SMOV \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W , S_H), \
|
|
|
|
|
QLF2(X , S_H), \
|
|
|
|
|
QLF2(X , S_S), \
|
|
|
|
|
QLF2(W , S_B), \
|
|
|
|
|
QLF2(X , S_B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. UMOV <Wd>, <Vn>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_UMOV \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W , S_H), \
|
|
|
|
|
QLF2(W , S_S), \
|
|
|
|
|
QLF2(X , S_D), \
|
|
|
|
|
QLF2(W , S_B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MOV <Wd>, <Vn>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_MOV \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W , S_S), \
|
|
|
|
|
QLF2(X , S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SUQADD <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAME \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8B ), \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
QLF2(V_4H , V_4H ), \
|
|
|
|
|
QLF2(V_8H , V_8H ), \
|
|
|
|
|
QLF2(V_2S , V_2S ), \
|
|
|
|
|
QLF2(V_4S , V_4S ), \
|
|
|
|
|
QLF2(V_2D , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. URSQRTE <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAMES \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S , V_2S ), \
|
|
|
|
|
QLF2(V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. REV32 <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAMEBH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8B ), \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
QLF2(V_4H , V_4H ), \
|
|
|
|
|
QLF2(V_8H , V_8H ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FRINTN <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAMESD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S , V_2S ), \
|
|
|
|
|
QLF2(V_4S , V_4S ), \
|
|
|
|
|
QLF2(V_2D , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. REV64 <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAMEBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8B ), \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
QLF2(V_4H , V_4H ), \
|
|
|
|
|
QLF2(V_8H , V_8H ), \
|
|
|
|
|
QLF2(V_2S , V_2S ), \
|
|
|
|
|
QLF2(V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 4/14] Support FP16 Vector Two Register Misc. instructions.
ARMv8.2 adds 16-bit floating point operations as an optional extension
to the floating point and Adv.SIMD support. This patch adds FP16
instructions to the group Vector Two Register Misc, making them
available when +simd+fp16 is enabled.
The instructions added are: FCMGT, FCMGE, FCMEQ, FCMLE, FCMLT, FABS,
FNEG, FRINTN, FRINTA, FRINTP, FRINTM, FRINTX, FRINTZ, FRINTI, FCVTNS,
FCVTNU, FCVTPS, FCVTPU, FCVTMS, FCVTMU, FCVTZS, FCVTZU, FCVTAS, FCVTAU,
SCVTF, UCVTF, FRECPE, FRSQRTE and FSQRT.
The general form for these instructions is
<OP> <Vd>.<T>, <Vs>.<T>
where T is 4h or 8h.
gas/testsuite/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/advsimd-fp16.d: Update expected output.
* gas/aarch64/advsimd-fp16.s: Add tests for vector two register misc.
instructions.
opcodes/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_V2SAMEH): New.
(aarch64_opcode_table): Add fp16 versions of frintn, frintm,
fcvtns, fcvtms, fcvtas, scvtf, fcmgt, fcmeq, fcmlt, fabs, frintp,
frintz, fcvtps, fcvtzs, frecpe, frinta, frintx, fcvtnu, fcvtmu,
fcvtau, ucvtf, fcmge, fcmle, fneg, frinti, fcvtpu, fcvtzu, frsqrte
and fsqrt to the vector register misc. group.
Change-Id: I0267511a7f7ea14247504d29fe4752e84c9af9ad
2015-12-14 17:54:38 +01:00
|
|
|
|
/* e.g. FCMGT <Vd>.<T>, <Vd>.<T>>, #0.0. */
|
|
|
|
|
#define QL_V2SAMEH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (V_4H, V_4H), \
|
|
|
|
|
QLF2 (V_8H, V_8H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. REV16 <Vd>.<T>, <Vn>.<T>. */
|
|
|
|
|
#define QL_V2SAMEB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8B ), \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SADDLP <Vd>.<Ta>, <Vn>.<Tb>. */
|
|
|
|
|
#define QL_V2PAIRWISELONGBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4H , V_8B ), \
|
|
|
|
|
QLF2(V_8H , V_16B), \
|
|
|
|
|
QLF2(V_2S , V_4H ), \
|
|
|
|
|
QLF2(V_4S , V_8H ), \
|
|
|
|
|
QLF2(V_1D , V_2S ), \
|
|
|
|
|
QLF2(V_2D , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHLL<Q> <Vd>.<Ta>, <Vn>.<Tb>, #<shift>. */
|
|
|
|
|
#define QL_V2LONGBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8H , V_8B ), \
|
|
|
|
|
QLF2(V_4S , V_4H ), \
|
|
|
|
|
QLF2(V_2D , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHLL<Q> <Vd>.<Ta>, <Vn>.<Tb>, #<shift>. */
|
|
|
|
|
#define QL_V2LONGBHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8H , V_16B), \
|
|
|
|
|
QLF2(V_4S , V_8H ), \
|
|
|
|
|
QLF2(V_2D , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* */
|
|
|
|
|
#define QL_V3SAME \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_16B, V_16B, V_16B), \
|
|
|
|
|
QLF3(V_4H , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_8H , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_2D ) \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHADD. */
|
|
|
|
|
#define QL_V3SAMEBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_16B, V_16B, V_16B), \
|
|
|
|
|
QLF3(V_4H , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_8H , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTXN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTXN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4H , V_4S ), \
|
|
|
|
|
QLF2(V_2S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8H , V_4S ), \
|
|
|
|
|
QLF2(V_4S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTL<Q> <Vd>.<Ta>, <Vn>.<Tb>. */
|
|
|
|
|
#define QL_V2LONGHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4S , V_4H ), \
|
|
|
|
|
QLF2(V_2D , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FCVTL<Q> <Vd>.<Ta>, <Vn>.<Tb>. */
|
|
|
|
|
#define QL_V2LONGHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4S , V_8H ), \
|
|
|
|
|
QLF2(V_2D , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. XTN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8H ), \
|
|
|
|
|
QLF2(V_4H , V_4S ), \
|
|
|
|
|
QLF2(V_2S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. XTN<Q> <Vd>.<Tb>, <Vn>.<Ta>. */
|
|
|
|
|
#define QL_V2NARRBHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_16B, V_8H ), \
|
|
|
|
|
QLF2(V_8H , V_4S ), \
|
|
|
|
|
QLF2(V_4S , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ORR. */
|
|
|
|
|
#define QL_V2SAMEB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B , V_8B ), \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. AESE. */
|
|
|
|
|
#define QL_V2SAME16B \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA1SU1. */
|
|
|
|
|
#define QL_V2SAME4S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA1SU0. */
|
|
|
|
|
#define QL_V3SAME4S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_4S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHADD. */
|
|
|
|
|
#define QL_V3SAMEB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_16B, V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. EXT <Vd>.<T>, <Vn>.<T>, <Vm>.<T>, #<index>. */
|
|
|
|
|
#define QL_VEXT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(V_8B , V_8B , V_8B , imm_0_7), \
|
|
|
|
|
QLF4(V_16B, V_16B, V_16B, imm_0_15), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. . */
|
|
|
|
|
#define QL_V3SAMEHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4H , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_8H , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* */
|
|
|
|
|
#define QL_V3SAMESD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S , V_2S , V_2S ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_2D ) \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64] Add ARMv8.3 FCMLA and FCADD instructions
Add support for FCMLA and FCADD complex arithmetic SIMD instructions.
FCMLA has an indexed element variant where the index range has to be
treated specially because a complex number takes two elements and the
indexed vector size depends on the other operands.
These complex number SIMD instructions are part of ARMv8.3
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
include/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* opcode/aarch64.h (enum aarch64_opnd): Add AARCH64_OPND_IMM_ROT1,
AARCH64_OPND_IMM_ROT2, AARCH64_OPND_IMM_ROT3.
(enum aarch64_op): Add OP_FCMLA_ELEM.
opcodes/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (QL_V3SAMEHSD_ROT, QL_ELEMENT_ROT): Define.
(aarch64_feature_simd_v8_3, SIMD_V8_3): Define.
(aarch64_opcode_table): Add fcmla and fcadd.
(AARCH64_OPERANDS): Add IMM_ROT{1,2,3}.
* aarch64-asm.h (aarch64_ins_imm_rotate): Declare.
* aarch64-asm.c (aarch64_ins_imm_rotate): Define.
* aarch64-dis.h (aarch64_ext_imm_rotate): Declare.
* aarch64-dis.c (aarch64_ext_imm_rotate): Define.
* aarch64-opc.h (enum aarch64_field_kind): Add FLD_rotate{1,2,3}.
* aarch64-opc.c (fields): Add FLD_rotate{1,2,3}.
(operand_general_constraint_met_p): Rotate and index range check.
(aarch64_print_operand): Handle rotate operand.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
gas/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* config/tc-aarch64.c (parse_operands): Handle AARCH64_OPND_IMM_ROT*.
* testsuite/gas/aarch64/advsimd-armv8_3.d: New.
* testsuite/gas/aarch64/advsimd-armv8_3.s: New.
* testsuite/gas/aarch64/illegal-fcmla.s: New.
* testsuite/gas/aarch64/illegal-fcmla.l: New.
* testsuite/gas/aarch64/illegal-fcmla.d: New.
2016-11-18 11:02:16 +01:00
|
|
|
|
/* e.g. FCMLA <Vd>.<T>, <Vn>.<T>, <Vm>.<T>, #<rotate>. */
|
|
|
|
|
#define QL_V3SAMEHSD_ROT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4 (V_4H, V_4H, V_4H, NIL), \
|
|
|
|
|
QLF4 (V_8H, V_8H, V_8H, NIL), \
|
|
|
|
|
QLF4 (V_2S, V_2S, V_2S, NIL), \
|
|
|
|
|
QLF4 (V_4S, V_4S, V_4S, NIL), \
|
|
|
|
|
QLF4 (V_2D, V_2D, V_2D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64][PATCH 2/14] Support ARMv8.2 FP16 Vector Three Same instructions.
ARMv8.2 adds 16-bit floating point operations as an optional extension
to floating point and Adv.SIMD support. This patch adds FP16
instructions to the group Vector Three Register Same, making them
available when +simd+fp16 is enabled.
The instructions added are: FMAXNM, FMAXNMP, FNMINNM, FMINNMP, FMLA,
FMLS, FADD, FADDP, FSUB, FABD, FMULX, FMUL, FCMEQ, FCMGE, FCMGT, FACGE,
FACGT, FMAX, FMAXP, FMIN, FMINP, FRECPS, FDIV and FRSQRTS.
The general form for these instructions is
<OP> <Vd>.<T>, <Vs>.<T>, <Vm>.<T>
where T is 4h or 8h.
gas/testsuite/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* gas/aarch64/advsimd-fp16.d: New.
* gas/aarch64/advsimd-fp16.s: New.
opcodes/
2015-12-14 Matthew Wahab <matthew.wahab@arm.com>
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
* aarch64-tbl.h (QL_V3SAMEH): New.
(aarch64_opcode_table): Add fp16 versions of fmaxnm, fmla, fadd,
fmulx, fcmeq, fmax, frecps, fminnm, fmls, fsub, fmin, frsqrts,
fmaxnmp, faddp, fmul, fcmge, facge, fmaxp, fdiv, fminnmp, fabd,
fcmgt, facgt and fminp to the vector three same group.
Change-Id: I3f1c5fe82ca73f7a17fe5329cf2b0de03c94328c
2015-12-14 17:44:02 +01:00
|
|
|
|
/* e.g. FMAXNM <Vd>.<T>, <Vn>.<T>, <Vm>.<T>. */
|
|
|
|
|
#define QL_V3SAMEH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (V_4H , V_4H , V_4H ), \
|
|
|
|
|
QLF3 (V_8H , V_8H , V_8H ), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SQDMLAL<Q> <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3LONGHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_2D , V_2S , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SQDMLAL<Q> <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3LONGHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2D , V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SADDL<Q> <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3LONGBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_8B , V_8B ), \
|
|
|
|
|
QLF3(V_4S , V_4H , V_4H ), \
|
|
|
|
|
QLF3(V_2D , V_2S , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SADDL<Q> <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3LONGBHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_16B , V_16B ), \
|
|
|
|
|
QLF3(V_4S , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_2D , V_4S , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SADDW<Q> <Vd>.<Ta>, <Vn>.<Ta>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3WIDEBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_8H , V_8B ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_4H ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_2S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SADDW<Q> <Vd>.<Ta>, <Vn>.<Ta>, <Vm>.<Tb>. */
|
|
|
|
|
#define QL_V3WIDEBHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_8H , V_16B ), \
|
|
|
|
|
QLF3(V_4S , V_4S , V_8H ), \
|
|
|
|
|
QLF3(V_2D , V_2D , V_4S ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ADDHN<Q> <Vd>.<Tb>, <Vn>.<Ta>, <Vm>.<Ta>. */
|
|
|
|
|
#define QL_V3NARRBHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8B , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_4H , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_2S , V_2D , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ADDHN<Q> <Vd>.<Tb>, <Vn>.<Ta>, <Vm>.<Ta>. */
|
|
|
|
|
#define QL_V3NARRBHS2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_16B , V_8H , V_8H ), \
|
|
|
|
|
QLF3(V_8H , V_4S , V_4S ), \
|
|
|
|
|
QLF3(V_4S , V_2D , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PMULL. */
|
|
|
|
|
#define QL_V3LONGB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_8B , V_8B ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PMULL crypto. */
|
|
|
|
|
#define QL_V3LONGD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_1Q , V_1D , V_1D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PMULL2. */
|
|
|
|
|
#define QL_V3LONGB2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_8H , V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PMULL2 crypto. */
|
|
|
|
|
#define QL_V3LONGD2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_1Q , V_2D , V_2D ), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA1C. */
|
|
|
|
|
#define QL_SHAUPT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_Q, S_S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA256H2. */
|
|
|
|
|
#define QL_SHA256UPT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_Q, S_Q, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDXRB <Wt>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_W1_LDST_EXC \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDXR <Xt>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_R1NIL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, NIL), \
|
|
|
|
|
QLF2(X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STXRB <Ws>, <Wt>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_W2_LDST_EXC \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W, W, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STXR <Ws>, <Xt>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_R2_LDST_EXC \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W, W, NIL), \
|
|
|
|
|
QLF3(W, X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2016-11-18 10:49:06 +01:00
|
|
|
|
/* e.g. LDRAA <Xt>, [<Xn|SP>{,#imm}]. */
|
|
|
|
|
#define QL_X1NIL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. LDXP <Xt1>, <Xt2>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_R2NIL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W, W, NIL), \
|
|
|
|
|
QLF3(X, X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2014-09-03 15:40:41 +02:00
|
|
|
|
/* e.g. CASP <Xt1>, <Xt1+1>, <Xt2>, <Xt2+1>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_R4NIL \
|
|
|
|
|
{ \
|
|
|
|
|
QLF5(W, W, W, W, NIL), \
|
|
|
|
|
QLF5(X, X, X, X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. STXP <Ws>, <Xt1>, <Xt2>, [<Xn|SP>{,#0}]. */
|
|
|
|
|
#define QL_R3_LDST_EXC \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W, W, W, NIL), \
|
|
|
|
|
QLF4(W, X, X, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STR <Qt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B, S_B), \
|
|
|
|
|
QLF2(S_H, S_H), \
|
|
|
|
|
QLF2(S_S, S_S), \
|
|
|
|
|
QLF2(S_D, S_D), \
|
|
|
|
|
QLF2(S_Q, S_Q), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STR <Xt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_R \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, S_S), \
|
|
|
|
|
QLF2(X, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STRB <Wt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_W8 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, S_B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDRSB <Wt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_R8 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, S_B), \
|
|
|
|
|
QLF2(X, S_B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STRH <Wt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_W16 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDRSW <Xt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_X32 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LDRSH <Wt>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_R16 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, S_H), \
|
|
|
|
|
QLF2(X, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. PRFM <prfop>, [<Xn|SP>, <R><m>{, <extend> {<amount>}}]. */
|
|
|
|
|
#define QL_LDST_PRFM \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(NIL, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
2018-11-12 14:14:00 +01:00
|
|
|
|
/* e.g. LDG <Xt>, [<Xn|SP>{, #<simm>}]. */
|
|
|
|
|
#define QL_LDG \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X, imm_tag), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. LDPSW <Xt1>, <Xt2>, [<Xn|SP>{, #<imm>}]. */
|
|
|
|
|
#define QL_LDST_PAIR_X32 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X, X, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
/* e.g. STGP <Xt1>, <Xt2>, [<Xn|SP>{, #<imm>}]. */
|
|
|
|
|
#define QL_STGP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X, X, imm_tag), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. STP <Wt1>, <Wt2>, [<Xn|SP>, #<imm>]!. */
|
|
|
|
|
#define QL_LDST_PAIR_R \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W, W, S_S), \
|
|
|
|
|
QLF3(X, X, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STNP <Qt1>, <Qt2>, [<Xn|SP>{, #<imm>}]. */
|
|
|
|
|
#define QL_LDST_PAIR_FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S, S_S, S_S), \
|
|
|
|
|
QLF3(S_D, S_D, S_D), \
|
|
|
|
|
QLF3(S_Q, S_Q, S_Q), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LD3 {<Vt>.<T>, <Vt2>.<T>, <Vt3>.<T>}, [<Xn|SP>]. */
|
|
|
|
|
#define QL_SIMD_LDST \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B, NIL), \
|
|
|
|
|
QLF2(V_16B, NIL), \
|
|
|
|
|
QLF2(V_4H, NIL), \
|
|
|
|
|
QLF2(V_8H, NIL), \
|
|
|
|
|
QLF2(V_2S, NIL), \
|
|
|
|
|
QLF2(V_4S, NIL), \
|
|
|
|
|
QLF2(V_2D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LD1 {<Vt>.<T>, <Vt2>.<T>, <Vt3>.<T>}, [<Xn|SP>]. */
|
|
|
|
|
#define QL_SIMD_LDST_ANY \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_8B, NIL), \
|
|
|
|
|
QLF2(V_16B, NIL), \
|
|
|
|
|
QLF2(V_4H, NIL), \
|
|
|
|
|
QLF2(V_8H, NIL), \
|
|
|
|
|
QLF2(V_2S, NIL), \
|
|
|
|
|
QLF2(V_4S, NIL), \
|
|
|
|
|
QLF2(V_1D, NIL), \
|
|
|
|
|
QLF2(V_2D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. LD4 {<Vt>.<T>, <Vt2a>.<T>, <Vt3a>.<T>, <Vt4a>.<T>}[<index>], [<Xn|SP>]. */
|
|
|
|
|
#define QL_SIMD_LDSTONE \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B, NIL), \
|
|
|
|
|
QLF2(S_H, NIL), \
|
|
|
|
|
QLF2(S_S, NIL), \
|
|
|
|
|
QLF2(S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. ADDV <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_XLANES \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B, V_8B), \
|
|
|
|
|
QLF2(S_B, V_16B), \
|
|
|
|
|
QLF2(S_H, V_4H), \
|
|
|
|
|
QLF2(S_H, V_8H), \
|
|
|
|
|
QLF2(S_S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMINV <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_XLANES_FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:16:50 +01:00
|
|
|
|
/* e.g. FMINV <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_XLANES_FP_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (S_H, V_4H), \
|
|
|
|
|
QLF2 (S_H, V_8H), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. SADDLV <V><d>, <Vn>.<T>. */
|
|
|
|
|
#define QL_XLANES_L \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H, V_8B), \
|
|
|
|
|
QLF2(S_H, V_16B), \
|
|
|
|
|
QLF2(S_S, V_4H), \
|
|
|
|
|
QLF2(S_S, V_8H), \
|
|
|
|
|
QLF2(S_D, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MUL <Vd>.<T>, <Vn>.<T>, <Vm>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_ELEMENT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4H, V_4H, S_H), \
|
|
|
|
|
QLF3(V_8H, V_8H, S_H), \
|
|
|
|
|
QLF3(V_2S, V_2S, S_S), \
|
|
|
|
|
QLF3(V_4S, V_4S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SMLAL <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_ELEMENT_L \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_4H, S_H), \
|
|
|
|
|
QLF3(V_2D, V_2S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SMLAL2 <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_ELEMENT_L2 \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_8H, S_H), \
|
|
|
|
|
QLF3(V_2D, V_4S, S_S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMLA <V><d>, <V><n>, <Vm>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_ELEMENT_FP \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S, V_2S, S_S), \
|
|
|
|
|
QLF3(V_4S, V_4S, S_S), \
|
|
|
|
|
QLF3(V_2D, V_2D, S_D), \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:01:56 +01:00
|
|
|
|
/* e.g. FMLA <V><d>, <V><n>, <Vm>.<Ts>[<index>]. */
|
|
|
|
|
#define QL_ELEMENT_FP_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3 (V_4H, V_4H, S_H), \
|
|
|
|
|
QLF3 (V_8H, V_8H, S_H), \
|
|
|
|
|
}
|
|
|
|
|
|
[AArch64] Add ARMv8.3 FCMLA and FCADD instructions
Add support for FCMLA and FCADD complex arithmetic SIMD instructions.
FCMLA has an indexed element variant where the index range has to be
treated specially because a complex number takes two elements and the
indexed vector size depends on the other operands.
These complex number SIMD instructions are part of ARMv8.3
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
include/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* opcode/aarch64.h (enum aarch64_opnd): Add AARCH64_OPND_IMM_ROT1,
AARCH64_OPND_IMM_ROT2, AARCH64_OPND_IMM_ROT3.
(enum aarch64_op): Add OP_FCMLA_ELEM.
opcodes/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (QL_V3SAMEHSD_ROT, QL_ELEMENT_ROT): Define.
(aarch64_feature_simd_v8_3, SIMD_V8_3): Define.
(aarch64_opcode_table): Add fcmla and fcadd.
(AARCH64_OPERANDS): Add IMM_ROT{1,2,3}.
* aarch64-asm.h (aarch64_ins_imm_rotate): Declare.
* aarch64-asm.c (aarch64_ins_imm_rotate): Define.
* aarch64-dis.h (aarch64_ext_imm_rotate): Declare.
* aarch64-dis.c (aarch64_ext_imm_rotate): Define.
* aarch64-opc.h (enum aarch64_field_kind): Add FLD_rotate{1,2,3}.
* aarch64-opc.c (fields): Add FLD_rotate{1,2,3}.
(operand_general_constraint_met_p): Rotate and index range check.
(aarch64_print_operand): Handle rotate operand.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
gas/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* config/tc-aarch64.c (parse_operands): Handle AARCH64_OPND_IMM_ROT*.
* testsuite/gas/aarch64/advsimd-armv8_3.d: New.
* testsuite/gas/aarch64/advsimd-armv8_3.s: New.
* testsuite/gas/aarch64/illegal-fcmla.s: New.
* testsuite/gas/aarch64/illegal-fcmla.l: New.
* testsuite/gas/aarch64/illegal-fcmla.d: New.
2016-11-18 11:02:16 +01:00
|
|
|
|
/* e.g. FCMLA <Vd>.<T>, <Vn>.<T>, <Vm>.<Ts>[<index>], #<rotate>. */
|
|
|
|
|
#define QL_ELEMENT_ROT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4 (V_4H, V_4H, S_H, NIL), \
|
|
|
|
|
QLF4 (V_8H, V_8H, S_H, NIL), \
|
|
|
|
|
QLF4 (V_4S, V_4S, S_S, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. MOVI <Vd>.4S, #<imm8> {, LSL #<amount>}. */
|
|
|
|
|
#define QL_SIMD_IMM_S0W \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S, LSL), \
|
|
|
|
|
QLF2(V_4S, LSL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MOVI <Vd>.4S, #<imm8>, MSL #<amount>. */
|
|
|
|
|
#define QL_SIMD_IMM_S1W \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S, MSL), \
|
|
|
|
|
QLF2(V_4S, MSL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. MOVI <Vd>.4H, #<imm8> {, LSL #<amount>}. */
|
|
|
|
|
#define QL_SIMD_IMM_S0H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_4H, LSL), \
|
|
|
|
|
QLF2(V_8H, LSL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMOV <Vd>.<T>, #<imm>. */
|
|
|
|
|
#define QL_SIMD_IMM_S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2S, NIL), \
|
|
|
|
|
QLF2(V_4S, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2013-01-17 17:09:44 +01:00
|
|
|
|
/* e.g. MOVI <Vd>.8B, #<imm8> {, LSL #<amount>}. */
|
2012-08-13 16:52:54 +02:00
|
|
|
|
#define QL_SIMD_IMM_B \
|
|
|
|
|
{ \
|
2013-01-17 17:09:44 +01:00
|
|
|
|
QLF2(V_8B, LSL), \
|
|
|
|
|
QLF2(V_16B, LSL), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
}
|
|
|
|
|
/* e.g. MOVI <Dd>, #<imm>. */
|
|
|
|
|
#define QL_SIMD_IMM_D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-14 18:22:36 +01:00
|
|
|
|
/* e.g. FMOV <Vd>.<T>, #<imm>. */
|
|
|
|
|
#define QL_SIMD_IMM_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2 (V_4H, NIL), \
|
|
|
|
|
QLF2 (V_8H, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* e.g. MOVI <Vd>.2D, #<imm>. */
|
|
|
|
|
#define QL_SIMD_IMM_V2D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2D, NIL), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
|
|
|
|
|
/* The naming convention for SVE macros is:
|
|
|
|
|
|
|
|
|
|
OP_SVE_<operands>[_<sizes>]*
|
|
|
|
|
|
|
|
|
|
<operands> contains one character per operand, using the following scheme:
|
|
|
|
|
|
|
|
|
|
- U: the operand is unqualified (NIL).
|
|
|
|
|
|
|
|
|
|
- [BHSD]: the operand has a S_[BHSD] qualifier and the choice of
|
|
|
|
|
qualifier is the same for all variants. This is used for both
|
|
|
|
|
.[BHSD] suffixes on an SVE predicate or vector register and
|
|
|
|
|
scalar FPRs of the form [BHSD]<number>.
|
|
|
|
|
|
|
|
|
|
- [WX]: the operand has a [WX] qualifier and the choice of qualifier
|
|
|
|
|
is the same for all variants.
|
|
|
|
|
|
|
|
|
|
- [ZM]: the operand has a /[ZM] suffix and the choice of suffix
|
|
|
|
|
is the same for all variants.
|
|
|
|
|
|
|
|
|
|
- V: the operand has a S_[BHSD] qualifier and the choice of qualifier
|
|
|
|
|
is not the same for all variants.
|
|
|
|
|
|
|
|
|
|
- R: the operand has a [WX] qualifier and the choice of qualifier is
|
|
|
|
|
not the same for all variants.
|
|
|
|
|
|
|
|
|
|
- P: the operand has a /[ZM] suffix and the choice of suffix is not
|
|
|
|
|
the same for all variants.
|
|
|
|
|
|
|
|
|
|
The _<sizes>, if present, give the subset of [BHSD] that are accepted
|
|
|
|
|
by the V entries in <operands>. */
|
|
|
|
|
#define OP_SVE_B \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BBBU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,S_B,S_B,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BMB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_M,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BPB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_Z,S_B), \
|
|
|
|
|
QLF3(S_B,P_M,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BUB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,NIL,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BUBB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,NIL,S_B,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BUU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BZ \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,P_Z), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BZB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_Z,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BZBB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_Z,S_B,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_BZU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_Z,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DDD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,S_D,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DMD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DMH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,P_M,S_H), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DMS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,P_M,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DUD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DUU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DUV_BHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,NIL,S_B), \
|
|
|
|
|
QLF3(S_D,NIL,S_H), \
|
|
|
|
|
QLF3(S_D,NIL,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DUV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,NIL,S_B), \
|
|
|
|
|
QLF3(S_D,NIL,S_H), \
|
|
|
|
|
QLF3(S_D,NIL,S_S), \
|
|
|
|
|
QLF3(S_D,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DZD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,P_Z,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_DZU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,P_Z,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_HB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H,S_B), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_HMH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,P_M,S_H), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_HMD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_HMS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,P_M,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_HU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_HUU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_HZU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,P_Z,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_RR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,W), \
|
|
|
|
|
QLF2(X,X), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_RURV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(W,NIL,W,S_B), \
|
|
|
|
|
QLF4(W,NIL,W,S_H), \
|
|
|
|
|
QLF4(W,NIL,W,S_S), \
|
|
|
|
|
QLF4(X,NIL,X,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_RUV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(W,NIL,S_B), \
|
|
|
|
|
QLF3(W,NIL,S_H), \
|
|
|
|
|
QLF3(W,NIL,S_S), \
|
|
|
|
|
QLF3(X,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SMD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SMH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_M,S_H), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SMS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_M,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SUS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,NIL,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SUU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SZS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_Z,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_SZU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_Z,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_UB \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(NIL,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_UUD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(NIL,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_UUS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(NIL,NIL,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMR_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_M,W), \
|
|
|
|
|
QLF3(S_H,P_M,W), \
|
|
|
|
|
QLF3(S_S,P_M,W), \
|
|
|
|
|
QLF3(S_D,P_M,X), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VMU_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF3(S_H,P_M,NIL), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF3(S_S,P_M,NIL), \
|
|
|
|
|
QLF3(S_D,P_M,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMVD_BHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_M,S_B,S_D), \
|
|
|
|
|
QLF4(S_H,P_M,S_H,S_D), \
|
|
|
|
|
QLF4(S_S,P_M,S_S,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMVU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_M,S_B,NIL), \
|
|
|
|
|
QLF4(S_H,P_M,S_H,NIL), \
|
|
|
|
|
QLF4(S_S,P_M,S_S,NIL), \
|
|
|
|
|
QLF4(S_D,P_M,S_D,NIL), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VMVU_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF4(S_H,P_M,S_H,NIL), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF4(S_S,P_M,S_S,NIL), \
|
|
|
|
|
QLF4(S_D,P_M,S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMVV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_M,S_B,S_B), \
|
|
|
|
|
QLF4(S_H,P_M,S_H,S_H), \
|
|
|
|
|
QLF4(S_S,P_M,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,P_M,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VMVV_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_H,P_M,S_H,S_H), \
|
|
|
|
|
QLF4(S_S,P_M,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,P_M,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_VMVV_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_S,P_M,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,P_M,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VMVVU_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF5(S_H,P_M,S_H,S_H,NIL), \
|
|
|
|
|
QLF5(S_S,P_M,S_S,S_S,NIL), \
|
|
|
|
|
QLF5(S_D,P_M,S_D,S_D,NIL), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_VMV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_M,S_B), \
|
|
|
|
|
QLF3(S_H,P_M,S_H), \
|
|
|
|
|
QLF3(S_S,P_M,S_S), \
|
|
|
|
|
QLF3(S_D,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMV_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,P_M,S_H), \
|
|
|
|
|
QLF3(S_S,P_M,S_S), \
|
|
|
|
|
QLF3(S_D,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VMV_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,P_M,S_S), \
|
|
|
|
|
QLF3(S_D,P_M,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VM_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF2(S_H,P_M), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF2(S_S,P_M), \
|
|
|
|
|
QLF2(S_D,P_M), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VPU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_Z,NIL), \
|
|
|
|
|
QLF3(S_B,P_M,NIL), \
|
|
|
|
|
QLF3(S_H,P_Z,NIL), \
|
|
|
|
|
QLF3(S_H,P_M,NIL), \
|
|
|
|
|
QLF3(S_S,P_Z,NIL), \
|
|
|
|
|
QLF3(S_S,P_M,NIL), \
|
|
|
|
|
QLF3(S_D,P_Z,NIL), \
|
|
|
|
|
QLF3(S_D,P_M,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VPV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,P_Z,S_B), \
|
|
|
|
|
QLF3(S_B,P_M,S_B), \
|
|
|
|
|
QLF3(S_H,P_Z,S_H), \
|
|
|
|
|
QLF3(S_H,P_M,S_H), \
|
|
|
|
|
QLF3(S_S,P_Z,S_S), \
|
|
|
|
|
QLF3(S_S,P_M,S_S), \
|
|
|
|
|
QLF3(S_D,P_Z,S_D), \
|
|
|
|
|
QLF3(S_D,P_M,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VRR_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,W,W), \
|
|
|
|
|
QLF3(S_H,W,W), \
|
|
|
|
|
QLF3(S_S,W,W), \
|
|
|
|
|
QLF3(S_D,X,X), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VRU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,W,NIL), \
|
|
|
|
|
QLF3(S_H,W,NIL), \
|
|
|
|
|
QLF3(S_S,W,NIL), \
|
|
|
|
|
QLF3(S_D,X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VR_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,W), \
|
|
|
|
|
QLF2(S_H,W), \
|
|
|
|
|
QLF2(S_S,W), \
|
|
|
|
|
QLF2(S_D,X), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VUR_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,NIL,W), \
|
|
|
|
|
QLF3(S_H,NIL,W), \
|
|
|
|
|
QLF3(S_S,NIL,W), \
|
|
|
|
|
QLF3(S_D,NIL,X), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VUU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,NIL,NIL), \
|
|
|
|
|
QLF3(S_H,NIL,NIL), \
|
|
|
|
|
QLF3(S_S,NIL,NIL), \
|
|
|
|
|
QLF3(S_D,NIL,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VUVV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,NIL,S_B,S_B), \
|
|
|
|
|
QLF4(S_H,NIL,S_H,S_H), \
|
|
|
|
|
QLF4(S_S,NIL,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,NIL,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VUVV_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF4(S_H,NIL,S_H,S_H), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF4(S_S,NIL,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,NIL,S_D,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VUV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,NIL,S_B), \
|
|
|
|
|
QLF3(S_H,NIL,S_H), \
|
|
|
|
|
QLF3(S_S,NIL,S_S), \
|
|
|
|
|
QLF3(S_D,NIL,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VUV_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,NIL,S_H), \
|
|
|
|
|
QLF3(S_S,NIL,S_S), \
|
|
|
|
|
QLF3(S_D,NIL,S_D), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_VUV_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,NIL,S_S), \
|
|
|
|
|
QLF3(S_D,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,NIL), \
|
|
|
|
|
QLF2(S_H,NIL), \
|
|
|
|
|
QLF2(S_S,NIL), \
|
|
|
|
|
QLF2(S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VU_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H,NIL), \
|
|
|
|
|
QLF2(S_S,NIL), \
|
|
|
|
|
QLF2(S_D,NIL), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VU_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF2(S_H,NIL), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF2(S_S,NIL), \
|
|
|
|
|
QLF2(S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVD_BHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,S_B,S_D), \
|
|
|
|
|
QLF3(S_H,S_H,S_D), \
|
|
|
|
|
QLF3(S_S,S_S,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,S_B,NIL), \
|
|
|
|
|
QLF3(S_H,S_H,NIL), \
|
|
|
|
|
QLF3(S_S,S_S,NIL), \
|
|
|
|
|
QLF3(S_D,S_D,NIL), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VVVU_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_H,S_H,S_H,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVVU_S \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_S,S_S,S_S,NIL), \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVVU_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_H,S_H,S_H,NIL), \
|
|
|
|
|
QLF4(S_S,S_S,S_S,NIL), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF4(S_D,S_D,S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,S_B,S_B), \
|
|
|
|
|
QLF3(S_H,S_H,S_H), \
|
|
|
|
|
QLF3(S_S,S_S,S_S), \
|
|
|
|
|
QLF3(S_D,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VVV_D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,S_D,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_D_H \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_D,S_H,S_H), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_H \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF3(S_H,S_H,S_H), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_H,S_H,S_H), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF3(S_S,S_S,S_S), \
|
|
|
|
|
QLF3(S_D,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VVV_S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,S_S,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_S_B \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,S_B,S_B), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VVV_SD_BH \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_S,S_B,S_B), \
|
|
|
|
|
QLF3(S_D,S_H,S_H), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_VV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,S_B), \
|
|
|
|
|
QLF2(S_H,S_H), \
|
|
|
|
|
QLF2(S_S,S_S), \
|
|
|
|
|
QLF2(S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VV_BHSDQ \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_B,S_B), \
|
|
|
|
|
QLF2(S_H,S_H), \
|
|
|
|
|
QLF2(S_S,S_S), \
|
|
|
|
|
QLF2(S_D,S_D), \
|
|
|
|
|
QLF2(S_Q,S_Q), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VV_HSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H,S_H), \
|
|
|
|
|
QLF2(S_S,S_S), \
|
|
|
|
|
QLF2(S_D,S_D), \
|
|
|
|
|
}
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define OP_SVE_VV_HSD_BHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_H,S_B), \
|
|
|
|
|
QLF2(S_S,S_H), \
|
|
|
|
|
QLF2(S_D,S_S), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VV_SD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(S_S,S_S), \
|
|
|
|
|
QLF2(S_D,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VWW_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,W,W), \
|
|
|
|
|
QLF3(S_H,W,W), \
|
|
|
|
|
QLF3(S_S,W,W), \
|
|
|
|
|
QLF3(S_D,W,W), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VXX_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_B,X,X), \
|
|
|
|
|
QLF3(S_H,X,X), \
|
|
|
|
|
QLF3(S_S,X,X), \
|
|
|
|
|
QLF3(S_D,X,X), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VZVD_BHS \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_Z,S_B,S_D), \
|
|
|
|
|
QLF4(S_H,P_Z,S_H,S_D), \
|
|
|
|
|
QLF4(S_S,P_Z,S_S,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VZVU_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_Z,S_B,NIL), \
|
|
|
|
|
QLF4(S_H,P_Z,S_H,NIL), \
|
|
|
|
|
QLF4(S_S,P_Z,S_S,NIL), \
|
|
|
|
|
QLF4(S_D,P_Z,S_D,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_VZVV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(S_B,P_Z,S_B,S_B), \
|
|
|
|
|
QLF4(S_H,P_Z,S_H,S_H), \
|
|
|
|
|
QLF4(S_S,P_Z,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,P_Z,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VZVV_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF4(S_H,P_Z,S_H,S_H), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF4(S_S,P_Z,S_S,S_S), \
|
|
|
|
|
QLF4(S_D,P_Z,S_D,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_VZV_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF3(S_H,P_Z,S_H), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF3(S_S,P_Z,S_S), \
|
|
|
|
|
QLF3(S_D,P_Z,S_D), \
|
|
|
|
|
}
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
#define OP_SVE_V_HSD \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
{ \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
QLF1(S_H), \
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
QLF1(S_S), \
|
|
|
|
|
QLF1(S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_WU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_WV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W,S_B), \
|
|
|
|
|
QLF2(W,S_H), \
|
|
|
|
|
QLF2(W,S_S), \
|
|
|
|
|
QLF2(W,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XUV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,NIL,S_B), \
|
|
|
|
|
QLF3(X,NIL,S_H), \
|
|
|
|
|
QLF3(X,NIL,S_S), \
|
|
|
|
|
QLF3(X,NIL,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XVW_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,S_B,W), \
|
|
|
|
|
QLF3(X,S_H,W), \
|
|
|
|
|
QLF3(X,S_S,W), \
|
|
|
|
|
QLF3(X,S_D,W), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XV_BHSD \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X,S_B), \
|
|
|
|
|
QLF2(X,S_H), \
|
|
|
|
|
QLF2(X,S_S), \
|
|
|
|
|
QLF2(X,S_D), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XWU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,W,NIL), \
|
|
|
|
|
}
|
|
|
|
|
#define OP_SVE_XXU \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X,X,NIL), \
|
|
|
|
|
}
|
2017-06-28 12:09:01 +02:00
|
|
|
|
/* e.g. UDOT <Vd>.2S, <Vn>.8B, <Vm>.8B. */
|
|
|
|
|
#define QL_V3DOT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S, V_8B, V_8B), \
|
|
|
|
|
QLF3(V_4S, V_16B, V_16B),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. UDOT <Vd>.2S, <Vn>.8B, <Vm>.4B[<index>]. */
|
|
|
|
|
#define QL_V2DOT \
|
|
|
|
|
{ \
|
2017-12-19 13:05:20 +01:00
|
|
|
|
QLF3(V_2S, V_8B, S_4B),\
|
|
|
|
|
QLF3(V_4S, V_16B, S_4B),\
|
2017-06-28 12:09:01 +02:00
|
|
|
|
}
|
Add the operand encoding types for the new Armv8.2-a back-ported instructions. These are to be used later when the new instructions are added.
opcodes/
* aarch64-tbl.h (QL_SHA512UPT, QL_V2SAME2D, QL_V3SAME2D): New.
(QL_V4SAME16B, QL_V4SAME4S, QL_XAR, QL_SM3TT, QL_V3FML2S): New.
(QL_V3FML4S, QL_V2FML2S, QL_V2FML4S, QL_RMIF, QL_SETF): New.
(QL_STLW, QL_STLX): New.
2017-11-09 16:45:00 +01:00
|
|
|
|
|
|
|
|
|
/* e.g. SHA512H <Qd>, <Qn>, <Vm>.2D. */
|
|
|
|
|
#define QL_SHA512UPT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(S_Q, S_Q, V_2D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA512SU0 <Vd.2D>, <Vn>.2D. */
|
|
|
|
|
#define QL_V2SAME2D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(V_2D, V_2D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SHA512SU1 <Vd>.2D, <Vn>.2D, <Vm>.2D>. */
|
|
|
|
|
#define QL_V3SAME2D \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2D, V_2D, V_2D), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. EOR3 <Vd>.16B, <Vn>.16B, <Vm>.16B, <Va>.16B. */
|
|
|
|
|
#define QL_V4SAME16B \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(V_16B, V_16B, V_16B, V_16B), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SM3SS1 <Vd>.4S, <Vn>.4S, <Vm>.4S, <Va>.4S. */
|
|
|
|
|
#define QL_V4SAME4S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(V_4S, V_4S, V_4S, V_4S), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. XAR <Vd>.2D, <Vn>.2D, <Vm>.2D, #<imm6>. */
|
|
|
|
|
#define QL_XAR \
|
|
|
|
|
{ \
|
|
|
|
|
QLF4(V_2D, V_2D, V_2D, imm_0_63), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SM3TT1A <Vd>.4S, <Vn>.4S, <Vm>.S[<imm2>]. */
|
|
|
|
|
#define QL_SM3TT \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_4S, S_S),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMLAL <Vd>.2S, <Vn>.2H, <Vm>.2H. */
|
|
|
|
|
#define QL_V3FML2S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S, V_2H, V_2H),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMLAL <Vd>.4S, <Vn>.4H, <Vm>.4H. */
|
|
|
|
|
#define QL_V3FML4S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_4H, V_4H),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMLAL <Vd>.2S, <Vn>.2H, <Vm>.H[<index>]. */
|
|
|
|
|
#define QL_V2FML2S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_2S, V_2H, S_H),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. FMLAL <Vd>.4S, <Vn>.4H, <Vm>.H[<index>]. */
|
|
|
|
|
#define QL_V2FML4S \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(V_4S, V_4H, S_H),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. RMIF <Xn>, #<shift>, #<mask>. */
|
|
|
|
|
#define QL_RMIF \
|
|
|
|
|
{ \
|
|
|
|
|
QLF3(X, imm_0_63, imm_0_15),\
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. SETF8 <Wn>. */
|
|
|
|
|
#define QL_SETF \
|
|
|
|
|
{ \
|
|
|
|
|
QLF1(W), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STLURB <Wt>, [<Xn|SP>{,#<simm>}]. */
|
|
|
|
|
#define QL_STLW \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(W, NIL), \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* e.g. STLURB <Xt>, [<Xn|SP>{,#<simm>}]. */
|
|
|
|
|
#define QL_STLX \
|
|
|
|
|
{ \
|
|
|
|
|
QLF2(X, NIL), \
|
|
|
|
|
}
|
2012-08-13 16:52:54 +02:00
|
|
|
|
|
|
|
|
|
/* Opcode table. */
|
|
|
|
|
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_v8 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_fp =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_FP, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_simd =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_SIMD, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_crypto =
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_CRYPTO | AARCH64_FEATURE_AES
|
2017-11-16 17:15:51 +01:00
|
|
|
|
| AARCH64_FEATURE_SHA2 | AARCH64_FEATURE_SIMD | AARCH64_FEATURE_FP, 0);
|
2013-02-28 20:18:40 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_crc =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_CRC, 0);
|
2014-09-03 15:40:41 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_lse =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_LSE, 0);
|
2015-06-02 12:29:15 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_lor =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_LOR, 0);
|
2015-06-02 13:20:00 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_rdma =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_RDMA, 0);
|
2015-12-10 15:05:01 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_ras =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_RAS, 0);
|
2015-11-27 16:25:08 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_v8_2 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2, 0);
|
2015-11-27 16:47:53 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_fp_f16 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_F16 | AARCH64_FEATURE_FP, 0);
|
2015-12-14 17:34:47 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_simd_f16 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_F16 | AARCH64_FEATURE_SIMD, 0);
|
2015-12-11 11:22:40 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_stat_profile =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_PROFILE, 0);
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_sve =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_SVE, 0);
|
[AArch64] Add ARMv8.3 instructions which are in the NOP space
This patch adds support for a subset of the ARMv8.3 pointer authentication
instructions: XPACLRI, PACIA1716, PACIB1716, AUTIA1716, AUTIA1716, PACIAZ,
PACIASP, PACIBZ, PACISP, AUTIAZ, AUTIASP, AUTIBZ, AUTIBSP.
These are aliases to HINT #0x7, HINT #0x8, HINT #0xa, HINT #0xc, HINT #0xe,
HINT #0x18, HINT #0x19, ..., HINT #0x1f respectively.
For more details about pointer authentication in ARMv8.3 see
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (aarch64_feature_v8_3, ARMV8_3, V8_3_INSN): New.
(arch64_opcode_table): Add xpaclri, pacia1716, pacib1716, autia1716,
autib1716, paciaz, paciasp, pacibz, pacibsp, autiaz, autiasp, autibz,
autibsp.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
gas/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/system-3.s: New.
* testsuite/gas/aarch64/system-3.d: New.
* testsuite/gas/aarch64/system.d: Update expected output.
2016-11-11 11:29:07 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_v8_3 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_3, 0);
|
2016-11-18 10:53:45 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_fp_v8_3 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_3 | AARCH64_FEATURE_FP, 0);
|
2017-02-24 19:27:26 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_compnum =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_COMPNUM, 0);
|
2017-01-04 13:27:10 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_rcpc =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_RCPC, 0);
|
2017-06-28 12:09:01 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_dotprod =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2 | AARCH64_FEATURE_DOTPROD, 0);
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_sha2 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8 | AARCH64_FEATURE_SHA2, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_aes =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8 | AARCH64_FEATURE_AES, 0);
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_v8_4 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_4, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_crypto_v8_2 =
|
2017-11-16 17:15:51 +01:00
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2 | AARCH64_FEATURE_CRYPTO
|
|
|
|
|
| AARCH64_FEATURE_SIMD | AARCH64_FEATURE_FP, 0);
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_sm4 =
|
2017-11-16 17:15:51 +01:00
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2 | AARCH64_FEATURE_SM4
|
|
|
|
|
| AARCH64_FEATURE_SIMD | AARCH64_FEATURE_FP, 0);
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_sha3 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2 | AARCH64_FEATURE_SHA2
|
2017-11-16 17:15:51 +01:00
|
|
|
|
| AARCH64_FEATURE_SHA3 | AARCH64_FEATURE_SIMD | AARCH64_FEATURE_FP, 0);
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_fp_16_v8_2 =
|
2017-11-16 17:19:37 +01:00
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_2 | AARCH64_FEATURE_F16_FML
|
|
|
|
|
| AARCH64_FEATURE_F16 | AARCH64_FEATURE_FP, 0);
|
2018-09-26 11:38:59 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_v8_5 =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_5, 0);
|
[PATCH, BINUTILS, AARCH64, 2/9] Add Data procoessing instructions for ARMv8.5-A
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order)
This patch adds the data processing instructions that are new to
ARMv8.5-A.
1) There are 2 instructions: xaflag, axflag, that are added to
manipulate the states of the flag and are used to convert between the
Arm representation and the fcmp representation.
2) The other instructions are rounding instructions which have 8
versions based on whether the floating-point number is a
Single-Precision or Double-Precision number, whether the target
integer is a 32-bit or 64-bit integer and whether the rounding mode is
the ambient rounding mode or to zero. Each of these instruction is
available in both Scalar and Vector forms.
Since both 1) and 2) have separate identification mechanism and it is
permissible that a ARMv8.4 compliant implementation may include any
arbitrary subset of the ARMv8.5 features unless otherwise specified,
new feature bits are added.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_FLAGMANIP): New.
(AARCH64_FEATURE_FRINTTS): New.
(AARCH64_ARCH_V8_5): Add both by default.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_feature_flagmanip): New.
(aarch64_feature_frintts): New.
(FLAGMANIP, FRINTTS): New.
(aarch64_opcode_table): Add entries for xaflag, axflag
and frint[32,64][x,z] instructions.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-dp.s: New.
* testsuite/gas/aarch64/armv8_5-a-dp.d: New.
2018-09-26 11:45:35 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_flagmanip =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_FLAGMANIP, 0);
|
|
|
|
|
static const aarch64_feature_set aarch64_feature_frintts =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_FRINTTS, 0);
|
2018-09-26 11:47:40 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_sb =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_SB, 0);
|
2018-09-26 11:52:51 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_predres =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_PREDRES, 0);
|
[PATCH, BINUTILS, AARCH64, 7/9] Add BTI instruction
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order/bti-branch-target-identification)
The Branch Target Identification instructions (BTI) are allocated to
existing HINT space, using HINT numbers 32, 34, 36, 38, such that
bits[7:6] of the instruction identify the compatibility of the BTI
instruction to different branches.
BTI {<targets>}
where <targets> one of the following, specifying which type of
indirection is allowed:
j : Can be a target of any BR Xn isntruction.
c : Can be a target of any BLR Xn and BR {X16|X17}.
jc: Can be a target of any free branch.
A BTI instruction without any <targets> is the strictest of all and
can not be a target of nay free branch.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_BTI): New.
(AARCH64_ARCH_V8_5): Add AARCH64_FEATURE_BTI by default.
(aarch64_opnd): Add AARCH64_OPND_BTI_TARGET.
(HINT_OPD_CSYNC, HINT_OPD_C, HINT_OPD_J): New macros to
define HINT #imm values.
(HINT_OPD_JC, HINT_OPD_NULL): Likewise.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (HINT_OPD_NOPRINT, HINT_ENCODE): New.
(HINT_FLAG, HINT_VALUE): New macros to encode NO_PRINT flag
with the hint immediate.
* aarch64-opc.c (aarch64_hint_options): New entries for
c, j, jc and default (with HINT_OPD_F_NOPRINT flag) for BTI.
(aarch64_print_operand): Add case for AARCH64_OPND_BTI_TARGET
while checking for HINT_OPD_F_NOPRINT flag.
* aarch64-dis.c (aarch64_ext_hint): Use new HINT_VALUE to
extract value.
* aarch64-tbl.h (aarch64_feature_bti, BTI, BTI_INSN): New.
(aarch64_opcode_table): Add entry for BTI.
(AARCH64_OPERANDS): Add new description for BTI targets.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_bti_operand): New.
(process_omitted_operand): Add case for AARCH64_OPND_BTI_TARGET.
(parse_operands): Likewise.
* testsuite/gas/aarch64/system.d: Update for BTI.
* testsuite/gas/aarch64/bti.s: New.
* testsuite/gas/aarch64/bti.d: New.
* testsuite/gas/aarch64/illegal-bti.d: New.
* testsuite/gas/aarch64/illegal-bti.l: New.
2018-09-26 12:00:49 +02:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_bti =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_BTI, 0);
|
2018-11-12 13:45:30 +01:00
|
|
|
|
static const aarch64_feature_set aarch64_feature_memtag =
|
|
|
|
|
AARCH64_FEATURE (AARCH64_FEATURE_V8_5 | AARCH64_FEATURE_MEMTAG, 0);
|
2018-09-26 11:38:59 +02:00
|
|
|
|
|
2012-08-13 16:52:54 +02:00
|
|
|
|
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define CORE &aarch64_feature_v8
|
|
|
|
|
#define FP &aarch64_feature_fp
|
|
|
|
|
#define SIMD &aarch64_feature_simd
|
|
|
|
|
#define CRYPTO &aarch64_feature_crypto
|
|
|
|
|
#define CRC &aarch64_feature_crc
|
|
|
|
|
#define LSE &aarch64_feature_lse
|
|
|
|
|
#define LOR &aarch64_feature_lor
|
|
|
|
|
#define RDMA &aarch64_feature_rdma
|
|
|
|
|
#define FP_F16 &aarch64_feature_fp_f16
|
2015-12-14 17:34:47 +01:00
|
|
|
|
#define SIMD_F16 &aarch64_feature_simd_f16
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define RAS &aarch64_feature_ras
|
2015-12-11 11:22:40 +01:00
|
|
|
|
#define STAT_PROFILE &aarch64_feature_stat_profile
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define ARMV8_2 &aarch64_feature_v8_2
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define SVE &aarch64_feature_sve
|
[AArch64] Add ARMv8.3 instructions which are in the NOP space
This patch adds support for a subset of the ARMv8.3 pointer authentication
instructions: XPACLRI, PACIA1716, PACIB1716, AUTIA1716, AUTIA1716, PACIAZ,
PACIASP, PACIBZ, PACISP, AUTIAZ, AUTIASP, AUTIBZ, AUTIBSP.
These are aliases to HINT #0x7, HINT #0x8, HINT #0xa, HINT #0xc, HINT #0xe,
HINT #0x18, HINT #0x19, ..., HINT #0x1f respectively.
For more details about pointer authentication in ARMv8.3 see
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (aarch64_feature_v8_3, ARMV8_3, V8_3_INSN): New.
(arch64_opcode_table): Add xpaclri, pacia1716, pacib1716, autia1716,
autib1716, paciaz, paciasp, pacibz, pacibsp, autiaz, autiasp, autibz,
autibsp.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
gas/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/system-3.s: New.
* testsuite/gas/aarch64/system-3.d: New.
* testsuite/gas/aarch64/system.d: Update expected output.
2016-11-11 11:29:07 +01:00
|
|
|
|
#define ARMV8_3 &aarch64_feature_v8_3
|
2016-11-18 10:53:45 +01:00
|
|
|
|
#define FP_V8_3 &aarch64_feature_fp_v8_3
|
2017-02-24 19:27:26 +01:00
|
|
|
|
#define COMPNUM &aarch64_feature_compnum
|
2017-01-04 13:27:10 +01:00
|
|
|
|
#define RCPC &aarch64_feature_rcpc
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
#define SHA2 &aarch64_feature_sha2
|
|
|
|
|
#define AES &aarch64_feature_aes
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
#define ARMV8_4 &aarch64_feature_v8_4
|
|
|
|
|
#define SHA3 &aarch64_feature_sha3
|
|
|
|
|
#define SM4 &aarch64_feature_sm4
|
|
|
|
|
#define CRYPTO_V8_2 &aarch64_feature_crypto_v8_2
|
|
|
|
|
#define FP_F16_V8_2 &aarch64_feature_fp_16_v8_2
|
2017-06-28 12:09:01 +02:00
|
|
|
|
#define DOTPROD &aarch64_feature_dotprod
|
2018-09-26 11:38:59 +02:00
|
|
|
|
#define ARMV8_5 &aarch64_feature_v8_5
|
[PATCH, BINUTILS, AARCH64, 2/9] Add Data procoessing instructions for ARMv8.5-A
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order)
This patch adds the data processing instructions that are new to
ARMv8.5-A.
1) There are 2 instructions: xaflag, axflag, that are added to
manipulate the states of the flag and are used to convert between the
Arm representation and the fcmp representation.
2) The other instructions are rounding instructions which have 8
versions based on whether the floating-point number is a
Single-Precision or Double-Precision number, whether the target
integer is a 32-bit or 64-bit integer and whether the rounding mode is
the ambient rounding mode or to zero. Each of these instruction is
available in both Scalar and Vector forms.
Since both 1) and 2) have separate identification mechanism and it is
permissible that a ARMv8.4 compliant implementation may include any
arbitrary subset of the ARMv8.5 features unless otherwise specified,
new feature bits are added.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_FLAGMANIP): New.
(AARCH64_FEATURE_FRINTTS): New.
(AARCH64_ARCH_V8_5): Add both by default.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_feature_flagmanip): New.
(aarch64_feature_frintts): New.
(FLAGMANIP, FRINTTS): New.
(aarch64_opcode_table): Add entries for xaflag, axflag
and frint[32,64][x,z] instructions.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-dp.s: New.
* testsuite/gas/aarch64/armv8_5-a-dp.d: New.
2018-09-26 11:45:35 +02:00
|
|
|
|
#define FLAGMANIP &aarch64_feature_flagmanip
|
|
|
|
|
#define FRINTTS &aarch64_feature_frintts
|
2018-09-26 11:47:40 +02:00
|
|
|
|
#define SB &aarch64_feature_sb
|
2018-09-26 11:52:51 +02:00
|
|
|
|
#define PREDRES &aarch64_feature_predres
|
[PATCH, BINUTILS, AARCH64, 7/9] Add BTI instruction
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order/bti-branch-target-identification)
The Branch Target Identification instructions (BTI) are allocated to
existing HINT space, using HINT numbers 32, 34, 36, 38, such that
bits[7:6] of the instruction identify the compatibility of the BTI
instruction to different branches.
BTI {<targets>}
where <targets> one of the following, specifying which type of
indirection is allowed:
j : Can be a target of any BR Xn isntruction.
c : Can be a target of any BLR Xn and BR {X16|X17}.
jc: Can be a target of any free branch.
A BTI instruction without any <targets> is the strictest of all and
can not be a target of nay free branch.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_BTI): New.
(AARCH64_ARCH_V8_5): Add AARCH64_FEATURE_BTI by default.
(aarch64_opnd): Add AARCH64_OPND_BTI_TARGET.
(HINT_OPD_CSYNC, HINT_OPD_C, HINT_OPD_J): New macros to
define HINT #imm values.
(HINT_OPD_JC, HINT_OPD_NULL): Likewise.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (HINT_OPD_NOPRINT, HINT_ENCODE): New.
(HINT_FLAG, HINT_VALUE): New macros to encode NO_PRINT flag
with the hint immediate.
* aarch64-opc.c (aarch64_hint_options): New entries for
c, j, jc and default (with HINT_OPD_F_NOPRINT flag) for BTI.
(aarch64_print_operand): Add case for AARCH64_OPND_BTI_TARGET
while checking for HINT_OPD_F_NOPRINT flag.
* aarch64-dis.c (aarch64_ext_hint): Use new HINT_VALUE to
extract value.
* aarch64-tbl.h (aarch64_feature_bti, BTI, BTI_INSN): New.
(aarch64_opcode_table): Add entry for BTI.
(AARCH64_OPERANDS): Add new description for BTI targets.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_bti_operand): New.
(process_omitted_operand): Add case for AARCH64_OPND_BTI_TARGET.
(parse_operands): Likewise.
* testsuite/gas/aarch64/system.d: Update for BTI.
* testsuite/gas/aarch64/bti.s: New.
* testsuite/gas/aarch64/bti.d: New.
* testsuite/gas/aarch64/illegal-bti.d: New.
* testsuite/gas/aarch64/illegal-bti.l: New.
2018-09-26 12:00:49 +02:00
|
|
|
|
#define BTI &aarch64_feature_bti
|
2018-11-12 13:45:30 +01:00
|
|
|
|
#define MEMTAG &aarch64_feature_memtag
|
2016-04-28 10:11:03 +02:00
|
|
|
|
|
2016-08-23 10:01:54 +02:00
|
|
|
|
#define CORE_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, CORE, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-08-23 10:01:54 +02:00
|
|
|
|
#define __FP_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, FP, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-08-23 10:01:54 +02:00
|
|
|
|
#define SIMD_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, SIMD, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define CRYP_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, CRYPTO, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define _CRC_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, CRC, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define _LSE_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, LSE, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define _LOR_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, LOR, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define RDMA_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, RDMA, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define FF16_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, FP_F16, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-04-28 10:11:03 +02:00
|
|
|
|
#define SF16_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, SIMD_F16, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2016-08-23 10:03:19 +02:00
|
|
|
|
#define V8_2_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, ARMV8_2, OPS, QUALS, FLAGS, 0, 0, NULL }
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
#define _SVE_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS,TIED) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, SVE, OPS, QUALS, \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
FLAGS | F_STRICT, 0, TIED, NULL }
|
|
|
|
|
#define _SVE_INSNC(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS,CONSTRAINTS,TIED) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, SVE, OPS, QUALS, \
|
|
|
|
|
FLAGS | F_STRICT, CONSTRAINTS, TIED, NULL }
|
[AArch64] Add ARMv8.3 instructions which are in the NOP space
This patch adds support for a subset of the ARMv8.3 pointer authentication
instructions: XPACLRI, PACIA1716, PACIB1716, AUTIA1716, AUTIA1716, PACIAZ,
PACIASP, PACIBZ, PACISP, AUTIAZ, AUTIASP, AUTIBZ, AUTIBSP.
These are aliases to HINT #0x7, HINT #0x8, HINT #0xa, HINT #0xc, HINT #0xe,
HINT #0x18, HINT #0x19, ..., HINT #0x1f respectively.
For more details about pointer authentication in ARMv8.3 see
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (aarch64_feature_v8_3, ARMV8_3, V8_3_INSN): New.
(arch64_opcode_table): Add xpaclri, pacia1716, pacib1716, autia1716,
autib1716, paciaz, paciasp, pacibz, pacibsp, autiaz, autiasp, autibz,
autibsp.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
gas/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/system-3.s: New.
* testsuite/gas/aarch64/system-3.d: New.
* testsuite/gas/aarch64/system.d: Update expected output.
2016-11-11 11:29:07 +01:00
|
|
|
|
#define V8_3_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, ARMV8_3, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2017-02-24 19:27:26 +01:00
|
|
|
|
#define CNUM_INSN(NAME,OPCODE,MASK,CLASS,OP,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, OP, COMPNUM, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2017-01-04 13:27:10 +01:00
|
|
|
|
#define RCPC_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, RCPC, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
#define SHA2_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, SHA2, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
#define AES_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, AES, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
#define V8_4_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, ARMV8_4, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
#define CRYPTO_V8_2_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, CRYPTO_V8_2, OPS, QUALS, FLAGS, 0, NULL }
|
|
|
|
|
#define SHA3_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, SHA3, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
#define SM4_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, SM4, OPS, QUALS, FLAGS, 0, 0, NULL }
|
Split the ARM Crypto ISA extensions for AES and SHA1+2 into their own options (+aes and +sha2). The reason for the split is because with the introduction of Armv8.4-a the implementation of AES has explicitly been made independent of the implementation of the other crypto extensions.
gas * config/tc-aarch64.c (aarch64_arch_option_table): Add armv8.4-a.
(aarch64_features): Added SM4 and SHA3.
include * opcode/aarch64.h:
(AARCH64_FEATURE_V8_4, AARCH64_FEATURE_SM4): New.
(AARCH64_ARCH_V8_4, AARCH64_FEATURE_SHA3): New.
opcodes * aarch64-tbl.h
(aarch64_feature_v8_4, aarch64_feature_crypto_v8_2): New.
(aarch64_feature_sm4, aarch64_feature_sha3): New.
(aarch64_feature_fp_16_v8_2): New.
(ARMV8_4, SHA3, SM4, CRYPTO_V8_2, FP_F16_V8_2): New.
(V8_4_INSN, CRYPTO_V8_2_INSN): New.
(SHA3_INSN, SM4_INSN, FP16_V8_2_INSN): New.
2017-11-09 12:21:31 +01:00
|
|
|
|
#define FP16_V8_2_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, FP_F16_V8_2, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2017-06-28 12:09:01 +02:00
|
|
|
|
#define DOT_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, DOTPROD, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2018-09-26 11:38:59 +02:00
|
|
|
|
#define V8_5_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, ARMV8_5, OPS, QUALS, FLAGS, 0, 0, NULL }
|
[PATCH, BINUTILS, AARCH64, 2/9] Add Data procoessing instructions for ARMv8.5-A
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order)
This patch adds the data processing instructions that are new to
ARMv8.5-A.
1) There are 2 instructions: xaflag, axflag, that are added to
manipulate the states of the flag and are used to convert between the
Arm representation and the fcmp representation.
2) The other instructions are rounding instructions which have 8
versions based on whether the floating-point number is a
Single-Precision or Double-Precision number, whether the target
integer is a 32-bit or 64-bit integer and whether the rounding mode is
the ambient rounding mode or to zero. Each of these instruction is
available in both Scalar and Vector forms.
Since both 1) and 2) have separate identification mechanism and it is
permissible that a ARMv8.4 compliant implementation may include any
arbitrary subset of the ARMv8.5 features unless otherwise specified,
new feature bits are added.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_FLAGMANIP): New.
(AARCH64_FEATURE_FRINTTS): New.
(AARCH64_ARCH_V8_5): Add both by default.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_feature_flagmanip): New.
(aarch64_feature_frintts): New.
(FLAGMANIP, FRINTTS): New.
(aarch64_opcode_table): Add entries for xaflag, axflag
and frint[32,64][x,z] instructions.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-dp.s: New.
* testsuite/gas/aarch64/armv8_5-a-dp.d: New.
2018-09-26 11:45:35 +02:00
|
|
|
|
#define FLAGMANIP_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, FLAGMANIP, OPS, QUALS, FLAGS, 0, 0, NULL }
|
|
|
|
|
#define FRINTTS_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, FRINTTS, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2018-09-26 11:47:40 +02:00
|
|
|
|
#define SB_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, SB, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2018-09-26 11:52:51 +02:00
|
|
|
|
#define PREDRES_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, PREDRES, OPS, QUALS, FLAGS, 0, 0, NULL }
|
[PATCH, BINUTILS, AARCH64, 7/9] Add BTI instruction
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order/bti-branch-target-identification)
The Branch Target Identification instructions (BTI) are allocated to
existing HINT space, using HINT numbers 32, 34, 36, 38, such that
bits[7:6] of the instruction identify the compatibility of the BTI
instruction to different branches.
BTI {<targets>}
where <targets> one of the following, specifying which type of
indirection is allowed:
j : Can be a target of any BR Xn isntruction.
c : Can be a target of any BLR Xn and BR {X16|X17}.
jc: Can be a target of any free branch.
A BTI instruction without any <targets> is the strictest of all and
can not be a target of nay free branch.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_BTI): New.
(AARCH64_ARCH_V8_5): Add AARCH64_FEATURE_BTI by default.
(aarch64_opnd): Add AARCH64_OPND_BTI_TARGET.
(HINT_OPD_CSYNC, HINT_OPD_C, HINT_OPD_J): New macros to
define HINT #imm values.
(HINT_OPD_JC, HINT_OPD_NULL): Likewise.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (HINT_OPD_NOPRINT, HINT_ENCODE): New.
(HINT_FLAG, HINT_VALUE): New macros to encode NO_PRINT flag
with the hint immediate.
* aarch64-opc.c (aarch64_hint_options): New entries for
c, j, jc and default (with HINT_OPD_F_NOPRINT flag) for BTI.
(aarch64_print_operand): Add case for AARCH64_OPND_BTI_TARGET
while checking for HINT_OPD_F_NOPRINT flag.
* aarch64-dis.c (aarch64_ext_hint): Use new HINT_VALUE to
extract value.
* aarch64-tbl.h (aarch64_feature_bti, BTI, BTI_INSN): New.
(aarch64_opcode_table): Add entry for BTI.
(AARCH64_OPERANDS): Add new description for BTI targets.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_bti_operand): New.
(process_omitted_operand): Add case for AARCH64_OPND_BTI_TARGET.
(parse_operands): Likewise.
* testsuite/gas/aarch64/system.d: Update for BTI.
* testsuite/gas/aarch64/bti.s: New.
* testsuite/gas/aarch64/bti.d: New.
* testsuite/gas/aarch64/illegal-bti.d: New.
* testsuite/gas/aarch64/illegal-bti.l: New.
2018-09-26 12:00:49 +02:00
|
|
|
|
#define BTI_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, BTI, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2018-11-12 13:45:30 +01:00
|
|
|
|
#define MEMTAG_INSN(NAME,OPCODE,MASK,CLASS,OPS,QUALS,FLAGS) \
|
|
|
|
|
{ NAME, OPCODE, MASK, CLASS, 0, MEMTAG, OPS, QUALS, FLAGS, 0, 0, NULL }
|
2012-08-13 16:52:54 +02:00
|
|
|
|
|
|
|
|
|
struct aarch64_opcode aarch64_opcode_table[] =
|
|
|
|
|
{
|
|
|
|
|
/* Add/subtract (with carry). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("adc", 0x1a000000, 0x7fe0fc00, addsub_carry, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("adcs", 0x3a000000, 0x7fe0fc00, addsub_carry, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("sbc", 0x5a000000, 0x7fe0fc00, addsub_carry, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("ngc", 0x5a0003e0, 0x7fe0ffe0, addsub_carry, 0, OP2 (Rd, Rm), QL_I2SAME, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("sbcs", 0x7a000000, 0x7fe0fc00, addsub_carry, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("ngcs", 0x7a0003e0, 0x7fe0ffe0, addsub_carry, 0, OP2 (Rd, Rm), QL_I2SAME, F_ALIAS | F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Add/subtract (extended register). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("add", 0x0b200000, 0x7fe00000, addsub_ext, 0, OP3 (Rd_SP, Rn_SP, Rm_EXT), QL_I3_EXT, F_SF),
|
|
|
|
|
CORE_INSN ("adds", 0x2b200000, 0x7fe00000, addsub_ext, 0, OP3 (Rd, Rn_SP, Rm_EXT), QL_I3_EXT, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("cmn", 0x2b20001f, 0x7fe0001f, addsub_ext, 0, OP2 (Rn_SP, Rm_EXT), QL_I2_EXT, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("sub", 0x4b200000, 0x7fe00000, addsub_ext, 0, OP3 (Rd_SP, Rn_SP, Rm_EXT), QL_I3_EXT, F_SF),
|
|
|
|
|
CORE_INSN ("subs", 0x6b200000, 0x7fe00000, addsub_ext, 0, OP3 (Rd, Rn_SP, Rm_EXT), QL_I3_EXT, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("cmp", 0x6b20001f, 0x7fe0001f, addsub_ext, 0, OP2 (Rn_SP, Rm_EXT), QL_I2_EXT, F_ALIAS | F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Add/subtract (immediate). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("add", 0x11000000, 0x7f000000, addsub_imm, OP_ADD, OP3 (Rd_SP, Rn_SP, AIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("mov", 0x11000000, 0x7ffffc00, addsub_imm, 0, OP2 (Rd_SP, Rn_SP), QL_I2SP, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("adds", 0x31000000, 0x7f000000, addsub_imm, 0, OP3 (Rd, Rn_SP, AIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("cmn", 0x3100001f, 0x7f00001f, addsub_imm, 0, OP2 (Rn_SP, AIMM), QL_R1NIL, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("sub", 0x51000000, 0x7f000000, addsub_imm, 0, OP3 (Rd_SP, Rn_SP, AIMM), QL_R2NIL, F_SF),
|
|
|
|
|
CORE_INSN ("subs", 0x71000000, 0x7f000000, addsub_imm, 0, OP3 (Rd, Rn_SP, AIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("cmp", 0x7100001f, 0x7f00001f, addsub_imm, 0, OP2 (Rn_SP, AIMM), QL_R1NIL, F_ALIAS | F_SF),
|
[BINUTILS, AARCH64, 2/8] Add Tag generation instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag generation instructions from
MTE. These are the following instructions added in this patch:
- IRG <Xd|SP>, <Xn|SP>{, Xm}
- ADDG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- SUBG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- GMI <Xd>, <Xn|SP>, <Xm>
where
<Xd|SP> : Is the 64-bit destination GPR or Stack pointer.
<Xn|SP> : Is the 64-bit source GPR or Stack pointer.
<uimm6> : Is the unsigned immediate, a multiple of 16
in the range 0 to 1008.
<uimm4> : Is the unsigned immediate, in the range 0 to 15.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10 as new enums.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (aarch64_field_kind): New FLD_imm4_3.
(OPD_F_SHIFT_BY_4, operand_need_shift_by_four): New.
* aarch64-opc.c (fields): Add entry for imm4_3.
(operand_general_constraint_met_p): Add cases for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_ADDG): New.
(aarch64_opcode_table): Add addg, subg, irg and gmi.
(AARCH64_OPERANDS): Define UIMM4_ADDG and UIMM10.
* aarch64-asm.c (aarch64_ins_imm): Add case for
operand_need_shift_by_four.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: New.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
* testsuite/gas/aarch64/illegal-memtag.d: Likewise.
2018-11-12 13:52:55 +01:00
|
|
|
|
MEMTAG_INSN ("addg", 0x91800000, 0xffc0c000, addsub_imm, OP4 (Rd_SP, Rn_SP, UIMM10, UIMM4_ADDG), QL_ADDG, 0),
|
|
|
|
|
MEMTAG_INSN ("subg", 0xd1800000, 0xffc0c000, addsub_imm, OP4 (Rd_SP, Rn_SP, UIMM10, UIMM4_ADDG), QL_ADDG, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Add/subtract (shifted register). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("add", 0x0b000000, 0x7f200000, addsub_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("adds", 0x2b000000, 0x7f200000, addsub_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("cmn", 0x2b00001f, 0x7f20001f, addsub_shift, 0, OP2 (Rn, Rm_SFT), QL_I2SAME, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("sub", 0x4b000000, 0x7f200000, addsub_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("neg", 0x4b0003e0, 0x7f2003e0, addsub_shift, 0, OP2 (Rd, Rm_SFT), QL_I2SAME, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("subs", 0x6b000000, 0x7f200000, addsub_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
2018-06-22 13:27:53 +02:00
|
|
|
|
CORE_INSN ("cmp", 0x6b00001f, 0x7f20001f, addsub_shift, 0, OP2 (Rn, Rm_SFT), QL_I2SAME, F_ALIAS | F_SF | F_P1),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("negs", 0x6b0003e0, 0x7f2003e0, addsub_shift, 0, OP2 (Rd, Rm_SFT), QL_I2SAME, F_ALIAS | F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD across lanes. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("saddlv", 0x0e303800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smaxv", 0x0e30a800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sminv", 0x0e31a800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("addv", 0x0e31b800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaddlv", 0x2e303800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umaxv", 0x2e30a800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uminv", 0x2e31a800, 0xbf3ffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmaxnmv",0x2e30c800, 0xbfbffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_FP, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxnmv",0x0e30c800, 0xbffffc00, asimdall, OP2 (Fd, Vn), QL_XLANES_FP_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmaxv", 0x2e30f800, 0xbfbffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_FP, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxv", 0x0e30f800, 0xbffffc00, asimdall, OP2 (Fd, Vn), QL_XLANES_FP_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fminnmv",0x2eb0c800, 0xbfbffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_FP, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminnmv",0x0eb0c800, 0xbffffc00, asimdall, OP2 (Fd, Vn), QL_XLANES_FP_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fminv", 0x2eb0f800, 0xbfbffc00, asimdall, 0, OP2 (Fd, Vn), QL_XLANES_FP, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminv", 0x0eb0f800, 0xbffffc00, asimdall, OP2 (Fd, Vn), QL_XLANES_FP_H, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD three different. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("saddl", 0x0e200000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("saddl2", 0x4e200000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("saddw", 0x0e201000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("saddw2", 0x4e201000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("ssubl", 0x0e202000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("ssubl2", 0x4e202000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("ssubw", 0x0e203000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("ssubw2", 0x4e203000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("addhn", 0x0e204000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("addhn2", 0x4e204000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sabal", 0x0e205000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sabal2", 0x4e205000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("subhn", 0x0e206000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("subhn2", 0x4e206000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sabdl", 0x0e207000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sabdl2", 0x4e207000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlal", 0x0e208000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlal2", 0x4e208000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlal", 0x0e209000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlal2",0x4e209000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlsl", 0x0e20a000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlsl2", 0x4e20a000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlsl", 0x0e20b000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlsl2",0x4e20b000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smull", 0x0e20c000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smull2", 0x4e20c000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmull", 0x0e20d000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmull2",0x4e20d000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("pmull", 0x0e20e000, 0xffe0fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGB, 0),
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
AES_INSN ("pmull", 0x0ee0e000, 0xffe0fc00, asimddiff, OP3 (Vd, Vn, Vm), QL_V3LONGD, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("pmull2", 0x4e20e000, 0xffe0fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGB2, 0),
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
AES_INSN ("pmull2", 0x4ee0e000, 0xffe0fc00, asimddiff, OP3 (Vd, Vn, Vm), QL_V3LONGD2, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("uaddl", 0x2e200000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaddl2", 0x6e200000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaddw", 0x2e201000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaddw2", 0x6e201000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("usubl", 0x2e202000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("usubl2", 0x6e202000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("usubw", 0x2e203000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("usubw2", 0x6e203000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3WIDEBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("raddhn", 0x2e204000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("raddhn2", 0x6e204000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uabal", 0x2e205000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uabal2", 0x6e205000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("rsubhn", 0x2e206000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("rsubhn2", 0x6e206000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uabdl", 0x2e207000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uabdl2", 0x6e207000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlal", 0x2e208000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlal2", 0x6e208000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlsl", 0x2e20a000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlsl2", 0x6e20a000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umull", 0x2e20c000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umull2", 0x6e20c000, 0xff20fc00, asimddiff, 0, OP3 (Vd, Vn, Vm), QL_V3LONGBHS2, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD vector x indexed element. */
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SIMD_INSN ("smlal", 0x0f002000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlal2", 0x4f002000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
Add remainder of Em16 restrictions for AArch64 gas.
This adds the missing Em16 constraints the rest of the instructions requiring them
and also adds a testcase to test all the instructions so these are checked from
now on.
The Em16 operand constrains the valid registers to the lower 16 registers when used
with a half precision qualifier.
The list has been cross checked (by hand) through the Arm ARM version Ca.
opcodes/
PR binutils/23192
* aarch64-tbl.h (sqdmlal, sqdmlal2, smlsl, smlsl2, sqdmlsl, sqdmlsl2,
mul, smull, smull2, sqdmull, sqdmull2, sqdmulh, sqrdmulh, mla, umlal,
umlal2, mls, umlsl, umlsl2, umull, umull2, sqdmlal, sqdmlsl, sqdmull,
sqdmulh, sqrdmulh): Use Em16.
gas/
PR binutils/23192
* testsuite/gas/aarch64/illegal-by-element.s: New.
* testsuite/gas/aarch64/illegal-by-element.d: New.
* testsuite/gas/aarch64/illegal-by-element.l: New.
2018-07-12 11:28:46 +02:00
|
|
|
|
SIMD_INSN ("sqdmlal", 0x0f003000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlal2",0x4f003000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlsl", 0x0f006000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smlsl2", 0x4f006000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlsl", 0x0f007000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmlsl2",0x4f007000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mul", 0x0f008000, 0xbf00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smull", 0x0f00a000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smull2", 0x4f00a000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmull", 0x0f00b000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmull2",0x4f00b000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmulh", 0x0f00c000, 0xbf00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqrdmulh",0x0f00d000, 0xbf00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmla", 0x0f801000, 0xbf80f400, asimdelem, 0, OP3 (Vd, Vn, Em), QL_ELEMENT_FP, F_SIZEQ),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmla", 0x0f001000, 0xbfc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT_FP_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmls", 0x0f805000, 0xbf80f400, asimdelem, 0, OP3 (Vd, Vn, Em), QL_ELEMENT_FP, F_SIZEQ),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmls", 0x0f005000, 0xbfc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT_FP_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmul", 0x0f809000, 0xbf80f400, asimdelem, 0, OP3 (Vd, Vn, Em), QL_ELEMENT_FP, F_SIZEQ),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmul", 0x0f009000, 0xbfc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT_FP_H, F_SIZEQ),
|
Add remainder of Em16 restrictions for AArch64 gas.
This adds the missing Em16 constraints the rest of the instructions requiring them
and also adds a testcase to test all the instructions so these are checked from
now on.
The Em16 operand constrains the valid registers to the lower 16 registers when used
with a half precision qualifier.
The list has been cross checked (by hand) through the Arm ARM version Ca.
opcodes/
PR binutils/23192
* aarch64-tbl.h (sqdmlal, sqdmlal2, smlsl, smlsl2, sqdmlsl, sqdmlsl2,
mul, smull, smull2, sqdmull, sqdmull2, sqdmulh, sqrdmulh, mla, umlal,
umlal2, mls, umlsl, umlsl2, umull, umull2, sqdmlal, sqdmlsl, sqdmull,
sqdmulh, sqrdmulh): Use Em16.
gas/
PR binutils/23192
* testsuite/gas/aarch64/illegal-by-element.s: New.
* testsuite/gas/aarch64/illegal-by-element.d: New.
* testsuite/gas/aarch64/illegal-by-element.l: New.
2018-07-12 11:28:46 +02:00
|
|
|
|
SIMD_INSN ("mla", 0x2f000000, 0xbf00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlal", 0x2f002000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlal2", 0x6f002000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mls", 0x2f004000, 0xbf00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlsl", 0x2f006000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umlsl2", 0x6f006000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umull", 0x2f00a000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umull2", 0x6f00a000, 0xff00f400, asimdelem, 0, OP3 (Vd, Vn, Em16), QL_ELEMENT_L2, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmulx", 0x2f809000, 0xbf80f400, asimdelem, 0, OP3 (Vd, Vn, Em), QL_ELEMENT_FP, F_SIZEQ),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmulx", 0x2f009000, 0xbfc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT_FP_H, F_SIZEQ),
|
|
|
|
|
RDMA_INSN ("sqrdmlah",0x2f00d000, 0xbf00f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
|
|
|
|
RDMA_INSN ("sqrdmlsh",0x2f00f000, 0xbf00f400, asimdelem, OP3 (Vd, Vn, Em16), QL_ELEMENT, F_SIZEQ),
|
2017-02-24 19:27:26 +01:00
|
|
|
|
CNUM_INSN ("fcmla", 0x2f001000, 0xbf009400, asimdelem, OP_FCMLA_ELEM, OP4 (Vd, Vn, Em, IMM_ROT2), QL_ELEMENT_ROT, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD EXT. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ext", 0x2e000000, 0xbfe08400, asimdext, 0, OP4 (Vd, Vn, Vm, IDX), QL_VEXT, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD modified immediate. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("movi", 0x0f000400, 0xbff89c00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("orr", 0x0f001400, 0xbff89c00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("movi", 0x0f008400, 0xbff8dc00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0H, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("orr", 0x0f009400, 0xbff8dc00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0H, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("movi", 0x0f00c400, 0xbff8ec00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S1W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("movi", 0x0f00e400, 0xbff8fc00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_B, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmov", 0x0f00f400, 0xbff8fc00, asimdimm, 0, OP2 (Vd, SIMD_FPIMM), QL_SIMD_IMM_S, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmov", 0x0f00fc00, 0xbff8fc00, asimdimm, OP2 (Vd, SIMD_FPIMM), QL_SIMD_IMM_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("mvni", 0x2f000400, 0xbff89c00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("bic", 0x2f001400, 0xbff89c00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mvni", 0x2f008400, 0xbff8dc00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0H, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("bic", 0x2f009400, 0xbff8dc00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S0H, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mvni", 0x2f00c400, 0xbff8ec00, asimdimm, 0, OP2 (Vd, SIMD_IMM_SFT), QL_SIMD_IMM_S1W, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("movi", 0x2f00e400, 0xfff8fc00, asimdimm, 0, OP2 (Sd, SIMD_IMM), QL_SIMD_IMM_D, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("movi", 0x6f00e400, 0xfff8fc00, asimdimm, 0, OP2 (Vd, SIMD_IMM), QL_SIMD_IMM_V2D, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmov", 0x6f00f400, 0xfff8fc00, asimdimm, 0, OP2 (Vd, SIMD_FPIMM), QL_SIMD_IMM_V2D, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD copy. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("dup", 0x0e000400, 0xbfe0fc00, asimdins, 0, OP2 (Vd, En), QL_DUP_VX, F_T),
|
|
|
|
|
SIMD_INSN ("dup", 0x0e000c00, 0xbfe0fc00, asimdins, 0, OP2 (Vd, Rn), QL_DUP_VR, F_T),
|
|
|
|
|
SIMD_INSN ("smov",0x0e002c00, 0xbfe0fc00, asimdins, 0, OP2 (Rd, En), QL_SMOV, F_GPRSIZE_IN_Q),
|
|
|
|
|
SIMD_INSN ("umov",0x0e003c00, 0xbfe0fc00, asimdins, 0, OP2 (Rd, En), QL_UMOV, F_HAS_ALIAS | F_GPRSIZE_IN_Q),
|
|
|
|
|
SIMD_INSN ("mov", 0x0e003c00, 0xbfe0fc00, asimdins, 0, OP2 (Rd, En), QL_MOV, F_ALIAS | F_GPRSIZE_IN_Q),
|
|
|
|
|
SIMD_INSN ("ins", 0x4e001c00, 0xffe0fc00, asimdins, 0, OP2 (Ed, Rn), QL_INS_XR, F_HAS_ALIAS),
|
|
|
|
|
SIMD_INSN ("mov", 0x4e001c00, 0xffe0fc00, asimdins, 0, OP2 (Ed, Rn), QL_INS_XR, F_ALIAS),
|
|
|
|
|
SIMD_INSN ("ins", 0x6e000400, 0xffe08400, asimdins, 0, OP2 (Ed, En), QL_S_2SAME, F_HAS_ALIAS),
|
|
|
|
|
SIMD_INSN ("mov", 0x6e000400, 0xffe08400, asimdins, 0, OP2 (Ed, En), QL_S_2SAME, F_ALIAS),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD two-reg misc. */
|
[PATCH, BINUTILS, AARCH64, 2/9] Add Data procoessing instructions for ARMv8.5-A
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order)
This patch adds the data processing instructions that are new to
ARMv8.5-A.
1) There are 2 instructions: xaflag, axflag, that are added to
manipulate the states of the flag and are used to convert between the
Arm representation and the fcmp representation.
2) The other instructions are rounding instructions which have 8
versions based on whether the floating-point number is a
Single-Precision or Double-Precision number, whether the target
integer is a 32-bit or 64-bit integer and whether the rounding mode is
the ambient rounding mode or to zero. Each of these instruction is
available in both Scalar and Vector forms.
Since both 1) and 2) have separate identification mechanism and it is
permissible that a ARMv8.4 compliant implementation may include any
arbitrary subset of the ARMv8.5 features unless otherwise specified,
new feature bits are added.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_FLAGMANIP): New.
(AARCH64_FEATURE_FRINTTS): New.
(AARCH64_ARCH_V8_5): Add both by default.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_feature_flagmanip): New.
(aarch64_feature_frintts): New.
(FLAGMANIP, FRINTTS): New.
(aarch64_opcode_table): Add entries for xaflag, axflag
and frint[32,64][x,z] instructions.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-dp.s: New.
* testsuite/gas/aarch64/armv8_5-a-dp.d: New.
2018-09-26 11:45:35 +02:00
|
|
|
|
FRINTTS_INSN ("frint32z", 0x0e21e800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
|
|
|
|
FRINTTS_INSN ("frint32x", 0x2e21e800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
|
|
|
|
FRINTTS_INSN ("frint64z", 0x0e21f800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
|
|
|
|
FRINTTS_INSN ("frint64x", 0x2e21f800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("rev64", 0x0e200800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("rev16", 0x0e201800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("saddlp",0x0e202800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2PAIRWISELONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("suqadd",0x0e203800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cls", 0x0e204800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cnt", 0x0e205800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sadalp",0x0e206800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2PAIRWISELONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqabs", 0x0e207800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmgt", 0x0e208800, 0xbf3ffc00, asimdmisc, 0, OP3 (Vd, Vn, IMM0), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmeq", 0x0e209800, 0xbf3ffc00, asimdmisc, 0, OP3 (Vd, Vn, IMM0), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmlt", 0x0e20a800, 0xbf3ffc00, asimdmisc, 0, OP3 (Vd, Vn, IMM0), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("abs", 0x0e20b800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("xtn", 0x0e212800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("xtn2", 0x4e212800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqxtn", 0xe214800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqxtn2",0x4e214800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS2, F_SIZEQ),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("fcvtn", 0x0e216800, 0xffbffc00, asimdmisc, OP_FCVTN, OP2 (Vd, Vn), QL_V2NARRHS, F_MISC),
|
|
|
|
|
SIMD_INSN ("fcvtn2", 0x4e216800, 0xffbffc00, asimdmisc, OP_FCVTN2, OP2 (Vd, Vn), QL_V2NARRHS2, F_MISC),
|
|
|
|
|
SIMD_INSN ("fcvtl", 0x0e217800, 0xffbffc00, asimdmisc, OP_FCVTL, OP2 (Vd, Vn), QL_V2LONGHS, F_MISC),
|
|
|
|
|
SIMD_INSN ("fcvtl2", 0x4e217800, 0xffbffc00, asimdmisc, OP_FCVTL2, OP2 (Vd, Vn), QL_V2LONGHS2, F_MISC),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frintn", 0x0e218800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frintn", 0x0e798800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frintm", 0x0e219800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frintm", 0x0e799800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtns", 0x0e21a800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtns", 0x0e79a800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtms", 0x0e21b800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtms", 0x0e79b800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtas", 0x0e21c800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtas", 0x0e79c800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("scvtf", 0x0e21d800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("scvtf", 0x0e79d800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmgt", 0x0ea0c800, 0xbfbffc00, asimdmisc, 0, OP3 (Vd, Vn, FPIMM0), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmgt", 0x0ef8c800, 0xbffffc00, asimdmisc, OP3 (Vd, Vn, FPIMM0), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmeq", 0x0ea0d800, 0xbfbffc00, asimdmisc, 0, OP3 (Vd, Vn, FPIMM0), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmeq", 0x0ef8d800, 0xbffffc00, asimdmisc, OP3 (Vd, Vn, FPIMM0), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmlt", 0x0ea0e800, 0xbfbffc00, asimdmisc, 0, OP3 (Vd, Vn, FPIMM0), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmlt", 0x0ef8e800, 0xbffffc00, asimdmisc, OP3 (Vd, Vn, FPIMM0), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fabs", 0x0ea0f800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fabs", 0x0ef8f800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frintp", 0x0ea18800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frintp", 0x0ef98800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frintz", 0x0ea19800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frintz", 0x0ef99800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtps", 0x0ea1a800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtps", 0x0ef9a800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzs", 0x0ea1b800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzs", 0x0ef9b800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("urecpe", 0x0ea1c800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("frecpe", 0x0ea1d800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frecpe", 0x0ef9d800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("rev32", 0x2e200800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEBH, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaddlp", 0x2e202800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2PAIRWISELONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("usqadd", 0x2e203800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("clz", 0x2e204800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uadalp", 0x2e206800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2PAIRWISELONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqneg", 0x2e207800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmge", 0x2e208800, 0xbf3ffc00, asimdmisc, 0, OP3 (Vd, Vn, IMM0), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmle", 0x2e209800, 0xbf3ffc00, asimdmisc, 0, OP3 (Vd, Vn, IMM0), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("neg", 0x2e20b800, 0xbf3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqxtun", 0x2e212800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqxtun2",0x6e212800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("shll", 0x2e213800, 0xff3ffc00, asimdmisc, 0, OP3 (Vd, Vn, SHLL_IMM), QL_V2LONGBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("shll2", 0x6e213800, 0xff3ffc00, asimdmisc, 0, OP3 (Vd, Vn, SHLL_IMM), QL_V2LONGBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqxtn", 0x2e214800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqxtn2", 0x6e214800, 0xff3ffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRBHS2, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fcvtxn", 0x2e616800, 0xfffffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRS, 0),
|
|
|
|
|
SIMD_INSN ("fcvtxn2",0x6e616800, 0xfffffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2NARRS2, 0),
|
|
|
|
|
SIMD_INSN ("frinta", 0x2e218800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frinta", 0x2e798800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frintx", 0x2e219800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frintx", 0x2e799800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtnu", 0x2e21a800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtnu", 0x2e79a800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtmu", 0x2e21b800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtmu", 0x2e79b800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtau", 0x2e21c800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtau", 0x2e79c800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ucvtf", 0x2e21d800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("ucvtf", 0x2e79d800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("not", 0x2e205800, 0xbffffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEB, F_SIZEQ | F_HAS_ALIAS),
|
|
|
|
|
SIMD_INSN ("mvn", 0x2e205800, 0xbffffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEB, F_SIZEQ | F_ALIAS),
|
|
|
|
|
SIMD_INSN ("rbit", 0x2e605800, 0xbffffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fcmge", 0x2ea0c800, 0xbfbffc00, asimdmisc, 0, OP3 (Vd, Vn, FPIMM0), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmge", 0x2ef8c800, 0xbffffc00, asimdmisc, OP3 (Vd, Vn, FPIMM0), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmle", 0x2ea0d800, 0xbfbffc00, asimdmisc, 0, OP3 (Vd, Vn, FPIMM0), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmle", 0x2ef8d800, 0xbffffc00, asimdmisc, OP3 (Vd, Vn, FPIMM0), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fneg", 0x2ea0f800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fneg", 0x2ef8f800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frinti", 0x2ea19800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frinti", 0x2ef99800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtpu", 0x2ea1a800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtpu", 0x2ef9a800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzu", 0x2ea1b800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzu", 0x2ef9b800, 0xbffffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ursqrte",0x2ea1c800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMES, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("frsqrte",0x2ea1d800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frsqrte",0x2ef9d800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fsqrt", 0x2ea1f800, 0xbfbffc00, asimdmisc, 0, OP2 (Vd, Vn), QL_V2SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fsqrt", 0x2ef9f800, 0xbfbffc00, asimdmisc, OP2 (Vd, Vn), QL_V2SAMEH, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD ZIP/UZP/TRN. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("uzp1", 0xe001800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("trn1", 0xe002800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("zip1", 0xe003800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uzp2", 0xe005800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("trn2", 0xe006800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("zip2", 0xe007800, 0xbf20fc00, asimdperm, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD three same. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("shadd", 0xe200400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqadd", 0xe200c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("srhadd", 0xe201400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("shsub", 0xe202400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqsub", 0xe202c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmgt", 0xe203400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmge", 0xe203c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sshl", 0xe204400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqshl", 0xe204c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("srshl", 0xe205400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqrshl", 0xe205c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smax", 0xe206400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smin", 0xe206c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sabd", 0xe207400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("saba", 0xe207c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("add", 0xe208400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmtst", 0xe208c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mla", 0xe209400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mul", 0xe209c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("smaxp", 0xe20a400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sminp", 0xe20ac00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqdmulh", 0xe20b400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("addp", 0xe20bc00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmaxnm", 0xe20c400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxnm", 0xe400400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmla", 0xe20cc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmla", 0xe400c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fadd", 0xe20d400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fadd", 0xe401400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmulx", 0xe20dc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmulx", 0xe401c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmeq", 0xe20e400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmeq", 0xe402400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmax", 0xe20f400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmax", 0xe403400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frecps", 0xe20fc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frecps", 0xe403c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("and", 0xe201c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("bic", 0xe601c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fminnm", 0xea0c400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminnm", 0xec00400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmls", 0xea0cc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmls", 0xec00c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fsub", 0xea0d400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fsub", 0xec01400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmin", 0xea0f400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmin", 0xec03400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frsqrts", 0xea0fc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frsqrts", 0xec03c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("orr", 0xea01c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_HAS_ALIAS | F_SIZEQ),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("mov", 0xea01c00, 0xbfe0fc00, asimdsame, OP_MOV_V, OP2 (Vd, Vn), QL_V2SAMEB, F_ALIAS | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("orn", 0xee01c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uhadd", 0x2e200400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqadd", 0x2e200c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("urhadd", 0x2e201400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uhsub", 0x2e202400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqsub", 0x2e202c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmhi", 0x2e203400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmhs", 0x2e203c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("ushl", 0x2e204400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqshl", 0x2e204c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("urshl", 0x2e205400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uqrshl", 0x2e205c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umax", 0x2e206400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umin", 0x2e206c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uabd", 0x2e207400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uaba", 0x2e207c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sub", 0x2e208400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("cmeq", 0x2e208c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAME, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("mls", 0x2e209400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("pmul", 0x2e209c00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("umaxp", 0x2e20a400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("uminp", 0x2e20ac00, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEBHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("sqrdmulh", 0x2e20b400, 0xbf20fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEHS, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmaxnmp", 0x2e20c400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxnmp", 0x2e400400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("faddp", 0x2e20d400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("faddp", 0x2e401400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmul", 0x2e20dc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmul", 0x2e401c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmge", 0x2e20e400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmge", 0x2e402400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("facge", 0x2e20ec00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("facge", 0x2e402c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmaxp", 0x2e20f400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxp", 0x2e403400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fdiv", 0x2e20fc00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fdiv", 0x2e403c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("eor", 0x2e201c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("bsl", 0x2e601c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fminnmp", 0x2ea0c400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminnmp", 0x2ec00400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fabd", 0x2ea0d400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fabd", 0x2ec01400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmgt", 0x2ea0e400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmgt", 0x2ec02400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("facgt", 0x2ea0ec00, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("facgt", 0x2ec02c00, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fminp", 0x2ea0f400, 0xbfa0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMESD, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminp", 0x2ec03400, 0xbfe0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEH, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("bit", 0x2ea01c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("bif", 0x2ee01c00, 0xbfe0fc00, asimdsame, 0, OP3 (Vd, Vn, Vm), QL_V3SAMEB, F_SIZEQ),
|
2015-06-02 13:20:00 +02:00
|
|
|
|
/* AdvSIMD three same extension. */
|
2018-04-25 14:37:30 +02:00
|
|
|
|
RDMA_INSN ("sqrdmlah",0x2e008400, 0xbf20fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEHS, F_SIZEQ),
|
|
|
|
|
RDMA_INSN ("sqrdmlsh",0x2e008c00, 0xbf20fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3SAMEHS, F_SIZEQ),
|
2017-02-24 19:27:26 +01:00
|
|
|
|
CNUM_INSN ("fcmla", 0x2e00c400, 0xbf20e400, asimdsame, 0, OP4 (Vd, Vn, Vm, IMM_ROT1), QL_V3SAMEHSD_ROT, F_SIZEQ),
|
|
|
|
|
CNUM_INSN ("fcadd", 0x2e00e400, 0xbf20ec00, asimdsame, 0, OP4 (Vd, Vn, Vm, IMM_ROT3), QL_V3SAMEHSD_ROT, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD shift by immediate. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("sshr", 0xf000400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("ssra", 0xf001400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("srshr", 0xf002400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("srsra", 0xf003400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("shl", 0xf005400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sqshl", 0xf007400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("shrn", 0xf008400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("shrn2", 0x4f008400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("rshrn", 0xf008c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("rshrn2", 0x4f008c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("sqshrn", 0xf009400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqshrn2", 0x4f009400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrn", 0xf009c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrn2", 0x4f009c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("sshll", 0xf00a400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFTL, F_HAS_ALIAS),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("sxtl", 0xf00a400, 0xff87fc00, asimdshf, OP_SXTL, OP2 (Vd, Vn), QL_V2LONGBHS, F_ALIAS | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("sshll2", 0x4f00a400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFTL2, F_HAS_ALIAS),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("sxtl2", 0x4f00a400, 0xff87fc00, asimdshf, OP_SXTL2, OP2 (Vd, Vn), QL_V2LONGBHS2, F_ALIAS | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("scvtf", 0xf00e400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("scvtf", 0xf10e400, 0xbf80fc00, asimdshf, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzs", 0xf00fc00, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzs", 0xf10fc00, 0xbf80fc00, asimdshf, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ushr", 0x2f000400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("usra", 0x2f001400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("urshr", 0x2f002400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("ursra", 0x2f003400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sri", 0x2f004400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sli", 0x2f005400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sqshlu", 0x2f006400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("uqshl", 0x2f007400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sqshrun", 0x2f008400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqshrun2", 0x6f008400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrun", 0x2f008c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrun2", 0x6f008c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("uqshrn", 0x2f009400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("uqshrn2", 0x6f009400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("uqrshrn", 0x2f009c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("uqrshrn2", 0x6f009c00, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFTN2, 0),
|
|
|
|
|
SIMD_INSN ("ushll", 0x2f00a400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFTL, F_HAS_ALIAS),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("uxtl", 0x2f00a400, 0xff87fc00, asimdshf, OP_UXTL, OP2 (Vd, Vn), QL_V2LONGBHS, F_ALIAS | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ushll2", 0x6f00a400, 0xff80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSL), QL_VSHIFTL2, F_HAS_ALIAS),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("uxtl2", 0x6f00a400, 0xff87fc00, asimdshf, OP_UXTL2, OP2 (Vd, Vn), QL_V2LONGBHS2, F_ALIAS | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ucvtf", 0x2f00e400, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("ucvtf", 0x2f10e400, 0xbf80fc00, asimdshf, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzu", 0x2f00fc00, 0xbf80fc00, asimdshf, 0, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzu", 0x2f10fc00, 0xbf80fc00, asimdshf, OP3 (Vd, Vn, IMM_VLSR), QL_VSHIFT_H, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD TBL/TBX. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("tbl", 0xe000000, 0xbfe09c00, asimdtbl, 0, OP3 (Vd, LVn, Vm), QL_TABLE, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("tbx", 0xe001000, 0xbfe09c00, asimdtbl, 0, OP3 (Vd, LVn, Vm), QL_TABLE, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar three different. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("sqdmlal", 0x5e209000, 0xff20fc00, asisddiff, 0, OP3 (Sd, Sn, Sm), QL_SISDL_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmlsl", 0x5e20b000, 0xff20fc00, asisddiff, 0, OP3 (Sd, Sn, Sm), QL_SISDL_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmull", 0x5e20d000, 0xff20fc00, asisddiff, 0, OP3 (Sd, Sn, Sm), QL_SISDL_HS, F_SSIZE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar x indexed element. */
|
Add remainder of Em16 restrictions for AArch64 gas.
This adds the missing Em16 constraints the rest of the instructions requiring them
and also adds a testcase to test all the instructions so these are checked from
now on.
The Em16 operand constrains the valid registers to the lower 16 registers when used
with a half precision qualifier.
The list has been cross checked (by hand) through the Arm ARM version Ca.
opcodes/
PR binutils/23192
* aarch64-tbl.h (sqdmlal, sqdmlal2, smlsl, smlsl2, sqdmlsl, sqdmlsl2,
mul, smull, smull2, sqdmull, sqdmull2, sqdmulh, sqrdmulh, mla, umlal,
umlal2, mls, umlsl, umlsl2, umull, umull2, sqdmlal, sqdmlsl, sqdmull,
sqdmulh, sqrdmulh): Use Em16.
gas/
PR binutils/23192
* testsuite/gas/aarch64/illegal-by-element.s: New.
* testsuite/gas/aarch64/illegal-by-element.d: New.
* testsuite/gas/aarch64/illegal-by-element.l: New.
2018-07-12 11:28:46 +02:00
|
|
|
|
SIMD_INSN ("sqdmlal", 0x5f003000, 0xff00f400, asisdelem, 0, OP3 (Sd, Sn, Em16), QL_SISDL_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmlsl", 0x5f007000, 0xff00f400, asisdelem, 0, OP3 (Sd, Sn, Em16), QL_SISDL_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmull", 0x5f00b000, 0xff00f400, asisdelem, 0, OP3 (Sd, Sn, Em16), QL_SISDL_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmulh", 0x5f00c000, 0xff00f400, asisdelem, 0, OP3 (Sd, Sn, Em16), QL_SISD_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqrdmulh", 0x5f00d000, 0xff00f400, asisdelem, 0, OP3 (Sd, Sn, Em16), QL_SISD_HS, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmla", 0x5f801000, 0xff80f400, asisdelem, 0, OP3 (Sd, Sn, Em), QL_FP3, F_SSIZE),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmla", 0x5f001000, 0xffc0f400, asisdelem, OP3 (Sd, Sn, Em16), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmls", 0x5f805000, 0xff80f400, asisdelem, 0, OP3 (Sd, Sn, Em), QL_FP3, F_SSIZE),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmls", 0x5f005000, 0xffc0f400, asisdelem, OP3 (Sd, Sn, Em16), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmul", 0x5f809000, 0xff80f400, asisdelem, 0, OP3 (Sd, Sn, Em), QL_FP3, F_SSIZE),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmul", 0x5f009000, 0xffc0f400, asisdelem, OP3 (Sd, Sn, Em16), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmulx", 0x7f809000, 0xff80f400, asisdelem, 0, OP3 (Sd, Sn, Em), QL_FP3, F_SSIZE),
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
SF16_INSN ("fmulx", 0x7f009000, 0xffc0f400, asisdelem, OP3 (Sd, Sn, Em16), QL_FP3_H, F_SSIZE),
|
|
|
|
|
RDMA_INSN ("sqrdmlah", 0x7f00d000, 0xff00f400, asisdelem, OP3 (Sd, Sn, Em16), QL_SISD_HS, F_SSIZE),
|
|
|
|
|
RDMA_INSN ("sqrdmlsh", 0x7f00f000, 0xff00f400, asisdelem, OP3 (Sd, Sn, Em16), QL_SISD_HS, F_SSIZE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD load/store multiple structures. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("st4", 0xc000000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(4)),
|
|
|
|
|
SIMD_INSN ("st1", 0xc000000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("st2", 0xc000000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("st3", 0xc000000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(3)),
|
|
|
|
|
SIMD_INSN ("ld4", 0xc400000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(4)),
|
|
|
|
|
SIMD_INSN ("ld1", 0xc400000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld2", 0xc400000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld3", 0xc400000, 0xbfff0000, asisdlse, 0, OP2 (LVt, SIMD_ADDR_SIMPLE), QL_SIMD_LDST, F_SIZEQ | F_OD(3)),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD load/store multiple structures (post-indexed). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("st4", 0xc800000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(4)),
|
|
|
|
|
SIMD_INSN ("st1", 0xc800000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("st2", 0xc800000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("st3", 0xc800000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(3)),
|
|
|
|
|
SIMD_INSN ("ld4", 0xcc00000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(4)),
|
|
|
|
|
SIMD_INSN ("ld1", 0xcc00000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld2", 0xcc00000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld3", 0xcc00000, 0xbfe00000, asisdlsep, 0, OP2 (LVt, SIMD_ADDR_POST), QL_SIMD_LDST, F_SIZEQ | F_OD(3)),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD load/store single structure. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("st1", 0xd000000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(1)),
|
|
|
|
|
SIMD_INSN ("st3", 0xd002000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(3)),
|
|
|
|
|
SIMD_INSN ("st2", 0xd200000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(2)),
|
|
|
|
|
SIMD_INSN ("st4", 0xd202000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(4)),
|
|
|
|
|
SIMD_INSN ("ld1", 0xd400000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld3", 0xd402000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(3)),
|
2017-04-21 13:18:06 +02:00
|
|
|
|
SIMD_INSN ("ld1r", 0xd40c000, 0xbffff000, asisdlso, 0, OP2 (LVt_AL, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld3r", 0xd40e000, 0xbffff000, asisdlso, 0, OP2 (LVt_AL, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(3)),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ld2", 0xd600000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld4", 0xd602000, 0xbfff2000, asisdlso, 0, OP2 (LEt, SIMD_ADDR_SIMPLE), QL_SIMD_LDSTONE, F_OD(4)),
|
2017-04-21 13:18:06 +02:00
|
|
|
|
SIMD_INSN ("ld2r", 0xd60c000, 0xbffff000, asisdlso, 0, OP2 (LVt_AL, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld4r", 0xd60e000, 0xbffff000, asisdlso, 0, OP2 (LVt_AL, SIMD_ADDR_SIMPLE), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(4)),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD load/store single structure (post-indexed). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("st1", 0xd800000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(1)),
|
|
|
|
|
SIMD_INSN ("st3", 0xd802000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(3)),
|
|
|
|
|
SIMD_INSN ("st2", 0xda00000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(2)),
|
|
|
|
|
SIMD_INSN ("st4", 0xda02000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(4)),
|
|
|
|
|
SIMD_INSN ("ld1", 0xdc00000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld3", 0xdc02000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(3)),
|
2017-04-21 13:18:06 +02:00
|
|
|
|
SIMD_INSN ("ld1r", 0xdc0c000, 0xbfe0f000, asisdlsop, 0, OP2 (LVt_AL, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(1)),
|
|
|
|
|
SIMD_INSN ("ld3r", 0xdc0e000, 0xbfe0f000, asisdlsop, 0, OP2 (LVt_AL, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(3)),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ld2", 0xde00000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld4", 0xde02000, 0xbfe02000, asisdlsop, 0, OP2 (LEt, SIMD_ADDR_POST), QL_SIMD_LDSTONE, F_OD(4)),
|
2017-04-21 13:18:06 +02:00
|
|
|
|
SIMD_INSN ("ld2r", 0xde0c000, 0xbfe0f000, asisdlsop, 0, OP2 (LVt_AL, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(2)),
|
|
|
|
|
SIMD_INSN ("ld4r", 0xde0e000, 0xbfe0f000, asisdlsop, 0, OP2 (LVt_AL, SIMD_ADDR_POST), QL_SIMD_LDST_ANY, F_SIZEQ | F_OD(4)),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar two-reg misc. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("suqadd", 0x5e203800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqabs", 0x5e207800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmgt", 0x5e208800, 0xff3ffc00, asisdmisc, 0, OP3 (Sd, Sn, IMM0), QL_SISD_CMP_0, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmeq", 0x5e209800, 0xff3ffc00, asisdmisc, 0, OP3 (Sd, Sn, IMM0), QL_SISD_CMP_0, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmlt", 0x5e20a800, 0xff3ffc00, asisdmisc, 0, OP3 (Sd, Sn, IMM0), QL_SISD_CMP_0, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("abs", 0x5e20b800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_2SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqxtn", 0x5e214800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_SISD_NARROW, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("fcvtns", 0x5e21a800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtns", 0x5e79a800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtms", 0x5e21b800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtms", 0x5e79b800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtas", 0x5e21c800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtas", 0x5e79c800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("scvtf", 0x5e21d800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("scvtf", 0x5e79d800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmgt", 0x5ea0c800, 0xffbffc00, asisdmisc, 0, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_0, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmgt", 0x5ef8c800, 0xfffffc00, asisdmisc, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmeq", 0x5ea0d800, 0xffbffc00, asisdmisc, 0, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_0, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmeq", 0x5ef8d800, 0xfffffc00, asisdmisc, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmlt", 0x5ea0e800, 0xffbffc00, asisdmisc, 0, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_0, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmlt", 0x5ef8e800, 0xfffffc00, asisdmisc, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtps", 0x5ea1a800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtps", 0x5ef9a800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzs", 0x5ea1b800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzs", 0x5ef9b800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frecpe", 0x5ea1d800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frecpe", 0x5ef9d800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frecpx", 0x5ea1f800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frecpx", 0x5ef9f800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("usqadd", 0x7e203800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqneg", 0x7e207800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmge", 0x7e208800, 0xff3ffc00, asisdmisc, 0, OP3 (Sd, Sn, IMM0), QL_SISD_CMP_0, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmle", 0x7e209800, 0xff3ffc00, asisdmisc, 0, OP3 (Sd, Sn, IMM0), QL_SISD_CMP_0, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("neg", 0x7e20b800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_2SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqxtun", 0x7e212800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_SISD_NARROW, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("uqxtn", 0x7e214800, 0xff3ffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_SISD_NARROW, F_SSIZE),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
SIMD_INSN ("fcvtxn", 0x7e216800, 0xffbffc00, asisdmisc, OP_FCVTXN_S, OP2 (Sd, Sn), QL_SISD_NARROW_S, F_MISC),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtnu", 0x7e21a800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtnu", 0x7e79a800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtmu", 0x7e21b800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtmu", 0x7e79b800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtau", 0x7e21c800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtau", 0x7e79c800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ucvtf", 0x7e21d800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("ucvtf", 0x7e79d800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmge", 0x7ea0c800, 0xffbffc00, asisdmisc, 0, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_0, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmge", 0x7ef8c800, 0xfffffc00, asisdmisc, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmle", 0x7ea0d800, 0xffbffc00, asisdmisc, 0, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_0, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmle", 0x7ef8d800, 0xfffffc00, asisdmisc, OP3 (Sd, Sn, FPIMM0), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtpu", 0x7ea1a800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtpu", 0x7ef9a800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_SISD_FCMP_H_0, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzu", 0x7ea1b800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzu", 0x7ef9b800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frsqrte", 0x7ea1d800, 0xffbffc00, asisdmisc, 0, OP2 (Sd, Sn), QL_S_2SAMESD, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frsqrte", 0x7ef9d800, 0xfffffc00, asisdmisc, OP2 (Sd, Sn), QL_S_2SAMEH, F_SSIZE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar copy. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("dup", 0x5e000400, 0xffe0fc00, asisdone, 0, OP2 (Sd, En), QL_S_2SAME, F_HAS_ALIAS),
|
|
|
|
|
SIMD_INSN ("mov", 0x5e000400, 0xffe0fc00, asisdone, 0, OP2 (Sd, En), QL_S_2SAME, F_ALIAS),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar pairwise. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("addp", 0x5e31b800, 0xff3ffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR_D, F_SIZEQ),
|
|
|
|
|
SIMD_INSN ("fmaxnmp", 0x7e30c800, 0xffbffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxnmp", 0x5e30c800, 0xfffffc00, asisdpair, OP2 (Sd, Vn), QL_SISD_PAIR_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("faddp", 0x7e30d800, 0xffbffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("faddp", 0x5e30d800, 0xfffffc00, asisdpair, OP2 (Sd, Vn), QL_SISD_PAIR_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fmaxp", 0x7e30f800, 0xffbffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmaxp", 0x5e30f800, 0xfffffc00, asisdpair, OP2 (Sd, Vn), QL_SISD_PAIR_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fminnmp", 0x7eb0c800, 0xffbffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminnmp", 0x5eb0c800, 0xfffffc00, asisdpair, OP2 (Sd, Vn), QL_SISD_PAIR_H, F_SIZEQ),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fminp", 0x7eb0f800, 0xffbffc00, asisdpair, 0, OP2 (Sd, Vn), QL_SISD_PAIR, F_SIZEQ),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fminp", 0x5eb0f800, 0xfffffc00, asisdpair, OP2 (Sd, Vn), QL_SISD_PAIR_H, F_SIZEQ),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar three same. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("sqadd", 0x5e200c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqsub", 0x5e202c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqshl", 0x5e204c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqrshl", 0x5e205c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqdmulh", 0x5e20b400, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_SISD_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("fmulx", 0x5e20dc00, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fmulx", 0x5e401c00, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmeq", 0x5e20e400, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmeq", 0x5e402400, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frecps", 0x5e20fc00, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frecps", 0x5e403c00, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("frsqrts", 0x5ea0fc00, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("frsqrts", 0x5ec03c00, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("cmgt", 0x5ee03400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmge", 0x5ee03c00, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sshl", 0x5ee04400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("srshl", 0x5ee05400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("add", 0x5ee08400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmtst", 0x5ee08c00, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("uqadd", 0x7e200c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("uqsub", 0x7e202c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("uqshl", 0x7e204c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("uqrshl", 0x7e205c00, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAME, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sqrdmulh", 0x7e20b400, 0xff20fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_SISD_HS, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("fcmge", 0x7e20e400, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmge", 0x7e402400, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("facge", 0x7e20ec00, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("facge", 0x7e402c00, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fabd", 0x7ea0d400, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fabd", 0x7ec01400, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcmgt", 0x7ea0e400, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcmgt", 0x7ec02400, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("facgt", 0x7ea0ec00, 0xffa0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_FP3, F_SSIZE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("facgt", 0x7ec02c00, 0xffe0fc00, asisdsame, OP3 (Sd, Sn, Sm), QL_FP3_H, F_SSIZE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("cmhi", 0x7ee03400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmhs", 0x7ee03c00, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("ushl", 0x7ee04400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("urshl", 0x7ee05400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("sub", 0x7ee08400, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
|
|
|
|
SIMD_INSN ("cmeq", 0x7ee08c00, 0xffe0fc00, asisdsame, 0, OP3 (Sd, Sn, Sm), QL_S_3SAMED, F_SSIZE),
|
2015-06-02 13:20:00 +02:00
|
|
|
|
/* AdvSIMDs scalar three same extension. */
|
2016-04-28 10:11:03 +02:00
|
|
|
|
RDMA_INSN ("sqrdmlah", 0x7e008400, 0xff20fc00, asimdsame, OP3 (Sd, Sn, Sm), QL_SISD_HS, F_SSIZE),
|
|
|
|
|
RDMA_INSN ("sqrdmlsh", 0x7e008c00, 0xff20fc00, asimdsame, OP3 (Sd, Sn, Sm), QL_SISD_HS, F_SSIZE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* AdvSIMD scalar shift by immediate. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("sshr", 0x5f000400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("ssra", 0x5f001400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("srshr", 0x5f002400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("srsra", 0x5f003400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("shl", 0x5f005400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSL), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("sqshl", 0x5f007400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSL), QL_SSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sqshrn", 0x5f009400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrn", 0x5f009c00, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("scvtf", 0x5f00e400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("scvtf", 0x5f10e400, 0xff80fc00, asisdshf, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzs", 0x5f00fc00, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzs", 0x5f10fc00, 0xff80fc00, asisdshf, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("ushr", 0x7f000400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("usra", 0x7f001400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("urshr", 0x7f002400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("ursra", 0x7f003400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("sri", 0x7f004400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("sli", 0x7f005400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSL), QL_SSHIFT_D, 0),
|
|
|
|
|
SIMD_INSN ("sqshlu", 0x7f006400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSL), QL_SSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("uqshl", 0x7f007400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSL), QL_SSHIFT, 0),
|
|
|
|
|
SIMD_INSN ("sqshrun", 0x7f008400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("sqrshrun", 0x7f008c00, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("uqshrn", 0x7f009400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("uqrshrn", 0x7f009c00, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFTN, 0),
|
|
|
|
|
SIMD_INSN ("ucvtf", 0x7f00e400, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("ucvtf", 0x7f10e400, 0xff80fc00, asisdshf, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_H, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
SIMD_INSN ("fcvtzu", 0x7f00fc00, 0xff80fc00, asisdshf, 0, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_SD, 0),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
SF16_INSN ("fcvtzu", 0x7f10fc00, 0xff80fc00, asisdshf, OP3 (Sd, Sn, IMM_VLSR), QL_SSHIFT_H, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Bitfield. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("sbfm", 0x13000000, 0x7f800000, bitfield, 0, OP4 (Rd, Rn, IMMR, IMMS), QL_BF, F_HAS_ALIAS | F_SF | F_N),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("sbfiz", 0x13000000, 0x7f800000, bitfield, OP_SBFIZ, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
|
|
|
|
CORE_INSN ("sbfx", 0x13000000, 0x7f800000, bitfield, OP_SBFX, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("sxtb", 0x13001c00, 0x7fbffc00, bitfield, 0, OP2 (Rd, Rn), QL_EXT, F_ALIAS | F_P3 | F_SF | F_N),
|
|
|
|
|
CORE_INSN ("sxth", 0x13003c00, 0x7fbffc00, bitfield, 0, OP2 (Rd, Rn), QL_EXT, F_ALIAS | F_P3 | F_SF | F_N),
|
|
|
|
|
CORE_INSN ("sxtw", 0x93407c00, 0xfffffc00, bitfield, 0, OP2 (Rd, Rn), QL_EXT_W, F_ALIAS | F_P3),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("asr", 0x13000000, 0x7f800000, bitfield, OP_ASR_IMM, OP3 (Rd, Rn, IMM), QL_SHIFT, F_ALIAS | F_P2 | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("bfm", 0x33000000, 0x7f800000, bitfield, 0, OP4 (Rd, Rn, IMMR, IMMS), QL_BF, F_HAS_ALIAS | F_SF | F_N),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("bfi", 0x33000000, 0x7f800000, bitfield, OP_BFI, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
2016-08-23 10:03:19 +02:00
|
|
|
|
V8_2_INSN ("bfc", 0x330003e0, 0x7f8003e0, bitfield, OP_BFC, OP3 (Rd, IMM, WIDTH), QL_BF1, F_ALIAS | F_P2 | F_CONV),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("bfxil", 0x33000000, 0x7f800000, bitfield, OP_BFXIL, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("ubfm", 0x53000000, 0x7f800000, bitfield, 0, OP4 (Rd, Rn, IMMR, IMMS), QL_BF, F_HAS_ALIAS | F_SF | F_N),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("ubfiz", 0x53000000, 0x7f800000, bitfield, OP_UBFIZ, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
|
|
|
|
CORE_INSN ("ubfx", 0x53000000, 0x7f800000, bitfield, OP_UBFX, OP4 (Rd, Rn, IMM, WIDTH), QL_BF2, F_ALIAS | F_P1 | F_CONV),
|
|
|
|
|
CORE_INSN ("uxtb", 0x53001c00, 0xfffffc00, bitfield, OP_UXTB, OP2 (Rd, Rn), QL_I2SAMEW, F_ALIAS | F_P3),
|
|
|
|
|
CORE_INSN ("uxth", 0x53003c00, 0xfffffc00, bitfield, OP_UXTH, OP2 (Rd, Rn), QL_I2SAMEW, F_ALIAS | F_P3),
|
|
|
|
|
CORE_INSN ("lsl", 0x53000000, 0x7f800000, bitfield, OP_LSL_IMM, OP3 (Rd, Rn, IMM), QL_SHIFT, F_ALIAS | F_P2 | F_CONV),
|
|
|
|
|
CORE_INSN ("lsr", 0x53000000, 0x7f800000, bitfield, OP_LSR_IMM, OP3 (Rd, Rn, IMM), QL_SHIFT, F_ALIAS | F_P2 | F_CONV),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Unconditional branch (immediate). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("b", 0x14000000, 0xfc000000, branch_imm, OP_B, OP1 (ADDR_PCREL26), QL_PCREL_26, 0),
|
|
|
|
|
CORE_INSN ("bl", 0x94000000, 0xfc000000, branch_imm, OP_BL, OP1 (ADDR_PCREL26), QL_PCREL_26, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Unconditional branch (register). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("br", 0xd61f0000, 0xfffffc1f, branch_reg, 0, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
CORE_INSN ("blr", 0xd63f0000, 0xfffffc1f, branch_reg, 0, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
CORE_INSN ("ret", 0xd65f0000, 0xfffffc1f, branch_reg, 0, OP1 (Rn), QL_I1X, F_OPD0_OPT | F_DEFAULT (30)),
|
|
|
|
|
CORE_INSN ("eret", 0xd69f03e0, 0xffffffff, branch_reg, 0, OP0 (), {}, 0),
|
|
|
|
|
CORE_INSN ("drps", 0xd6bf03e0, 0xffffffff, branch_reg, 0, OP0 (), {}, 0),
|
[AArch64] Add ARMv8.3 combined pointer authentication branch instructions
Add support for ARMv8.3 pointer authentication instructions
that are encoded as unconditional branch instructions.
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (arch64_opcode_table): Add braa, brab, blraa, blrab, braaz,
brabz, blraaz, blrabz, retaa, retab, eretaa, eretab.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
gas/
2016-11-08 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/pac.s: Add ARMv8.3 branch instruction tests.
* testsuite/gas/aarch64/pac.d: Likewise.
2016-11-11 11:43:15 +01:00
|
|
|
|
V8_3_INSN ("braa", 0xd71f0800, 0xfffffc00, branch_reg, OP2 (Rn, Rd_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("brab", 0xd71f0c00, 0xfffffc00, branch_reg, OP2 (Rn, Rd_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("blraa", 0xd73f0800, 0xfffffc00, branch_reg, OP2 (Rn, Rd_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("blrab", 0xd73f0c00, 0xfffffc00, branch_reg, OP2 (Rn, Rd_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("braaz", 0xd61f081f, 0xfffffc1f, branch_reg, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("brabz", 0xd61f0c1f, 0xfffffc1f, branch_reg, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("blraaz", 0xd63f081f, 0xfffffc1f, branch_reg, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("blrabz", 0xd63f0c1f, 0xfffffc1f, branch_reg, OP1 (Rn), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("retaa", 0xd65f0bff, 0xffffffff, branch_reg, OP0 (), {}, 0),
|
|
|
|
|
V8_3_INSN ("retab", 0xd65f0fff, 0xffffffff, branch_reg, OP0 (), {}, 0),
|
|
|
|
|
V8_3_INSN ("eretaa", 0xd69f0bff, 0xffffffff, branch_reg, OP0 (), {}, 0),
|
|
|
|
|
V8_3_INSN ("eretab", 0xd69f0fff, 0xffffffff, branch_reg, OP0 (), {}, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Compare & branch (immediate). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("cbz", 0x34000000, 0x7f000000, compbranch, 0, OP2 (Rt, ADDR_PCREL19), QL_R_PCREL, F_SF),
|
|
|
|
|
CORE_INSN ("cbnz", 0x35000000, 0x7f000000, compbranch, 0, OP2 (Rt, ADDR_PCREL19), QL_R_PCREL, F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Conditional branch (immediate). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("b.c", 0x54000000, 0xff000010, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_COND),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Conditional compare (immediate). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("ccmn", 0x3a400800, 0x7fe00c10, condcmp_imm, 0, OP4 (Rn, CCMP_IMM, NZCV, COND), QL_CCMP_IMM, F_SF),
|
|
|
|
|
CORE_INSN ("ccmp", 0x7a400800, 0x7fe00c10, condcmp_imm, 0, OP4 (Rn, CCMP_IMM, NZCV, COND), QL_CCMP_IMM, F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Conditional compare (register). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("ccmn", 0x3a400000, 0x7fe00c10, condcmp_reg, 0, OP4 (Rn, Rm, NZCV, COND), QL_CCMP, F_SF),
|
|
|
|
|
CORE_INSN ("ccmp", 0x7a400000, 0x7fe00c10, condcmp_reg, 0, OP4 (Rn, Rm, NZCV, COND), QL_CCMP, F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Conditional select. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("csel", 0x1a800000, 0x7fe00c00, condsel, 0, OP4 (Rd, Rn, Rm, COND), QL_CSEL, F_SF),
|
|
|
|
|
CORE_INSN ("csinc", 0x1a800400, 0x7fe00c00, condsel, 0, OP4 (Rd, Rn, Rm, COND), QL_CSEL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("cinc", 0x1a800400, 0x7fe00c00, condsel, OP_CINC, OP3 (Rd, Rn, COND1), QL_CSEL, F_ALIAS | F_SF | F_CONV),
|
|
|
|
|
CORE_INSN ("cset", 0x1a9f07e0, 0x7fff0fe0, condsel, OP_CSET, OP2 (Rd, COND1), QL_DST_R, F_ALIAS | F_P1 | F_SF | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("csinv", 0x5a800000, 0x7fe00c00, condsel, 0, OP4 (Rd, Rn, Rm, COND), QL_CSEL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("cinv", 0x5a800000, 0x7fe00c00, condsel, OP_CINV, OP3 (Rd, Rn, COND1), QL_CSEL, F_ALIAS | F_SF | F_CONV),
|
|
|
|
|
CORE_INSN ("csetm", 0x5a9f03e0, 0x7fff0fe0, condsel, OP_CSETM, OP2 (Rd, COND1), QL_DST_R, F_ALIAS | F_P1 | F_SF | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("csneg", 0x5a800400, 0x7fe00c00, condsel, 0, OP4 (Rd, Rn, Rm, COND), QL_CSEL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("cneg", 0x5a800400, 0x7fe00c00, condsel, OP_CNEG, OP3 (Rd, Rn, COND1), QL_CSEL, F_ALIAS | F_SF | F_CONV),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Crypto AES. */
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
AES_INSN ("aese", 0x4e284800, 0xfffffc00, cryptoaes, OP2 (Vd, Vn), QL_V2SAME16B, 0),
|
|
|
|
|
AES_INSN ("aesd", 0x4e285800, 0xfffffc00, cryptoaes, OP2 (Vd, Vn), QL_V2SAME16B, 0),
|
|
|
|
|
AES_INSN ("aesmc", 0x4e286800, 0xfffffc00, cryptoaes, OP2 (Vd, Vn), QL_V2SAME16B, 0),
|
|
|
|
|
AES_INSN ("aesimc", 0x4e287800, 0xfffffc00, cryptoaes, OP2 (Vd, Vn), QL_V2SAME16B, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Crypto two-reg SHA. */
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
SHA2_INSN ("sha1h", 0x5e280800, 0xfffffc00, cryptosha2, OP2 (Fd, Fn), QL_2SAMES, 0),
|
|
|
|
|
SHA2_INSN ("sha1su1", 0x5e281800, 0xfffffc00, cryptosha2, OP2 (Vd, Vn), QL_V2SAME4S, 0),
|
|
|
|
|
SHA2_INSN ("sha256su0",0x5e282800, 0xfffffc00, cryptosha2, OP2 (Vd, Vn), QL_V2SAME4S, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Crypto three-reg SHA. */
|
Split the AArch64 Crypto instructions for AES and SHA1+2 into their own options (+aes and +sha2).
The new options are:
+aes: Enables the AES instructions of Armv8-a,
enabled by default with +crypto.
+sha2: Enables the SHA1 and SHA2 instructions of Armv8-a,
enabled by default with +crypto.
These options have been turned on by default when +crypto
is used, as such no breakage is expected.
The reason for the split is because with the introduction of Armv8.4-a
the implementation of AES has explicitly been made independent of the
implementation of the other crypto extensions. Backporting the split does
not break any of the previous requirements and so is safe to do.
gas * config/tc-aarch64.c
(aarch64_features): Include AES and SHA2 in CRYPTO.
Add SHA2 and AES.
include * opcode/aarch64.h:
(AARCH64_FEATURE_SHA2, AARCH64_FEATURE_AES): New.
opcodes * aarch64-tbl.h (aarch64_feature_crypto): Add AES and SHA2.
(aarch64_feature_sha2, aarch64_feature_aes): New.
(SHA2, AES): New.
(AES_INSN, SHA2_INSN): New.
(pmull, pmull2, aese, aesd, aesmc, aesimc): Change to AES_INS.
(sha1h, sha1su1, sha256su0, sha1c, sha1p,
sha1m, sha1su0, sha256h, sha256h2, sha256su1):
Change to SHA2_INS.
2017-11-08 15:30:53 +01:00
|
|
|
|
SHA2_INSN ("sha1c", 0x5e000000, 0xffe0fc00, cryptosha3, OP3 (Fd, Fn, Vm), QL_SHAUPT, 0),
|
|
|
|
|
SHA2_INSN ("sha1p", 0x5e001000, 0xffe0fc00, cryptosha3, OP3 (Fd, Fn, Vm), QL_SHAUPT, 0),
|
|
|
|
|
SHA2_INSN ("sha1m", 0x5e002000, 0xffe0fc00, cryptosha3, OP3 (Fd, Fn, Vm), QL_SHAUPT, 0),
|
|
|
|
|
SHA2_INSN ("sha1su0", 0x5e003000, 0xffe0fc00, cryptosha3, OP3 (Vd, Vn, Vm), QL_V3SAME4S, 0),
|
|
|
|
|
SHA2_INSN ("sha256h", 0x5e004000, 0xffe0fc00, cryptosha3, OP3 (Fd, Fn, Vm), QL_SHA256UPT, 0),
|
|
|
|
|
SHA2_INSN ("sha256h2", 0x5e005000, 0xffe0fc00, cryptosha3, OP3 (Fd, Fn, Vm), QL_SHA256UPT, 0),
|
|
|
|
|
SHA2_INSN ("sha256su1",0x5e006000, 0xffe0fc00, cryptosha3, OP3 (Vd, Vn, Vm), QL_V3SAME4S, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Data-processing (1 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("rbit", 0x5ac00000, 0x7ffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAME, F_SF),
|
|
|
|
|
CORE_INSN ("rev16", 0x5ac00400, 0x7ffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAME, F_SF),
|
|
|
|
|
CORE_INSN ("rev", 0x5ac00800, 0xfffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAMEW, 0),
|
|
|
|
|
CORE_INSN ("rev", 0xdac00c00, 0xfffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAMEX, F_SF | F_HAS_ALIAS | F_P1),
|
2016-08-23 10:03:19 +02:00
|
|
|
|
V8_2_INSN ("rev64", 0xdac00c00, 0xfffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAMEX, F_SF | F_ALIAS),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("clz", 0x5ac01000, 0x7ffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAME, F_SF),
|
|
|
|
|
CORE_INSN ("cls", 0x5ac01400, 0x7ffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAME, F_SF),
|
|
|
|
|
CORE_INSN ("rev32", 0xdac00800, 0xfffffc00, dp_1src, 0, OP2 (Rd, Rn), QL_I2SAMEX, 0),
|
[AArch64] Add ARMv8.3 single source PAC instructions
Add support for ARMv8.3 pointer authentication instructions
that are encoded as single source data processing instructions.
opcodes/
2016-11-08 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (arch64_opcode_table): Add pacia, pacib, pacda, pacdb, autia,
autib, autda, autdb, paciza, pacizb, pacdza, pacdzb, autiza, autizb, autdza,
autdzb, xpaci, xpacd.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
gas/testsuite/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/pac.s: New.
* testsuite/gas/aarch64/pac.d: New.
2016-11-11 11:36:32 +01:00
|
|
|
|
V8_3_INSN ("pacia", 0xdac10000, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("pacib", 0xdac10400, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("pacda", 0xdac10800, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("pacdb", 0xdac10c00, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("autia", 0xdac11000, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("autib", 0xdac11400, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("autda", 0xdac11800, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("autdb", 0xdac11c00, 0xfffffc00, dp_1src, OP2 (Rd, Rn_SP), QL_I2SAMEX, 0),
|
|
|
|
|
V8_3_INSN ("paciza", 0xdac123e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("pacizb", 0xdac127e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("pacdza", 0xdac12be0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("pacdzb", 0xdac12fe0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("autiza", 0xdac133e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("autizb", 0xdac137e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("autdza", 0xdac13be0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("autdzb", 0xdac13fe0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("xpaci", 0xdac143e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
|
|
|
|
V8_3_INSN ("xpacd", 0xdac147e0, 0xffffffe0, dp_1src, OP1 (Rd), QL_I1X, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Data-processing (2 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("udiv", 0x1ac00800, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("sdiv", 0x1ac00c00, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("lslv", 0x1ac02000, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("lsl", 0x1ac02000, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_ALIAS),
|
|
|
|
|
CORE_INSN ("lsrv", 0x1ac02400, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("lsr", 0x1ac02400, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_ALIAS),
|
|
|
|
|
CORE_INSN ("asrv", 0x1ac02800, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("asr", 0x1ac02800, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_ALIAS),
|
|
|
|
|
CORE_INSN ("rorv", 0x1ac02c00, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("ror", 0x1ac02c00, 0x7fe0fc00, dp_2src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_SF | F_ALIAS),
|
[BINUTILS, AARCH64, 3/8] Add Pointer Arithmetic instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Pointer Arithmetic instructions from
MTE. These are the following instructions added in this patch:
- SUBP <Xd>, <Xn|SP>, <Xm|SP>
- SUBPS <Xd>, <Xn|SP>, <Xm|SP>
- CMPP <Xn|SP>, <Xm|SP>
where CMPP is an alias to SUBPS XZR, <Xn|SP>, <Xm|SP>
where
<Xd> : Is the 64-bit destination GPR.
<Xn|SP> : Is the 64-bit first source GPR or Stack pointer.
<Xm|SP> : Is the 64-bit second source GPR or Stack pointer.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_opcode_table): Add subp, subps and cmpp.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for subp,
subps and cmpp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 13:58:26 +01:00
|
|
|
|
MEMTAG_INSN ("subp", 0x9ac00000, 0xffe0fc00, dp_2src, OP3 (Rd, Rn_SP, Rm_SP), QL_I3SAMEX, 0),
|
|
|
|
|
MEMTAG_INSN ("subps", 0xbac00000, 0xffe0fc00, dp_2src, OP3 (Rd, Rn_SP, Rm_SP), QL_I3SAMEX, F_HAS_ALIAS),
|
|
|
|
|
MEMTAG_INSN ("cmpp", 0xbac0001f, 0xffe0fc1f, dp_2src, OP2 (Rn_SP, Rm_SP), QL_I2SAMEX, F_ALIAS),
|
[BINUTILS, AARCH64, 2/8] Add Tag generation instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag generation instructions from
MTE. These are the following instructions added in this patch:
- IRG <Xd|SP>, <Xn|SP>{, Xm}
- ADDG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- SUBG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- GMI <Xd>, <Xn|SP>, <Xm>
where
<Xd|SP> : Is the 64-bit destination GPR or Stack pointer.
<Xn|SP> : Is the 64-bit source GPR or Stack pointer.
<uimm6> : Is the unsigned immediate, a multiple of 16
in the range 0 to 1008.
<uimm4> : Is the unsigned immediate, in the range 0 to 15.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10 as new enums.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (aarch64_field_kind): New FLD_imm4_3.
(OPD_F_SHIFT_BY_4, operand_need_shift_by_four): New.
* aarch64-opc.c (fields): Add entry for imm4_3.
(operand_general_constraint_met_p): Add cases for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_ADDG): New.
(aarch64_opcode_table): Add addg, subg, irg and gmi.
(AARCH64_OPERANDS): Define UIMM4_ADDG and UIMM10.
* aarch64-asm.c (aarch64_ins_imm): Add case for
operand_need_shift_by_four.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: New.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
* testsuite/gas/aarch64/illegal-memtag.d: Likewise.
2018-11-12 13:52:55 +01:00
|
|
|
|
MEMTAG_INSN ("irg", 0x9ac01000, 0xffe0fc00, dp_2src, OP3 (Rd_SP, Rn_SP, Rm), QL_I3SAMEX, F_OPD2_OPT | F_DEFAULT (0x1f)),
|
|
|
|
|
MEMTAG_INSN ("gmi", 0x9ac01400, 0xffe0fc00, dp_2src, OP3 (Rd, Rn_SP, Rm), QL_I3SAMEX, 0),
|
2016-11-11 11:39:46 +01:00
|
|
|
|
V8_3_INSN ("pacga", 0x9ac03000, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm_SP), QL_I3SAMEX, 0),
|
2013-02-28 20:18:40 +01:00
|
|
|
|
/* CRC instructions. */
|
2016-04-28 10:11:03 +02:00
|
|
|
|
_CRC_INSN ("crc32b", 0x1ac04000, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32h", 0x1ac04400, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32w", 0x1ac04800, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32x", 0x9ac04c00, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3WWX, 0),
|
|
|
|
|
_CRC_INSN ("crc32cb",0x1ac05000, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32ch",0x1ac05400, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32cw",0x1ac05800, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3SAMEW, 0),
|
|
|
|
|
_CRC_INSN ("crc32cx",0x9ac05c00, 0xffe0fc00, dp_2src, OP3 (Rd, Rn, Rm), QL_I3WWX, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Data-processing (3 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("madd", 0x1b000000, 0x7fe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("mul", 0x1b007c00, 0x7fe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("msub", 0x1b008000, 0x7fe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("mneg", 0x1b00fc00, 0x7fe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMER, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("smaddl",0x9b200000, 0xffe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMEL, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("smull", 0x9b207c00, 0xffe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEL, F_ALIAS),
|
|
|
|
|
CORE_INSN ("smsubl",0x9b208000, 0xffe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMEL, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("smnegl",0x9b20fc00, 0xffe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEL, F_ALIAS),
|
|
|
|
|
CORE_INSN ("smulh", 0x9b407c00, 0xffe08000, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEX, 0),
|
|
|
|
|
CORE_INSN ("umaddl",0x9ba00000, 0xffe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMEL, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("umull", 0x9ba07c00, 0xffe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEL, F_ALIAS),
|
|
|
|
|
CORE_INSN ("umsubl",0x9ba08000, 0xffe08000, dp_3src, 0, OP4 (Rd, Rn, Rm, Ra), QL_I4SAMEL, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("umnegl",0x9ba0fc00, 0xffe0fc00, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEL, F_ALIAS),
|
|
|
|
|
CORE_INSN ("umulh", 0x9bc07c00, 0xffe08000, dp_3src, 0, OP3 (Rd, Rn, Rm), QL_I3SAMEX, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Excep'n generation. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("svc", 0xd4000001, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, 0),
|
|
|
|
|
CORE_INSN ("hvc", 0xd4000002, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, 0),
|
|
|
|
|
CORE_INSN ("smc", 0xd4000003, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, 0),
|
|
|
|
|
CORE_INSN ("brk", 0xd4200000, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, 0),
|
|
|
|
|
CORE_INSN ("hlt", 0xd4400000, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, 0),
|
|
|
|
|
CORE_INSN ("dcps1", 0xd4a00001, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, F_OPD0_OPT | F_DEFAULT (0)),
|
|
|
|
|
CORE_INSN ("dcps2", 0xd4a00002, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, F_OPD0_OPT | F_DEFAULT (0)),
|
|
|
|
|
CORE_INSN ("dcps3", 0xd4a00003, 0xffe0001f, exception, 0, OP1 (EXCEPTION), {}, F_OPD0_OPT | F_DEFAULT (0)),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Extract. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("extr", 0x13800000, 0x7fa00000, extract, 0, OP4 (Rd, Rn, Rm, IMMS), QL_EXTR, F_HAS_ALIAS | F_SF | F_N),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("ror", 0x13800000, 0x7fa00000, extract, OP_ROR_IMM, OP3 (Rd, Rm, IMMS), QL_SHIFT, F_ALIAS | F_CONV),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point<->fixed-point conversions. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("scvtf", 0x1e020000, 0x7f3f0000, float2fix, 0, OP3 (Fd, Rn, FBITS), QL_FIX2FP, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("scvtf", 0x1ec20000, 0x7f3f0000, float2fix, OP3 (Fd, Rn, FBITS), QL_FIX2FP_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("ucvtf", 0x1e030000, 0x7f3f0000, float2fix, 0, OP3 (Fd, Rn, FBITS), QL_FIX2FP, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("ucvtf", 0x1ec30000, 0x7f3f0000, float2fix, OP3 (Fd, Rn, FBITS), QL_FIX2FP_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtzs",0x1e180000, 0x7f3f0000, float2fix, 0, OP3 (Rd, Fn, FBITS), QL_FP2FIX, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtzs",0x1ed80000, 0x7f3f0000, float2fix, OP3 (Rd, Fn, FBITS), QL_FP2FIX_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtzu",0x1e190000, 0x7f3f0000, float2fix, 0, OP3 (Rd, Fn, FBITS), QL_FP2FIX, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtzu",0x1ed90000, 0x7f3f0000, float2fix, OP3 (Rd, Fn, FBITS), QL_FP2FIX_H, F_FPTYPE | F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point<->integer conversions. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtns",0x1e200000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtns",0x1ee00000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtnu",0x1e210000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtnu",0x1ee10000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("scvtf", 0x1e220000, 0x7f3ffc00, float2int, 0, OP2 (Fd, Rn), QL_INT2FP, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("scvtf", 0x1ee20000, 0x7f3ffc00, float2int, OP2 (Fd, Rn), QL_INT2FP_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("ucvtf", 0x1e230000, 0x7f3ffc00, float2int, 0, OP2 (Fd, Rn), QL_INT2FP, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("ucvtf", 0x1ee30000, 0x7f3ffc00, float2int, OP2 (Fd, Rn), QL_INT2FP_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtas",0x1e240000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtas",0x1ee40000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtau",0x1e250000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtau",0x1ee50000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2018-06-08 15:02:52 +02:00
|
|
|
|
__FP_INSN ("fmov", 0x1e260000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT_FMOV, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmov", 0x1ee60000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2018-06-08 15:02:52 +02:00
|
|
|
|
__FP_INSN ("fmov", 0x1e270000, 0x7f3ffc00, float2int, 0, OP2 (Fd, Rn), QL_INT2FP_FMOV, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmov", 0x1ee70000, 0x7f3ffc00, float2int, OP2 (Fd, Rn), QL_INT2FP_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtps",0x1e280000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtps",0x1ee80000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtpu",0x1e290000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtpu",0x1ee90000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtms",0x1e300000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtms",0x1ef00000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtmu",0x1e310000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtmu",0x1ef10000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtzs",0x1e380000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtzs",0x1ef80000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcvtzu",0x1e390000, 0x7f3ffc00, float2int, 0, OP2 (Rd, Fn), QL_FP2INT, F_FPTYPE | F_SF),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcvtzu",0x1ef90000, 0x7f3ffc00, float2int, OP2 (Rd, Fn), QL_FP2INT_H, F_FPTYPE | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmov", 0x9eae0000, 0xfffffc00, float2int, 0, OP2 (Rd, VnD1), QL_XVD1, 0),
|
|
|
|
|
__FP_INSN ("fmov", 0x9eaf0000, 0xfffffc00, float2int, 0, OP2 (VdD1, Rn), QL_VD1X, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{"fjcvtzs", 0x1e7e0000, 0xfffffc00, float2int, 0, FP_V8_3, OP2 (Rd, Fn), QL_FP2INT_W_D, 0, 0, 0, NULL },
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point conditional compare. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fccmp", 0x1e200400, 0xff200c10, floatccmp, 0, OP4 (Fn, Fm, NZCV, COND), QL_FCCMP, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fccmp", 0x1ee00400, 0xff200c10, floatccmp, OP4 (Fn, Fm, NZCV, COND), QL_FCCMP_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fccmpe",0x1e200410, 0xff200c10, floatccmp, 0, OP4 (Fn, Fm, NZCV, COND), QL_FCCMP, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fccmpe",0x1ee00410, 0xff200c10, floatccmp, OP4 (Fn, Fm, NZCV, COND), QL_FCCMP_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point compare. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcmp", 0x1e202000, 0xff20fc1f, floatcmp, 0, OP2 (Fn, Fm), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcmp", 0x1ee02000, 0xff20fc1f, floatcmp, OP2 (Fn, Fm), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcmpe", 0x1e202010, 0xff20fc1f, floatcmp, 0, OP2 (Fn, Fm), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcmpe", 0x1ee02010, 0xff20fc1f, floatcmp, OP2 (Fn, Fm), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcmp", 0x1e202008, 0xff20fc1f, floatcmp, 0, OP2 (Fn, FPIMM0), QL_DST_SD,F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcmp", 0x1ee02008, 0xff20fc1f, floatcmp, OP2 (Fn, FPIMM0), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcmpe", 0x1e202018, 0xff20fc1f, floatcmp, 0, OP2 (Fn, FPIMM0), QL_DST_SD,F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcmpe", 0x1ee02018, 0xff20fc1f, floatcmp, OP2 (Fn, FPIMM0), QL_FP2_H, F_FPTYPE),
|
[PATCH, BINUTILS, AARCH64, 2/9] Add Data procoessing instructions for ARMv8.5-A
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order)
This patch adds the data processing instructions that are new to
ARMv8.5-A.
1) There are 2 instructions: xaflag, axflag, that are added to
manipulate the states of the flag and are used to convert between the
Arm representation and the fcmp representation.
2) The other instructions are rounding instructions which have 8
versions based on whether the floating-point number is a
Single-Precision or Double-Precision number, whether the target
integer is a 32-bit or 64-bit integer and whether the rounding mode is
the ambient rounding mode or to zero. Each of these instruction is
available in both Scalar and Vector forms.
Since both 1) and 2) have separate identification mechanism and it is
permissible that a ARMv8.4 compliant implementation may include any
arbitrary subset of the ARMv8.5 features unless otherwise specified,
new feature bits are added.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_FLAGMANIP): New.
(AARCH64_FEATURE_FRINTTS): New.
(AARCH64_ARCH_V8_5): Add both by default.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-tbl.h (aarch64_feature_flagmanip): New.
(aarch64_feature_frintts): New.
(FLAGMANIP, FRINTTS): New.
(aarch64_opcode_table): Add entries for xaflag, axflag
and frint[32,64][x,z] instructions.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* testsuite/gas/aarch64/armv8_5-a-dp.s: New.
* testsuite/gas/aarch64/armv8_5-a-dp.d: New.
2018-09-26 11:45:35 +02:00
|
|
|
|
/* Data processing instructions ARMv8.5-A. */
|
|
|
|
|
FLAGMANIP_INSN ("xaflag", 0xd500403f, 0xffffffff, 0, OP0 (), {}, 0),
|
|
|
|
|
FLAGMANIP_INSN ("axflag", 0xd500405f, 0xffffffff, 0, OP0 (), {}, 0),
|
|
|
|
|
FRINTTS_INSN ("frint32z", 0x1e284000, 0xffbffc00, floatdp1, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
|
|
|
|
FRINTTS_INSN ("frint32x", 0x1e28c000, 0xffbffc00, floatdp1, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
|
|
|
|
FRINTTS_INSN ("frint64z", 0x1e294000, 0xffbffc00, floatdp1, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
|
|
|
|
FRINTTS_INSN ("frint64x", 0x1e29c000, 0xffbffc00, floatdp1, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point data-processing (1 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmov", 0x1e204000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmov", 0x1ee04000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fabs", 0x1e20c000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fabs", 0x1ee0c000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fneg", 0x1e214000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fneg", 0x1ee14000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fsqrt", 0x1e21c000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fsqrt", 0x1ee1c000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
__FP_INSN ("fcvt", 0x1e224000, 0xff3e7c00, floatdp1, OP_FCVT, OP2 (Fd, Fn), QL_FCVT, F_FPTYPE | F_MISC),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frintn",0x1e244000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frintn",0x1ee44000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frintp",0x1e24c000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frintp",0x1ee4c000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frintm",0x1e254000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frintm",0x1ee54000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frintz",0x1e25c000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frintz",0x1ee5c000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frinta",0x1e264000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frinta",0x1ee64000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frintx",0x1e274000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frintx",0x1ee74000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("frinti",0x1e27c000, 0xff3ffc00, floatdp1, 0, OP2 (Fd, Fn), QL_FP2, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("frinti",0x1ee7c000, 0xff3ffc00, floatdp1, OP2 (Fd, Fn), QL_FP2_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point data-processing (2 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmul", 0x1e200800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmul", 0x1ee00800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fdiv", 0x1e201800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fdiv", 0x1ee01800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fadd", 0x1e202800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fadd", 0x1ee02800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fsub", 0x1e203800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fsub", 0x1ee03800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmax", 0x1e204800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmax", 0x1ee04800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmin", 0x1e205800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmin", 0x1ee05800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmaxnm",0x1e206800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmaxnm",0x1ee06800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fminnm",0x1e207800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fminnm",0x1ee07800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fnmul", 0x1e208800, 0xff20fc00, floatdp2, 0, OP3 (Fd, Fn, Fm), QL_FP3, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fnmul", 0x1ee08800, 0xff20fc00, floatdp2, OP3 (Fd, Fn, Fm), QL_FP3_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point data-processing (3 source). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmadd", 0x1f000000, 0xff208000, floatdp3, 0, OP4 (Fd, Fn, Fm, Fa), QL_FP4, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmadd", 0x1fc00000, 0xff208000, floatdp3, OP4 (Fd, Fn, Fm, Fa), QL_FP4_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmsub", 0x1f008000, 0xff208000, floatdp3, 0, OP4 (Fd, Fn, Fm, Fa), QL_FP4, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmsub", 0x1fc08000, 0xff208000, floatdp3, OP4 (Fd, Fn, Fm, Fa), QL_FP4_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fnmadd",0x1f200000, 0xff208000, floatdp3, 0, OP4 (Fd, Fn, Fm, Fa), QL_FP4, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fnmadd",0x1fe00000, 0xff208000, floatdp3, OP4 (Fd, Fn, Fm, Fa), QL_FP4_H, F_FPTYPE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fnmsub",0x1f208000, 0xff208000, floatdp3, 0, OP4 (Fd, Fn, Fm, Fa), QL_FP4, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fnmsub",0x1fe08000, 0xff208000, floatdp3, OP4 (Fd, Fn, Fm, Fa), QL_FP4_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point immediate. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fmov", 0x1e201000, 0xff201fe0, floatimm, 0, OP2 (Fd, FPIMM), QL_DST_SD, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fmov", 0x1ee01000, 0xff201fe0, floatimm, OP2 (Fd, FPIMM), QL_DST_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Floating-point conditional select. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
__FP_INSN ("fcsel", 0x1e200c00, 0xff200c00, floatsel, 0, OP4 (Fd, Fn, Fm, COND), QL_FP_COND, F_FPTYPE),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
FF16_INSN ("fcsel", 0x1ee00c00, 0xff200c00, floatsel, OP4 (Fd, Fn, Fm, COND), QL_FP_COND_H, F_FPTYPE),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register (immediate indexed). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("strb", 0x38000400, 0xffe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrb", 0x38400400, 0xffe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrsb", 0x38800400, 0xffa00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R8, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0x3c000400, 0x3f600400, ldst_imm9, 0, OP2 (Ft, ADDR_SIMM9), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("ldr", 0x3c400400, 0x3f600400, ldst_imm9, 0, OP2 (Ft, ADDR_SIMM9), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("strh", 0x78000400, 0xffe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrh", 0x78400400, 0xffe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrsh", 0x78800400, 0xffa00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R16, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0xb8000400, 0xbfe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldr", 0xb8400400, 0xbfe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldrsw", 0xb8800400, 0xffe00400, ldst_imm9, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_X32, 0),
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
/* Load/store Allocation Tag instructions. */
|
|
|
|
|
MEMTAG_INSN ("stg", 0xd920081f, 0xffe00c1f, ldst_unscaled, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("stzg", 0xd960081f, 0xffe00c1f, ldst_unscaled, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("st2g", 0xd9a0081f, 0xffe00c1f, ldst_unscaled, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("stz2g",0xd9e0081f, 0xffe00c1f, ldst_unscaled, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("stg", 0xd920041f, 0xffe0041f, ldst_imm9, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("stzg", 0xd960041f, 0xffe0041f, ldst_imm9, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("st2g", 0xd9a0041f, 0xffe0041f, ldst_imm9, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
|
|
|
|
MEMTAG_INSN ("stz2g",0xd9e0041f, 0xffe0041f, ldst_imm9, OP1 (ADDR_SIMM13), QL_LDST_AT, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register (unsigned immediate). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("strb", 0x39000000, 0xffc00000, ldst_pos, OP_STRB_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrb", 0x39400000, 0xffc00000, ldst_pos, OP_LDRB_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrsb", 0x39800000, 0xff800000, ldst_pos, OP_LDRSB_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_R8, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0x3d000000, 0x3f400000, ldst_pos, OP_STRF_POS, OP2 (Ft, ADDR_UIMM12), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("ldr", 0x3d400000, 0x3f400000, ldst_pos, OP_LDRF_POS, OP2 (Ft, ADDR_UIMM12), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("strh", 0x79000000, 0xffc00000, ldst_pos, OP_STRH_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrh", 0x79400000, 0xffc00000, ldst_pos, OP_LDRH_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrsh", 0x79800000, 0xff800000, ldst_pos, OP_LDRSH_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_R16, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0xb9000000, 0xbfc00000, ldst_pos, OP_STR_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldr", 0xb9400000, 0xbfc00000, ldst_pos, OP_LDR_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldrsw", 0xb9800000, 0xffc00000, ldst_pos, OP_LDRSW_POS, OP2 (Rt, ADDR_UIMM12), QL_LDST_X32, 0),
|
|
|
|
|
CORE_INSN ("prfm", 0xf9800000, 0xffc00000, ldst_pos, OP_PRFM_POS, OP2 (PRFOP, ADDR_UIMM12), QL_LDST_PRFM, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register (register offset). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("strb", 0x38200800, 0xffe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrb", 0x38600800, 0xffe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldrsb", 0x38a00800, 0xffa00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_R8, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0x3c200800, 0x3f600c00, ldst_regoff, 0, OP2 (Ft, ADDR_REGOFF), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("ldr", 0x3c600800, 0x3f600c00, ldst_regoff, 0, OP2 (Ft, ADDR_REGOFF), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("strh", 0x78200800, 0xffe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrh", 0x78600800, 0xffe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldrsh", 0x78a00800, 0xffa00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_R16, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("str", 0xb8200800, 0xbfe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldr", 0xb8600800, 0xbfe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldrsw", 0xb8a00800, 0xffe00c00, ldst_regoff, 0, OP2 (Rt, ADDR_REGOFF), QL_LDST_X32, 0),
|
|
|
|
|
CORE_INSN ("prfm", 0xf8a00800, 0xffe00c00, ldst_regoff, 0, OP2 (PRFOP, ADDR_REGOFF), QL_LDST_PRFM, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register (unprivileged). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("sttrb", 0x38000800, 0xffe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldtrb", 0x38400800, 0xffe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldtrsb", 0x38800800, 0xffa00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R8, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("sttrh", 0x78000800, 0xffe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldtrh", 0x78400800, 0xffe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldtrsh", 0x78800800, 0xffa00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R16, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("sttr", 0xb8000800, 0xbfe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldtr", 0xb8400800, 0xbfe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldtrsw", 0xb8800800, 0xffe00c00, ldst_unpriv, 0, OP2 (Rt, ADDR_SIMM9), QL_LDST_X32, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register (unscaled immediate). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("sturb", 0x38000000, 0xffe00c00, ldst_unscaled, OP_STURB, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldurb", 0x38400000, 0xffe00c00, ldst_unscaled, OP_LDURB, OP2 (Rt, ADDR_SIMM9), QL_LDST_W8, 0),
|
|
|
|
|
CORE_INSN ("ldursb", 0x38800000, 0xffa00c00, ldst_unscaled, OP_LDURSB, OP2 (Rt, ADDR_SIMM9), QL_LDST_R8, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("stur", 0x3c000000, 0x3f600c00, ldst_unscaled, OP_STURV, OP2 (Ft, ADDR_SIMM9), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("ldur", 0x3c400000, 0x3f600c00, ldst_unscaled, OP_LDURV, OP2 (Ft, ADDR_SIMM9), QL_LDST_FP, 0),
|
|
|
|
|
CORE_INSN ("sturh", 0x78000000, 0xffe00c00, ldst_unscaled, OP_STURH, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldurh", 0x78400000, 0xffe00c00, ldst_unscaled, OP_LDURH, OP2 (Rt, ADDR_SIMM9), QL_LDST_W16, 0),
|
|
|
|
|
CORE_INSN ("ldursh", 0x78800000, 0xffa00c00, ldst_unscaled, OP_LDURSH, OP2 (Rt, ADDR_SIMM9), QL_LDST_R16, F_LDS_SIZE),
|
|
|
|
|
CORE_INSN ("stur", 0xb8000000, 0xbfe00c00, ldst_unscaled, OP_STUR, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldur", 0xb8400000, 0xbfe00c00, ldst_unscaled, OP_LDUR, OP2 (Rt, ADDR_SIMM9), QL_LDST_R, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldursw", 0xb8800000, 0xffe00c00, ldst_unscaled, OP_LDURSW, OP2 (Rt, ADDR_SIMM9), QL_LDST_X32, 0),
|
|
|
|
|
CORE_INSN ("prfum", 0xf8800000, 0xffe00c00, ldst_unscaled, OP_PRFUM, OP2 (PRFOP, ADDR_SIMM9), QL_LDST_PRFM, 0),
|
2018-11-12 14:14:00 +01:00
|
|
|
|
MEMTAG_INSN ("ldg", 0xd9600000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_SIMM13), QL_LDG, 0),
|
2016-11-18 10:49:06 +01:00
|
|
|
|
/* Load/store register (scaled signed immediate). */
|
|
|
|
|
V8_3_INSN ("ldraa", 0xf8200400, 0xffa00400, ldst_imm10, OP2 (Rt, ADDR_SIMM10), QL_X1NIL, 0),
|
|
|
|
|
V8_3_INSN ("ldrab", 0xf8a00400, 0xffa00400, ldst_imm10, OP2 (Rt, ADDR_SIMM10), QL_X1NIL, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store exclusive. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("stxrb", 0x8007c00, 0xffe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("stlxrb", 0x800fc00, 0xffe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("ldxrb", 0x85f7c00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("ldaxrb", 0x85ffc00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("stlrb", 0x89ffc00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("ldarb", 0x8dffc00, 0xffeffc00, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("stxrh", 0x48007c00, 0xffe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("stlxrh", 0x4800fc00, 0xffe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("ldxrh", 0x485f7c00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("ldaxrh", 0x485ffc00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
CORE_INSN ("stlrh", 0x489ffc00, 0xffe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
2018-07-06 17:18:19 +02:00
|
|
|
|
CORE_INSN ("ldarh", 0x48dffc00, 0xfffffc00, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("stxr", 0x88007c00, 0xbfe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2_LDST_EXC, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("stlxr", 0x8800fc00, 0xbfe08000, ldstexcl, 0, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2_LDST_EXC, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("stxp", 0x88200000, 0xbfe08000, ldstexcl, 0, OP4 (Rs, Rt, Rt2, ADDR_SIMPLE), QL_R3_LDST_EXC, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("stlxp", 0x88208000, 0xbfe08000, ldstexcl, 0, OP4 (Rs, Rt, Rt2, ADDR_SIMPLE), QL_R3_LDST_EXC, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldxr", 0x885f7c00, 0xbfe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldaxr", 0x885ffc00, 0xbfe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldxp", 0x887f0000, 0xbfe08000, ldstexcl, 0, OP3 (Rt, Rt2, ADDR_SIMPLE), QL_R2NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldaxp", 0x887f8000, 0xbfe08000, ldstexcl, 0, OP3 (Rt, Rt2, ADDR_SIMPLE), QL_R2NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("stlr", 0x889ffc00, 0xbfe08000, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldar", 0x88dffc00, 0xbfeffc00, ldstexcl, 0, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
2017-01-04 13:27:10 +01:00
|
|
|
|
RCPC_INSN ("ldaprb", 0x38bfc000, 0xfffffc00, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
RCPC_INSN ("ldaprh", 0x78bfc000, 0xfffffc00, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
RCPC_INSN ("ldapr", 0xb8bfc000, 0xbffffc00, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
2019-01-25 15:15:45 +01:00
|
|
|
|
MEMTAG_INSN ("stzgm", 0xd9200000, 0xfffffc00, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_X1NIL, 0),
|
2015-06-02 12:29:15 +02:00
|
|
|
|
/* Limited Ordering Regions load/store instructions. */
|
2016-04-28 10:11:03 +02:00
|
|
|
|
_LOR_INSN ("ldlar", 0x88df7c00, 0xbfe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
_LOR_INSN ("ldlarb", 0x08df7c00, 0xffe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
_LOR_INSN ("ldlarh", 0x48df7c00, 0xffe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
_LOR_INSN ("stllr", 0x889f7c00, 0xbfe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_R1NIL, F_GPRSIZE_IN_Q),
|
|
|
|
|
_LOR_INSN ("stllrb", 0x089f7c00, 0xffe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
|
|
|
|
_LOR_INSN ("stllrh", 0x489f7c00, 0xbfe08000, ldstexcl, OP2 (Rt, ADDR_SIMPLE), QL_W1_LDST_EXC, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store no-allocate pair (offset). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("stnp", 0x28000000, 0x7fc00000, ldstnapair_offs, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("ldnp", 0x28400000, 0x7fc00000, ldstnapair_offs, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("stnp", 0x2c000000, 0x3fc00000, ldstnapair_offs, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
|
|
|
|
CORE_INSN ("ldnp", 0x2c400000, 0x3fc00000, ldstnapair_offs, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register pair (offset). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("stp", 0x29000000, 0x7ec00000, ldstpair_off, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("ldp", 0x29400000, 0x7ec00000, ldstpair_off, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("stp", 0x2d000000, 0x3fc00000, ldstpair_off, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
|
|
|
|
CORE_INSN ("ldp", 0x2d400000, 0x3fc00000, ldstpair_off, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{"ldpsw", 0x69400000, 0xffc00000, ldstpair_off, 0, CORE, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_X32, 0, 0, 0, VERIFIER (ldpsw)},
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
MEMTAG_INSN ("stgp", 0x69000000, 0xffc00000, ldstpair_off, OP3 (Rt, Rt2, ADDR_SIMM11), QL_STGP, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load/store register pair (indexed). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("stp", 0x28800000, 0x7ec00000, ldstpair_indexed, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("ldp", 0x28c00000, 0x7ec00000, ldstpair_indexed, 0, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_R, F_SF),
|
|
|
|
|
CORE_INSN ("stp", 0x2c800000, 0x3ec00000, ldstpair_indexed, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
|
|
|
|
CORE_INSN ("ldp", 0x2cc00000, 0x3ec00000, ldstpair_indexed, 0, OP3 (Ft, Ft2, ADDR_SIMM7), QL_LDST_PAIR_FP, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{"ldpsw", 0x68c00000, 0xfec00000, ldstpair_indexed, 0, CORE, OP3 (Rt, Rt2, ADDR_SIMM7), QL_LDST_PAIR_X32, 0, 0, 0, VERIFIER (ldpsw)},
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
MEMTAG_INSN ("stgp", 0x68800000, 0xfec00000, ldstpair_indexed, OP3 (Rt, Rt2, ADDR_SIMM11), QL_STGP, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Load register (literal). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("ldr", 0x18000000, 0xbf000000, loadlit, OP_LDR_LIT, OP2 (Rt, ADDR_PCREL19), QL_R_PCREL, F_GPRSIZE_IN_Q),
|
|
|
|
|
CORE_INSN ("ldr", 0x1c000000, 0x3f000000, loadlit, OP_LDRV_LIT, OP2 (Ft, ADDR_PCREL19), QL_FP_PCREL, 0),
|
|
|
|
|
CORE_INSN ("ldrsw", 0x98000000, 0xff000000, loadlit, OP_LDRSW_LIT, OP2 (Rt, ADDR_PCREL19), QL_X_PCREL, 0),
|
|
|
|
|
CORE_INSN ("prfm", 0xd8000000, 0xff000000, loadlit, OP_PRFM_LIT, OP2 (PRFOP, ADDR_PCREL19), QL_PRFM_PCREL, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Logical (immediate). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("and", 0x12000000, 0x7f800000, log_imm, 0, OP3 (Rd_SP, Rn, LIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("bic", 0x12000000, 0x7f800000, log_imm, OP_BIC, OP3 (Rd_SP, Rn, LIMM), QL_R2NIL, F_ALIAS | F_PSEUDO | F_SF),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("orr", 0x32000000, 0x7f800000, log_imm, 0, OP3 (Rd_SP, Rn, LIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("mov", 0x320003e0, 0x7f8003e0, log_imm, OP_MOV_IMM_LOG, OP2 (Rd_SP, IMM_MOV), QL_R1NIL, F_ALIAS | F_P1 | F_SF | F_CONV),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("eor", 0x52000000, 0x7f800000, log_imm, 0, OP3 (Rd_SP, Rn, LIMM), QL_R2NIL, F_SF),
|
|
|
|
|
CORE_INSN ("ands", 0x72000000, 0x7f800000, log_imm, 0, OP3 (Rd, Rn, LIMM), QL_R2NIL, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("tst", 0x7200001f, 0x7f80001f, log_imm, 0, OP2 (Rn, LIMM), QL_R1NIL, F_ALIAS | F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Logical (shifted register). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("and", 0xa000000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("bic", 0xa200000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("orr", 0x2a000000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
2018-12-03 18:31:44 +01:00
|
|
|
|
CORE_INSN ("mov", 0x2a0003e0, 0x7fe0ffe0, log_shift, 0, OP2 (Rd, Rm_SFT), QL_I2SAMER, F_ALIAS | F_SF),
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("uxtw", 0x2a0003e0, 0x7f2003e0, log_shift, OP_UXTW, OP2 (Rd, Rm), QL_I2SAMEW, F_ALIAS | F_PSEUDO),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("orn", 0x2a200000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("mvn", 0x2a2003e0, 0x7f2003e0, log_shift, 0, OP2 (Rd, Rm_SFT), QL_I2SAMER, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("eor", 0x4a000000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("eon", 0x4a200000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
|
|
|
|
CORE_INSN ("ands", 0x6a000000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_HAS_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("tst", 0x6a00001f, 0x7f20001f, log_shift, 0, OP2 (Rn, Rm_SFT), QL_I2SAMER, F_ALIAS | F_SF),
|
|
|
|
|
CORE_INSN ("bics", 0x6a200000, 0x7f200000, log_shift, 0, OP3 (Rd, Rn, Rm_SFT), QL_I3SAMER, F_SF),
|
2014-09-03 15:40:41 +02:00
|
|
|
|
/* LSE extension (atomic). */
|
2016-04-28 10:11:03 +02:00
|
|
|
|
_LSE_INSN ("casb", 0x8a07c00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("cash", 0x48a07c00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("cas", 0x88a07c00, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("casab", 0x8e07c00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("caslb", 0x8a0fc00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("casalb", 0x8e0fc00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("casah", 0x48e07c00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("caslh", 0x48a0fc00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("casalh", 0x48e0fc00, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("casa", 0x88e07c00, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("casl", 0x88a0fc00, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("casal", 0x88e0fc00, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("casp", 0x8207c00, 0xbfe0fc00, lse_atomic, OP5 (Rs, PAIRREG, Rt, PAIRREG, ADDR_SIMPLE), QL_R4NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("caspa", 0x8607c00, 0xbfe0fc00, lse_atomic, OP5 (Rs, PAIRREG, Rt, PAIRREG, ADDR_SIMPLE), QL_R4NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("caspl", 0x820fc00, 0xbfe0fc00, lse_atomic, OP5 (Rs, PAIRREG, Rt, PAIRREG, ADDR_SIMPLE), QL_R4NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("caspal", 0x860fc00, 0xbfe0fc00, lse_atomic, OP5 (Rs, PAIRREG, Rt, PAIRREG, ADDR_SIMPLE), QL_R4NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("swpb", 0x38208000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swph", 0x78208000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swp", 0xb8208000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("swpab", 0x38a08000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swplb", 0x38608000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swpalb", 0x38e08000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swpah", 0x78a08000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swplh", 0x78608000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swpalh", 0x78e08000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("swpa", 0xb8a08000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("swpl", 0xb8608000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("swpal", 0xb8e08000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldaddb", 0x38200000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldaddh", 0x78200000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldadd", 0xb8200000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldaddab", 0x38a00000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldaddlb", 0x38600000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldaddalb", 0x38e00000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldaddah", 0x78a00000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldaddlh", 0x78600000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldaddalh", 0x78e00000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldadda", 0xb8a00000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldaddl", 0xb8600000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldaddal", 0xb8e00000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldclrb", 0x38201000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclrh", 0x78201000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclr", 0xb8201000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclrab", 0x38a01000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldclrlb", 0x38601000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclralb", 0x38e01000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldclrah", 0x78a01000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldclrlh", 0x78601000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclralh", 0x78e01000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldclra", 0xb8a01000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldclrl", 0xb8601000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldclral", 0xb8e01000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldeorb", 0x38202000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeorh", 0x78202000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeor", 0xb8202000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeorab", 0x38a02000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldeorlb", 0x38602000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeoralb", 0x38e02000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldeorah", 0x78a02000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldeorlh", 0x78602000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeoralh", 0x78e02000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldeora", 0xb8a02000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldeorl", 0xb8602000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldeoral", 0xb8e02000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsetb", 0x38203000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldseth", 0x78203000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldset", 0xb8203000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsetab", 0x38a03000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsetlb", 0x38603000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsetalb", 0x38e03000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsetah", 0x78a03000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsetlh", 0x78603000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsetalh", 0x78e03000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldseta", 0xb8a03000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsetl", 0xb8603000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsetal", 0xb8e03000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsmaxb", 0x38204000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmaxh", 0x78204000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmax", 0xb8204000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmaxab", 0x38a04000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsmaxlb", 0x38604000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmaxalb", 0x38e04000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsmaxah", 0x78a04000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsmaxlh", 0x78604000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmaxalh", 0x78e04000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsmaxa", 0xb8a04000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsmaxl", 0xb8604000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmaxal", 0xb8e04000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsminb", 0x38205000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsminh", 0x78205000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsmin", 0xb8205000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsminab", 0x38a05000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsminlb", 0x38605000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsminalb", 0x38e05000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsminah", 0x78a05000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsminlh", 0x78605000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsminalh", 0x78e05000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldsmina", 0xb8a05000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldsminl", 0xb8605000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldsminal", 0xb8e05000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldumaxb", 0x38206000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumaxh", 0x78206000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumax", 0xb8206000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumaxab", 0x38a06000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldumaxlb", 0x38606000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumaxalb", 0x38e06000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldumaxah", 0x78a06000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldumaxlh", 0x78606000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumaxalh", 0x78e06000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldumaxa", 0xb8a06000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("ldumaxl", 0xb8606000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumaxal", 0xb8e06000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("lduminb", 0x38207000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("lduminh", 0x78207000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("ldumin", 0xb8207000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("lduminab", 0x38a07000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("lduminlb", 0x38607000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("lduminalb", 0x38e07000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("lduminah", 0x78a07000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("lduminlh", 0x78607000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("lduminalh", 0x78e07000, 0xffe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_W2_LDST_EXC, 0),
|
|
|
|
|
_LSE_INSN ("ldumina", 0xb8a07000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("lduminl", 0xb8607000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ | F_HAS_ALIAS),
|
|
|
|
|
_LSE_INSN ("lduminal", 0xb8e07000, 0xbfe0fc00, lse_atomic, OP3 (Rs, Rt, ADDR_SIMPLE), QL_R2NIL, F_LSE_SZ),
|
|
|
|
|
_LSE_INSN ("staddb", 0x3820001f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("staddh", 0x7820001f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stadd", 0xb820001f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("staddlb", 0x3860001f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("staddlh", 0x7860001f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("staddl", 0xb860001f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclrb", 0x3820101f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclrh", 0x7820101f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclr", 0xb820101f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclrlb", 0x3860101f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclrlh", 0x7860101f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stclrl", 0xb860101f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steorb", 0x3820201f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steorh", 0x7820201f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steor", 0xb820201f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steorlb", 0x3860201f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steorlh", 0x7860201f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("steorl", 0xb860201f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsetb", 0x3820301f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stseth", 0x7820301f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stset", 0xb820301f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsetlb", 0x3860301f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsetlh", 0x7860301f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsetl", 0xb860301f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmaxb", 0x3820401f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmaxh", 0x7820401f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmax", 0xb820401f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmaxlb", 0x3860401f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmaxlh", 0x7860401f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmaxl", 0xb860401f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsminb", 0x3820501f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsminh", 0x7820501f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsmin", 0xb820501f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsminlb", 0x3860501f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsminlh", 0x7860501f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stsminl", 0xb860501f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumaxb", 0x3820601f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumaxh", 0x7820601f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumax", 0xb820601f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumaxlb", 0x3860601f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumaxlh", 0x7860601f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumaxl", 0xb860601f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stuminb", 0x3820701f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stuminh", 0x7820701f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stumin", 0xb820701f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stuminlb", 0x3860701f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stuminlh", 0x7860701f, 0xffe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_W1_LDST_EXC, F_ALIAS),
|
|
|
|
|
_LSE_INSN ("stuminl", 0xb860701f, 0xbfe0fc1f, lse_atomic, OP2 (Rs, ADDR_SIMPLE), QL_R1NIL, F_LSE_SZ | F_ALIAS),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Move wide (immediate). */
|
2016-08-23 10:02:38 +02:00
|
|
|
|
CORE_INSN ("movn", 0x12800000, 0x7f800000, movewide, OP_MOVN, OP2 (Rd, HALF), QL_DST_R, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("mov", 0x12800000, 0x7f800000, movewide, OP_MOV_IMM_WIDEN, OP2 (Rd, IMM_MOV), QL_DST_R, F_SF | F_ALIAS | F_CONV),
|
|
|
|
|
CORE_INSN ("movz", 0x52800000, 0x7f800000, movewide, OP_MOVZ, OP2 (Rd, HALF), QL_DST_R, F_SF | F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("mov", 0x52800000, 0x7f800000, movewide, OP_MOV_IMM_WIDE, OP2 (Rd, IMM_MOV), QL_DST_R, F_SF | F_ALIAS | F_CONV),
|
|
|
|
|
CORE_INSN ("movk", 0x72800000, 0x7f800000, movewide, OP_MOVK, OP2 (Rd, HALF), QL_DST_R, F_SF),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* PC-rel. addressing. */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("adr", 0x10000000, 0x9f000000, pcreladdr, 0, OP2 (Rd, ADDR_PCREL21), QL_ADRP, 0),
|
|
|
|
|
CORE_INSN ("adrp", 0x90000000, 0x9f000000, pcreladdr, 0, OP2 (Rd, ADDR_ADRP), QL_ADRP, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* System. */
|
Implement Read/Write constraints on system registers on AArch64
This patch adds constraints for read and write only system registers with the
msr and mrs instructions. The code will treat having both flags set and none
set as the same. These flags add constraints that must be matched up. e.g. a
system register with a READ only flag set, can only be used with mrs. If The
constraint fails a warning is emitted.
Examples of the warnings generated:
test.s: Assembler messages:
test.s:5: Warning: specified register cannot be written to at operand 1 -- `msr dbgdtrrx_el0,x3'
test.s:7: Warning: specified register cannot be read from at operand 2 -- `mrs x3,dbgdtrtx_el0'
test.s:8: Warning: specified register cannot be written to at operand 1 -- `msr midr_el1,x3'
and disassembly notes:
0000000000000000 <main>:
0: d5130503 msr dbgdtrtx_el0, x3
4: d5130503 msr dbgdtrtx_el0, x3
8: d5330503 mrs x3, dbgdtrrx_el0
c: d5330503 mrs x3, dbgdtrrx_el0
10: d5180003 msr midr_el1, x3 ; note: writing to a read-only register.
Note that because dbgdtrrx_el0 and dbgdtrtx_el0 have the same encoding, during
disassembly the constraints are use to disambiguate between the two. An exact
constraint match is always prefered over partial ones if available.
As always the warnings can be suppressed with -w and also be made errors using
warnings as errors.
binutils/
PR binutils/21446
* doc/binutils.texi (-M): Document AArch64 options.
gas/
PR binutils/21446
* testsuite/gas/aarch64/illegal-sysreg-2.s: Fix pmbidr_el1 test.
* testsuite/gas/aarch64/illegal-sysreg-2.l: Likewise.
* testsuite/gas/aarch64/illegal-sysreg-2.d: Likewise.
* testsuite/gas/aarch64/sysreg-diagnostic.s: New.
* testsuite/gas/aarch64/sysreg-diagnostic.l: New.
* testsuite/gas/aarch64/sysreg-diagnostic.d: New.
include/
PR binutils/21446
* opcode/aarch64.h (F_SYS_READ, F_SYS_WRITE): New.
opcodes/
PR binutils/21446
* aarch64-asm.c (opintl.h): Include.
(aarch64_ins_sysreg): Enforce read/write constraints.
* aarch64-dis.c (aarch64_ext_sysreg): Likewise.
* aarch64-opc.h (F_DEPRECATED, F_ARCHEXT, F_HASXT): Moved here.
(F_REG_READ, F_REG_WRITE): New.
* aarch64-opc.c (aarch64_print_operand): Generate notes for
AARCH64_OPND_SYSREG.
(F_DEPRECATED, F_ARCHEXT, F_HASXT): Move to aarch64-opc.h.
(aarch64_sys_regs): Add constraints to currentel, midr_el1, ctr_el0,
mpidr_el1, revidr_el1, aidr_el1, dczid_el0, id_dfr0_el1, id_pfr0_el1,
id_pfr1_el1, id_afr0_el1, id_mmfr0_el1, id_mmfr1_el1, id_mmfr2_el1,
id_mmfr3_el1, id_mmfr4_el1, id_isar0_el1, id_isar1_el1, id_isar2_el1,
id_isar3_el1, id_isar4_el1, id_isar5_el1, mvfr0_el1, mvfr1_el1,
mvfr2_el1, ccsidr_el1, id_aa64pfr0_el1, id_aa64pfr1_el1,
id_aa64dfr0_el1, id_aa64dfr1_el1, id_aa64isar0_el1, id_aa64isar1_el1,
id_aa64mmfr0_el1, id_aa64mmfr1_el1, id_aa64mmfr2_el1, id_aa64afr0_el1,
id_aa64afr0_el1, id_aa64afr1_el1, id_aa64zfr0_el1, clidr_el1,
csselr_el1, vsesr_el2, erridr_el1, erxfr_el1, rvbar_el1, rvbar_el2,
rvbar_el3, isr_el1, tpidrro_el0, cntfrq_el0, cntpct_el0, cntvct_el0,
mdccsr_el0, dbgdtrrx_el0, dbgdtrtx_el0, osdtrrx_el1, osdtrtx_el1,
mdrar_el1, oslar_el1, oslsr_el1, dbgauthstatus_el1, pmbidr_el1,
pmsidr_el1, pmswinc_el0, pmceid0_el0, pmceid1_el0.
* aarch64-tbl.h (aarch64_opcode_table): Add constraints to
msr (F_SYS_WRITE), mrs (F_SYS_READ).
2018-05-15 17:37:20 +02:00
|
|
|
|
CORE_INSN ("msr", 0xd500401f, 0xfff8f01f, ic_system, 0, OP2 (PSTATEFIELD, UIMM4), {}, F_SYS_WRITE),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("hint",0xd503201f, 0xfffff01f, ic_system, 0, OP1 (UIMM7), {}, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("nop", 0xd503201f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
2018-01-09 12:28:04 +01:00
|
|
|
|
CORE_INSN ("csdb",0xd503229f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
[PATCH, BINUTILS, AARCH64, 7/9] Add BTI instruction
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order/bti-branch-target-identification)
The Branch Target Identification instructions (BTI) are allocated to
existing HINT space, using HINT numbers 32, 34, 36, 38, such that
bits[7:6] of the instruction identify the compatibility of the BTI
instruction to different branches.
BTI {<targets>}
where <targets> one of the following, specifying which type of
indirection is allowed:
j : Can be a target of any BR Xn isntruction.
c : Can be a target of any BLR Xn and BR {X16|X17}.
jc: Can be a target of any free branch.
A BTI instruction without any <targets> is the strictest of all and
can not be a target of nay free branch.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_BTI): New.
(AARCH64_ARCH_V8_5): Add AARCH64_FEATURE_BTI by default.
(aarch64_opnd): Add AARCH64_OPND_BTI_TARGET.
(HINT_OPD_CSYNC, HINT_OPD_C, HINT_OPD_J): New macros to
define HINT #imm values.
(HINT_OPD_JC, HINT_OPD_NULL): Likewise.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (HINT_OPD_NOPRINT, HINT_ENCODE): New.
(HINT_FLAG, HINT_VALUE): New macros to encode NO_PRINT flag
with the hint immediate.
* aarch64-opc.c (aarch64_hint_options): New entries for
c, j, jc and default (with HINT_OPD_F_NOPRINT flag) for BTI.
(aarch64_print_operand): Add case for AARCH64_OPND_BTI_TARGET
while checking for HINT_OPD_F_NOPRINT flag.
* aarch64-dis.c (aarch64_ext_hint): Use new HINT_VALUE to
extract value.
* aarch64-tbl.h (aarch64_feature_bti, BTI, BTI_INSN): New.
(aarch64_opcode_table): Add entry for BTI.
(AARCH64_OPERANDS): Add new description for BTI targets.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_bti_operand): New.
(process_omitted_operand): Add case for AARCH64_OPND_BTI_TARGET.
(parse_operands): Likewise.
* testsuite/gas/aarch64/system.d: Update for BTI.
* testsuite/gas/aarch64/bti.s: New.
* testsuite/gas/aarch64/bti.d: New.
* testsuite/gas/aarch64/illegal-bti.d: New.
* testsuite/gas/aarch64/illegal-bti.l: New.
2018-09-26 12:00:49 +02:00
|
|
|
|
BTI_INSN ("bti",0xd503241f, 0xffffff3f, ic_system, OP1 (BTI_TARGET), {}, F_ALIAS | F_OPD0_OPT | F_DEFAULT (0x0)),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("yield", 0xd503203f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
|
|
|
|
CORE_INSN ("wfe", 0xd503205f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
|
|
|
|
CORE_INSN ("wfi", 0xd503207f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
|
|
|
|
CORE_INSN ("sev", 0xd503209f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
|
|
|
|
CORE_INSN ("sevl",0xd50320bf, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
[AArch64] Add ARMv8.3 instructions which are in the NOP space
This patch adds support for a subset of the ARMv8.3 pointer authentication
instructions: XPACLRI, PACIA1716, PACIB1716, AUTIA1716, AUTIA1716, PACIAZ,
PACIASP, PACIBZ, PACISP, AUTIAZ, AUTIASP, AUTIBZ, AUTIBSP.
These are aliases to HINT #0x7, HINT #0x8, HINT #0xa, HINT #0xc, HINT #0xe,
HINT #0x18, HINT #0x19, ..., HINT #0x1f respectively.
For more details about pointer authentication in ARMv8.3 see
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (aarch64_feature_v8_3, ARMV8_3, V8_3_INSN): New.
(arch64_opcode_table): Add xpaclri, pacia1716, pacib1716, autia1716,
autib1716, paciaz, paciasp, pacibz, pacibsp, autiaz, autiasp, autibz,
autibsp.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
gas/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/system-3.s: New.
* testsuite/gas/aarch64/system-3.d: New.
* testsuite/gas/aarch64/system.d: Update expected output.
2016-11-11 11:29:07 +01:00
|
|
|
|
V8_3_INSN ("xpaclri", 0xd50320ff, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("pacia1716", 0xd503211f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("pacib1716", 0xd503215f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autia1716", 0xd503219f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autib1716", 0xd50321df, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{"esb", 0xd503221f, 0xffffffff, ic_system, 0, RAS, OP0 (), {}, F_ALIAS, 0, 0, NULL},
|
|
|
|
|
{"psb", 0xd503223f, 0xffffffff, ic_system, 0, STAT_PROFILE, OP1 (BARRIER_PSB), {}, F_ALIAS, 0, 0, NULL},
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("clrex", 0xd503305f, 0xfffff0ff, ic_system, 0, OP1 (UIMM4), {}, F_OPD0_OPT | F_DEFAULT (0xF)),
|
2018-07-12 16:46:17 +02:00
|
|
|
|
CORE_INSN ("dsb", 0xd503309f, 0xfffff0ff, ic_system, 0, OP1 (BARRIER), {}, F_HAS_ALIAS),
|
|
|
|
|
CORE_INSN ("ssbb", 0xd503309f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
|
|
|
|
CORE_INSN ("pssbb", 0xd503349f, 0xffffffff, ic_system, 0, OP0 (), {}, F_ALIAS),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("dmb", 0xd50330bf, 0xfffff0ff, ic_system, 0, OP1 (BARRIER), {}, 0),
|
|
|
|
|
CORE_INSN ("isb", 0xd50330df, 0xfffff0ff, ic_system, 0, OP1 (BARRIER_ISB), {}, F_OPD0_OPT | F_DEFAULT (0xF)),
|
2018-09-26 11:47:40 +02:00
|
|
|
|
SB_INSN ("sb", 0xd50330ff, 0xffffffff, ic_system, OP0 (), {}, 0),
|
2016-12-13 13:37:18 +01:00
|
|
|
|
CORE_INSN ("sys", 0xd5080000, 0xfff80000, ic_system, 0, OP5 (UIMM3_OP1, CRn, CRm, UIMM3_OP2, Rt), QL_SYS, F_HAS_ALIAS | F_OPD4_OPT | F_DEFAULT (0x1F)),
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("at", 0xd5080000, 0xfff80000, ic_system, 0, OP2 (SYSREG_AT, Rt), QL_SRC_X, F_ALIAS),
|
|
|
|
|
CORE_INSN ("dc", 0xd5080000, 0xfff80000, ic_system, 0, OP2 (SYSREG_DC, Rt), QL_SRC_X, F_ALIAS),
|
|
|
|
|
CORE_INSN ("ic", 0xd5080000, 0xfff80000, ic_system, 0, OP2 (SYSREG_IC, Rt_SYS), QL_SRC_X, F_ALIAS | F_OPD1_OPT | F_DEFAULT (0x1F)),
|
|
|
|
|
CORE_INSN ("tlbi",0xd5080000, 0xfff80000, ic_system, 0, OP2 (SYSREG_TLBI, Rt_SYS), QL_SRC_X, F_ALIAS | F_OPD1_OPT | F_DEFAULT (0x1F)),
|
2018-09-26 11:52:51 +02:00
|
|
|
|
PREDRES_INSN ("cfp", 0xd50b7380, 0xffffffe0, ic_system, OP2 (SYSREG_SR, Rt), QL_SRC_X, F_ALIAS),
|
|
|
|
|
PREDRES_INSN ("dvp", 0xd50b73a0, 0xffffffe0, ic_system, OP2 (SYSREG_SR, Rt), QL_SRC_X, F_ALIAS),
|
|
|
|
|
PREDRES_INSN ("cpp", 0xd50b73e0, 0xffffffe0, ic_system, OP2 (SYSREG_SR, Rt), QL_SRC_X, F_ALIAS),
|
Implement Read/Write constraints on system registers on AArch64
This patch adds constraints for read and write only system registers with the
msr and mrs instructions. The code will treat having both flags set and none
set as the same. These flags add constraints that must be matched up. e.g. a
system register with a READ only flag set, can only be used with mrs. If The
constraint fails a warning is emitted.
Examples of the warnings generated:
test.s: Assembler messages:
test.s:5: Warning: specified register cannot be written to at operand 1 -- `msr dbgdtrrx_el0,x3'
test.s:7: Warning: specified register cannot be read from at operand 2 -- `mrs x3,dbgdtrtx_el0'
test.s:8: Warning: specified register cannot be written to at operand 1 -- `msr midr_el1,x3'
and disassembly notes:
0000000000000000 <main>:
0: d5130503 msr dbgdtrtx_el0, x3
4: d5130503 msr dbgdtrtx_el0, x3
8: d5330503 mrs x3, dbgdtrrx_el0
c: d5330503 mrs x3, dbgdtrrx_el0
10: d5180003 msr midr_el1, x3 ; note: writing to a read-only register.
Note that because dbgdtrrx_el0 and dbgdtrtx_el0 have the same encoding, during
disassembly the constraints are use to disambiguate between the two. An exact
constraint match is always prefered over partial ones if available.
As always the warnings can be suppressed with -w and also be made errors using
warnings as errors.
binutils/
PR binutils/21446
* doc/binutils.texi (-M): Document AArch64 options.
gas/
PR binutils/21446
* testsuite/gas/aarch64/illegal-sysreg-2.s: Fix pmbidr_el1 test.
* testsuite/gas/aarch64/illegal-sysreg-2.l: Likewise.
* testsuite/gas/aarch64/illegal-sysreg-2.d: Likewise.
* testsuite/gas/aarch64/sysreg-diagnostic.s: New.
* testsuite/gas/aarch64/sysreg-diagnostic.l: New.
* testsuite/gas/aarch64/sysreg-diagnostic.d: New.
include/
PR binutils/21446
* opcode/aarch64.h (F_SYS_READ, F_SYS_WRITE): New.
opcodes/
PR binutils/21446
* aarch64-asm.c (opintl.h): Include.
(aarch64_ins_sysreg): Enforce read/write constraints.
* aarch64-dis.c (aarch64_ext_sysreg): Likewise.
* aarch64-opc.h (F_DEPRECATED, F_ARCHEXT, F_HASXT): Moved here.
(F_REG_READ, F_REG_WRITE): New.
* aarch64-opc.c (aarch64_print_operand): Generate notes for
AARCH64_OPND_SYSREG.
(F_DEPRECATED, F_ARCHEXT, F_HASXT): Move to aarch64-opc.h.
(aarch64_sys_regs): Add constraints to currentel, midr_el1, ctr_el0,
mpidr_el1, revidr_el1, aidr_el1, dczid_el0, id_dfr0_el1, id_pfr0_el1,
id_pfr1_el1, id_afr0_el1, id_mmfr0_el1, id_mmfr1_el1, id_mmfr2_el1,
id_mmfr3_el1, id_mmfr4_el1, id_isar0_el1, id_isar1_el1, id_isar2_el1,
id_isar3_el1, id_isar4_el1, id_isar5_el1, mvfr0_el1, mvfr1_el1,
mvfr2_el1, ccsidr_el1, id_aa64pfr0_el1, id_aa64pfr1_el1,
id_aa64dfr0_el1, id_aa64dfr1_el1, id_aa64isar0_el1, id_aa64isar1_el1,
id_aa64mmfr0_el1, id_aa64mmfr1_el1, id_aa64mmfr2_el1, id_aa64afr0_el1,
id_aa64afr0_el1, id_aa64afr1_el1, id_aa64zfr0_el1, clidr_el1,
csselr_el1, vsesr_el2, erridr_el1, erxfr_el1, rvbar_el1, rvbar_el2,
rvbar_el3, isr_el1, tpidrro_el0, cntfrq_el0, cntpct_el0, cntvct_el0,
mdccsr_el0, dbgdtrrx_el0, dbgdtrtx_el0, osdtrrx_el1, osdtrtx_el1,
mdrar_el1, oslar_el1, oslsr_el1, dbgauthstatus_el1, pmbidr_el1,
pmsidr_el1, pmswinc_el0, pmceid0_el0, pmceid1_el0.
* aarch64-tbl.h (aarch64_opcode_table): Add constraints to
msr (F_SYS_WRITE), mrs (F_SYS_READ).
2018-05-15 17:37:20 +02:00
|
|
|
|
CORE_INSN ("msr", 0xd5000000, 0xffe00000, ic_system, 0, OP2 (SYSREG, Rt), QL_SRC_X, F_SYS_WRITE),
|
2016-12-13 13:37:18 +01:00
|
|
|
|
CORE_INSN ("sysl",0xd5280000, 0xfff80000, ic_system, 0, OP5 (Rt, UIMM3_OP1, CRn, CRm, UIMM3_OP2), QL_SYSL, 0),
|
Implement Read/Write constraints on system registers on AArch64
This patch adds constraints for read and write only system registers with the
msr and mrs instructions. The code will treat having both flags set and none
set as the same. These flags add constraints that must be matched up. e.g. a
system register with a READ only flag set, can only be used with mrs. If The
constraint fails a warning is emitted.
Examples of the warnings generated:
test.s: Assembler messages:
test.s:5: Warning: specified register cannot be written to at operand 1 -- `msr dbgdtrrx_el0,x3'
test.s:7: Warning: specified register cannot be read from at operand 2 -- `mrs x3,dbgdtrtx_el0'
test.s:8: Warning: specified register cannot be written to at operand 1 -- `msr midr_el1,x3'
and disassembly notes:
0000000000000000 <main>:
0: d5130503 msr dbgdtrtx_el0, x3
4: d5130503 msr dbgdtrtx_el0, x3
8: d5330503 mrs x3, dbgdtrrx_el0
c: d5330503 mrs x3, dbgdtrrx_el0
10: d5180003 msr midr_el1, x3 ; note: writing to a read-only register.
Note that because dbgdtrrx_el0 and dbgdtrtx_el0 have the same encoding, during
disassembly the constraints are use to disambiguate between the two. An exact
constraint match is always prefered over partial ones if available.
As always the warnings can be suppressed with -w and also be made errors using
warnings as errors.
binutils/
PR binutils/21446
* doc/binutils.texi (-M): Document AArch64 options.
gas/
PR binutils/21446
* testsuite/gas/aarch64/illegal-sysreg-2.s: Fix pmbidr_el1 test.
* testsuite/gas/aarch64/illegal-sysreg-2.l: Likewise.
* testsuite/gas/aarch64/illegal-sysreg-2.d: Likewise.
* testsuite/gas/aarch64/sysreg-diagnostic.s: New.
* testsuite/gas/aarch64/sysreg-diagnostic.l: New.
* testsuite/gas/aarch64/sysreg-diagnostic.d: New.
include/
PR binutils/21446
* opcode/aarch64.h (F_SYS_READ, F_SYS_WRITE): New.
opcodes/
PR binutils/21446
* aarch64-asm.c (opintl.h): Include.
(aarch64_ins_sysreg): Enforce read/write constraints.
* aarch64-dis.c (aarch64_ext_sysreg): Likewise.
* aarch64-opc.h (F_DEPRECATED, F_ARCHEXT, F_HASXT): Moved here.
(F_REG_READ, F_REG_WRITE): New.
* aarch64-opc.c (aarch64_print_operand): Generate notes for
AARCH64_OPND_SYSREG.
(F_DEPRECATED, F_ARCHEXT, F_HASXT): Move to aarch64-opc.h.
(aarch64_sys_regs): Add constraints to currentel, midr_el1, ctr_el0,
mpidr_el1, revidr_el1, aidr_el1, dczid_el0, id_dfr0_el1, id_pfr0_el1,
id_pfr1_el1, id_afr0_el1, id_mmfr0_el1, id_mmfr1_el1, id_mmfr2_el1,
id_mmfr3_el1, id_mmfr4_el1, id_isar0_el1, id_isar1_el1, id_isar2_el1,
id_isar3_el1, id_isar4_el1, id_isar5_el1, mvfr0_el1, mvfr1_el1,
mvfr2_el1, ccsidr_el1, id_aa64pfr0_el1, id_aa64pfr1_el1,
id_aa64dfr0_el1, id_aa64dfr1_el1, id_aa64isar0_el1, id_aa64isar1_el1,
id_aa64mmfr0_el1, id_aa64mmfr1_el1, id_aa64mmfr2_el1, id_aa64afr0_el1,
id_aa64afr0_el1, id_aa64afr1_el1, id_aa64zfr0_el1, clidr_el1,
csselr_el1, vsesr_el2, erridr_el1, erxfr_el1, rvbar_el1, rvbar_el2,
rvbar_el3, isr_el1, tpidrro_el0, cntfrq_el0, cntpct_el0, cntvct_el0,
mdccsr_el0, dbgdtrrx_el0, dbgdtrtx_el0, osdtrrx_el1, osdtrtx_el1,
mdrar_el1, oslar_el1, oslsr_el1, dbgauthstatus_el1, pmbidr_el1,
pmsidr_el1, pmswinc_el0, pmceid0_el0, pmceid1_el0.
* aarch64-tbl.h (aarch64_opcode_table): Add constraints to
msr (F_SYS_WRITE), mrs (F_SYS_READ).
2018-05-15 17:37:20 +02:00
|
|
|
|
CORE_INSN ("mrs", 0xd5200000, 0xffe00000, ic_system, 0, OP2 (Rt, SYSREG), QL_DST_X, F_SYS_READ),
|
[AArch64] Add ARMv8.3 instructions which are in the NOP space
This patch adds support for a subset of the ARMv8.3 pointer authentication
instructions: XPACLRI, PACIA1716, PACIB1716, AUTIA1716, AUTIA1716, PACIAZ,
PACIASP, PACIBZ, PACISP, AUTIAZ, AUTIASP, AUTIBZ, AUTIBSP.
These are aliases to HINT #0x7, HINT #0x8, HINT #0xa, HINT #0xc, HINT #0xe,
HINT #0x18, HINT #0x19, ..., HINT #0x1f respectively.
For more details about pointer authentication in ARMv8.3 see
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
opcodes/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (aarch64_feature_v8_3, ARMV8_3, V8_3_INSN): New.
(arch64_opcode_table): Add xpaclri, pacia1716, pacib1716, autia1716,
autib1716, paciaz, paciasp, pacibz, pacibsp, autiaz, autiasp, autibz,
autibsp.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
gas/
2016-11-11 Szabolcs Nagy <szabolcs.nagy@arm.com>
* testsuite/gas/aarch64/system-3.s: New.
* testsuite/gas/aarch64/system-3.d: New.
* testsuite/gas/aarch64/system.d: Update expected output.
2016-11-11 11:29:07 +01:00
|
|
|
|
V8_3_INSN ("paciaz", 0xd503231f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("paciasp", 0xd503233f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("pacibz", 0xd503235f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("pacibsp", 0xd503237f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autiaz", 0xd503239f, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autiasp", 0xd50323bf, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autibz", 0xd50323df, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
|
|
|
|
V8_3_INSN ("autibsp", 0xd50323ff, 0xffffffff, ic_system, OP0 (), {}, F_ALIAS),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* Test & branch (immediate). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("tbz", 0x36000000, 0x7f000000, testbranch, 0, OP3 (Rt, BIT_NUM, ADDR_PCREL14), QL_PCREL_14, 0),
|
|
|
|
|
CORE_INSN ("tbnz",0x37000000, 0x7f000000, testbranch, 0, OP3 (Rt, BIT_NUM, ADDR_PCREL14), QL_PCREL_14, 0),
|
2012-08-13 16:52:54 +02:00
|
|
|
|
/* The old UAL conditional branch mnemonics (to aid portability). */
|
2016-08-23 10:01:54 +02:00
|
|
|
|
CORE_INSN ("beq", 0x54000000, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bne", 0x54000001, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bcs", 0x54000002, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bhs", 0x54000002, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bcc", 0x54000003, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("blo", 0x54000003, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bmi", 0x54000004, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bpl", 0x54000005, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bvs", 0x54000006, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bvc", 0x54000007, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bhi", 0x54000008, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bls", 0x54000009, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bge", 0x5400000a, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("blt", 0x5400000b, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("bgt", 0x5400000c, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
|
|
|
|
CORE_INSN ("ble", 0x5400000d, 0xff00001f, condbranch, 0, OP1 (ADDR_PCREL19), QL_PCREL_NIL, F_ALIAS | F_PSEUDO),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
/* SVE instructions. */
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fmov", 0x2539c000, 0xff3fe000, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_FPIMM8), OP_SVE_VU_HSD, F_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("fmov", 0x0510c000, 0xff30e000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg4_16, SVE_FPIMM8), OP_SVE_VMU_HSD, F_ALIAS, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("mov", 0x04603000, 0xffe0fc00, sve_misc, OP_MOV_Z_Z, OP2 (SVE_Zd, SVE_Zn), OP_SVE_DD, F_ALIAS | F_MISC, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("mov", 0x05202000, 0xff20fc00, sve_index, OP_MOV_Z_V, OP2 (SVE_Zd, SVE_VZn), OP_SVE_VV_BHSDQ, F_ALIAS | F_MISC, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("mov", 0x05203800, 0xff3ffc00, sve_size_bhsd, 0, OP2 (SVE_Zd, Rn_SP), OP_SVE_VR_BHSD, F_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("mov", 0x25804000, 0xfff0c210, sve_misc, OP_MOV_P_P, OP2 (SVE_Pd, SVE_Pn), OP_SVE_BB, F_ALIAS | F_MISC, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("mov", 0x05202000, 0xff20fc00, sve_index, OP_MOV_Z_Zi, OP2 (SVE_Zd, SVE_Zn_INDEX), OP_SVE_VV_BHSDQ, F_ALIAS | F_MISC, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("mov", 0x05c00000, 0xfffc0000, sve_limm, 0, OP2 (SVE_Zd, SVE_LIMM_MOV), OP_SVE_VU_BHSD, F_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("mov", 0x2538c000, 0xff3fc000, sve_size_bhsd, 0, OP2 (SVE_Zd, SVE_ASIMM), OP_SVE_VU_BHSD, F_ALIAS, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("mov", 0x05208000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Vn), OP_SVE_VMV_BHSD, F_ALIAS, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("mov", 0x0520c000, 0xff20c000, sve_size_bhsd, OP_MOV_Z_P_Z, OP3 (SVE_Zd, SVE_Pg4_10, SVE_Zn), OP_SVE_VMV_BHSD, F_ALIAS | F_MISC, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("mov", 0x0528a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, Rn_SP), OP_SVE_VMR_BHSD, F_ALIAS, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("mov", 0x25004000, 0xfff0c210, sve_misc, OP_MOVZ_P_P_P, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, F_ALIAS | F_MISC, 0),
|
|
|
|
|
_SVE_INSN ("mov", 0x25004210, 0xfff0c210, sve_misc, OP_MOVM_P_P_P, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BMB, F_ALIAS | F_MISC, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("mov", 0x05100000, 0xff308000, sve_cpy, 0, OP3 (SVE_Zd, SVE_Pg4_16, SVE_ASIMM), OP_SVE_VPU_BHSD, F_ALIAS, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("movs", 0x25c04000, 0xfff0c210, sve_misc, OP_MOVS_P_P, OP2 (SVE_Pd, SVE_Pn), OP_SVE_BB, F_ALIAS | F_MISC, 0),
|
|
|
|
|
_SVE_INSN ("movs", 0x25404000, 0xfff0c210, sve_misc, OP_MOVZS_P_P_P, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, F_ALIAS | F_MISC, 0),
|
|
|
|
|
_SVE_INSN ("not", 0x25004200, 0xfff0c210, sve_misc, OP_NOT_P_P_P_Z, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, F_ALIAS | F_MISC, 0),
|
|
|
|
|
_SVE_INSN ("nots", 0x25404200, 0xfff0c210, sve_misc, OP_NOTS_P_P_P_Z, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, F_ALIAS | F_MISC, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("abs", 0x0416a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("add", 0x04200000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("add", 0x2520c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("add", 0x04000000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("addpl", 0x04605000, 0xffe0f800, sve_misc, 0, OP3 (Rd_SP, SVE_Rn_SP, SVE_SIMM6), OP_SVE_XXU, 0, 0),
|
|
|
|
|
_SVE_INSN ("addvl", 0x04205000, 0xffe0f800, sve_misc, 0, OP3 (Rd_SP, SVE_Rn_SP, SVE_SIMM6), OP_SVE_XXU, 0, 0),
|
|
|
|
|
_SVE_INSN ("adr", 0x0420a000, 0xffe0f000, sve_misc, 0, OP2 (SVE_Zd, SVE_ADDR_ZZ_SXTW), OP_SVE_DD, 0, 0),
|
|
|
|
|
_SVE_INSN ("adr", 0x0460a000, 0xffe0f000, sve_misc, 0, OP2 (SVE_Zd, SVE_ADDR_ZZ_UXTW), OP_SVE_DD, 0, 0),
|
|
|
|
|
_SVE_INSN ("adr", 0x04a0a000, 0xffa0f000, sve_size_sd, 0, OP2 (SVE_Zd, SVE_ADDR_ZZ_LSL), OP_SVE_VV_SD, 0, 0),
|
|
|
|
|
_SVE_INSN ("and", 0x04203000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_DDD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("and", 0x05800000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_LIMM), OP_SVE_VVU_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("and", 0x041a0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("and", 0x25004000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("ands", 0x25404000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("andv", 0x041a2000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("asr", 0x04208000, 0xff20fc00, sve_size_bhs, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("asr", 0x04209000, 0xff20fc00, sve_shift_unpred, 0, OP3 (SVE_Zd, SVE_Zn, SVE_SHRIMM_UNPRED), OP_SVE_VVU_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("asr", 0x04108000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("asr", 0x04188000, 0xff3fe000, sve_size_bhs, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVD_BHS, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("asr", 0x04008000, 0xff3fe000, sve_shift_pred, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_SHRIMM_PRED), OP_SVE_VMVU_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("asrd", 0x04048000, 0xff3fe000, sve_shift_pred, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_SHRIMM_PRED), OP_SVE_VMVU_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("asrr", 0x04148000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("bic", 0x04e03000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_DDD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("bic", 0x041b0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("bic", 0x25004010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("bics", 0x25404010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brka", 0x25104000, 0xffffc200, sve_pred_zm, 0, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BPB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkas", 0x25504000, 0xffffc210, sve_misc, 0, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkb", 0x25904000, 0xffffc200, sve_pred_zm, 0, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BPB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkbs", 0x25d04000, 0xffffc210, sve_misc, 0, OP3 (SVE_Pd, SVE_Pg4_10, SVE_Pn), OP_SVE_BZB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkn", 0x25184000, 0xffffc210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pd), OP_SVE_BZBB, 0, 3),
|
|
|
|
|
_SVE_INSN ("brkns", 0x25584000, 0xffffc210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pd), OP_SVE_BZBB, 0, 3),
|
|
|
|
|
_SVE_INSN ("brkpa", 0x2500c000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkpas", 0x2540c000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkpb", 0x2500c010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("brkpbs", 0x2540c010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("clasta", 0x05288000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VUVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("clasta", 0x052a8000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Vd, SVE_Pg3, SVE_Vd, SVE_Zm_5), OP_SVE_VUVV_BHSD, 0, 2),
|
|
|
|
|
_SVE_INSN ("clasta", 0x0530a000, 0xff3fe000, sve_size_bhsd, 0, OP4 (Rd, SVE_Pg3, Rd, SVE_Zm_5), OP_SVE_RURV_BHSD, 0, 2),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("clastb", 0x05298000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VUVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("clastb", 0x052b8000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Vd, SVE_Pg3, SVE_Vd, SVE_Zm_5), OP_SVE_VUVV_BHSD, 0, 2),
|
|
|
|
|
_SVE_INSN ("clastb", 0x0531a000, 0xff3fe000, sve_size_bhsd, 0, OP4 (Rd, SVE_Pg3, Rd, SVE_Zm_5), OP_SVE_RURV_BHSD, 0, 2),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("cls", 0x0418a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("clz", 0x0419a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("cmpeq", 0x24002000, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpeq", 0x2400a000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpeq", 0x25008000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpge", 0x24004000, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpge", 0x24008000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("cmpge", 0x25000000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpgt", 0x24004010, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpgt", 0x24008010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("cmpgt", 0x25000010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmphi", 0x24000010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("cmphi", 0x2400c010, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmphi", 0x24200010, 0xff202010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_UIMM7), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmphs", 0x24000000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("cmphs", 0x2400c000, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmphs", 0x24200000, 0xff202010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_UIMM7), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmple", 0x24006010, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmple", 0x25002010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmplo", 0x2400e000, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmplo", 0x24202000, 0xff202010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_UIMM7), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpls", 0x2400e010, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpls", 0x24202010, 0xff202010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_UIMM7), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmplt", 0x24006000, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmplt", 0x25002000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpne", 0x24002010, 0xff20e010, sve_size_bhs, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpne", 0x2400a010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cmpne", 0x25008010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SIMM5), OP_SVE_VZVU_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("cnot", 0x041ba000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("cnt", 0x041aa000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("cntb", 0x0420e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("cntd", 0x04e0e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("cnth", 0x0460e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("cntp", 0x25208000, 0xff3fc200, sve_size_bhsd, 0, OP3 (Rd, SVE_Pg4_10, SVE_Pn), OP_SVE_XUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("cntw", 0x04a0e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("compact", 0x05a18000, 0xffbfe000, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_SD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("cpy", 0x05208000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Vn), OP_SVE_VMV_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("cpy", 0x0528a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, Rn_SP), OP_SVE_VMR_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("cpy", 0x05100000, 0xff308000, sve_cpy, 0, OP3 (SVE_Zd, SVE_Pg4_16, SVE_ASIMM), OP_SVE_VPU_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ctermeq", 0x25a02000, 0xffa0fc1f, sve_size_sd, 0, OP2 (Rn, Rm), OP_SVE_RR, 0, 0),
|
|
|
|
|
_SVE_INSN ("ctermne", 0x25a02010, 0xffa0fc1f, sve_size_sd, 0, OP2 (Rn, Rm), OP_SVE_RR, 0, 0),
|
|
|
|
|
_SVE_INSN ("decb", 0x0430e400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("decd", 0x04f0c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("decd", 0x04f0e400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("dech", 0x0470c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("dech", 0x0470e400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("decp", 0x252d8000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("decp", 0x252d8800, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("decw", 0x04b0c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("decw", 0x04b0e400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("dup", 0x05203800, 0xff3ffc00, sve_size_bhsd, 0, OP2 (SVE_Zd, Rn_SP), OP_SVE_VR_BHSD, F_HAS_ALIAS, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("dup", 0x05202000, 0xff20fc00, sve_index, 0, OP2 (SVE_Zd, SVE_Zn_INDEX), OP_SVE_VV_BHSDQ, F_HAS_ALIAS, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("dup", 0x2538c000, 0xff3fc000, sve_size_bhsd, 0, OP2 (SVE_Zd, SVE_ASIMM), OP_SVE_VU_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("dupm", 0x05c00000, 0xfffc0000, sve_limm, 0, OP2 (SVE_Zd, SVE_LIMM), OP_SVE_VU_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("eor", 0x04a03000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_DDD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("eor", 0x05400000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_LIMM), OP_SVE_VVU_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("eor", 0x04190000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("eor", 0x25004200, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("eors", 0x25404200, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("eorv", 0x04192000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("ext", 0x05200000, 0xffe0e000, sve_misc, 0, OP4 (SVE_Zd, SVE_Zd, SVE_Zm_5, SVE_UIMM8_53), OP_SVE_BBBU, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("fabd", 0x65088000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fabs", 0x041ca000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("facge", 0x6500c010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("facgt", 0x6500e010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("fadd", 0x65000000, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fadd", 0x65008000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fadd", 0x65188000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_HALF_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fadda", 0x65182000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Vd, SVE_Pg3, SVE_Vd, SVE_Zm_5), OP_SVE_VUVV_HSD, 0, 2),
|
|
|
|
|
_SVE_INSN ("faddv", 0x65002000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fcadd", 0x64008000, 0xff3ee000, sve_size_hsd, 0, OP5 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5, SVE_IMM_ROT1), OP_SVE_VMVVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fcmla", 0x64000000, 0xff208000, sve_size_hsd, 0, OP5 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16, IMM_ROT2), OP_SVE_VMVVU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcmla", 0x64a01000, 0xffe0f000, sve_misc, 0, OP4 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX, SVE_IMM_ROT2), OP_SVE_VVVU_H, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcmla", 0x64e01000, 0xffe0f000, sve_misc, 0, OP4 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX, SVE_IMM_ROT2), OP_SVE_VVVU_S, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fcmeq", 0x65122000, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmeq", 0x65006000, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmge", 0x65102000, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmge", 0x65004000, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("fcmgt", 0x65102010, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmgt", 0x65004010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("fcmle", 0x65112010, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmlt", 0x65112000, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmne", 0x65132000, 0xff3fe010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, FPIMM0), OP_SVE_VZV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmne", 0x65006010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fcmuo", 0x6500c000, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VZVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fcpy", 0x0510c000, 0xff30e000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg4_16, SVE_FPIMM8), OP_SVE_VMU_HSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x6588a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMS, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x6589a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMH, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x65c8a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMD, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x65c9a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMH, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x65caa000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMD, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvt", 0x65cba000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMS, 0, C_SCAN_MOVPRFX | C_MAX_ELEM, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x655aa000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x655ca000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x655ea000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x659ca000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x65d8a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x65dca000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzs", 0x65dea000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x655ba000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x655da000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x655fa000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x659da000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x65d9a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x65dda000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fcvtzu", 0x65dfa000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fdiv", 0x650d8000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fdivr", 0x650c8000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fdup", 0x2539c000, 0xff3fe000, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_FPIMM8), OP_SVE_VU_HSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("fexpa", 0x0420b800, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmad", 0x65208000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_5, SVE_Za_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmax", 0x65068000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fmax", 0x651e8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_ZERO_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fmaxnm", 0x65048000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fmaxnm", 0x651c8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_ZERO_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fmaxnmv", 0x65042000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fmaxv", 0x65062000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmin", 0x65078000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fmin", 0x651f8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_ZERO_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fminnm", 0x65058000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fminnm", 0x651d8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_ZERO_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fminnmv", 0x65052000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("fminv", 0x65072000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmla", 0x65200000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmla", 0x64200000, 0xffa0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_22_INDEX), OP_SVE_VVV_H, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmla", 0x64a00000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX), OP_SVE_VVV_S, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmla", 0x64e00000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX), OP_SVE_VVV_D, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmls", 0x65202000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmls", 0x64200400, 0xffa0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_22_INDEX), OP_SVE_VVV_H, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmls", 0x64a00400, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX), OP_SVE_VVV_S, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmls", 0x64e00400, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX), OP_SVE_VVV_D, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fmsb", 0x6520a000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_5, SVE_Za_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fmul", 0x65000800, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmul", 0x65028000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fmul", 0x651a8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_HALF_TWO), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fmul", 0x64202000, 0xffa0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_22_INDEX), OP_SVE_VVV_H, 0, 0),
|
|
|
|
|
_SVE_INSN ("fmul", 0x64a02000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX), OP_SVE_VVV_S, 0, 0),
|
|
|
|
|
_SVE_INSN ("fmul", 0x64e02000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX), OP_SVE_VVV_D, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmulx", 0x650a8000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fneg", 0x041da000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fnmad", 0x6520c000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_5, SVE_Za_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fnmla", 0x65204000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fnmls", 0x65206000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("fnmsb", 0x6520e000, 0xff20e000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_5, SVE_Za_16), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("frecpe", 0x650e3000, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("frecps", 0x65001800, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("frecpx", 0x650ca000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frinta", 0x6504a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frinti", 0x6507a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frintm", 0x6502a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frintn", 0x6500a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frintp", 0x6501a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frintx", 0x6506a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("frintz", 0x6503a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("frsqrte", 0x650f3000, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("frsqrts", 0x65001c00, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fscale", 0x65098000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fsqrt", 0x650da000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("fsub", 0x65000400, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fsub", 0x65018000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fsub", 0x65198000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_HALF_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fsubr", 0x65038000, 0xff3fe000, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("fsubr", 0x651b8000, 0xff3fe3c0, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_I1_HALF_ONE), OP_SVE_VMVU_HSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("ftmad", 0x65108000, 0xff38fc00, sve_size_hsd, 0, OP4 (SVE_Zd, SVE_Zd, SVE_Zm_5, SVE_UIMM3), OP_SVE_VVVU_HSD, 0, C_SCAN_MOVPRFX, 1),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("ftsmul", 0x65000c00, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("ftssel", 0x0420b000, 0xff20fc00, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_HSD, 0, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("incb", 0x0430e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("incd", 0x04f0c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("incd", 0x04f0e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("inch", 0x0470c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("inch", 0x0470e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("incp", 0x252c8000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("incp", 0x252c8800, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("incw", 0x04b0c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("incw", 0x04b0e000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("index", 0x04204c00, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, Rn, Rm), OP_SVE_VRR_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("index", 0x04204000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_SIMM5, SVE_SIMM5B), OP_SVE_VUU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("index", 0x04204400, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, Rn, SIMM5), OP_SVE_VRU_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("index", 0x04204800, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_SIMM5, Rm), OP_SVE_VUR_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("insr", 0x05243800, 0xff3ffc00, sve_size_bhsd, 0, OP2 (SVE_Zd, SVE_Rm), OP_SVE_VR_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("insr", 0x05343800, 0xff3ffc00, sve_size_bhsd, 0, OP2 (SVE_Zd, SVE_Vm), OP_SVE_VV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("lasta", 0x0520a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (Rd, SVE_Pg3, SVE_Zn), OP_SVE_RUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("lasta", 0x05228000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("lastb", 0x0521a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (Rd, SVE_Pg3, SVE_Zn), OP_SVE_RUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("lastb", 0x05238000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0x84004000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa4004000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa4204000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa4404000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa4604000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xc4004000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xc440c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0x8420c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa400a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa420a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa440a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xa460a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1b", 0xc420c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xa5e04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xc5804000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xc5a04000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW3_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xc5c0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xc5e0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL3), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xa5e0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1d", 0xc5a0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x8), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0x84804000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0x84a04000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4a04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4c04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4e04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xc4804000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xc4a04000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xc4c0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xc4e0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0x84a0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4a0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4c0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xa4e0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1h", 0xc4a0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rb", 0x84408000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rb", 0x8440a000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rb", 0x8440c000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rb", 0x8440e000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rd", 0x85c0e000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x8), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rh", 0x84c0a000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x2), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rh", 0x84c0c000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x2), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rh", 0x84c0e000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x2), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("ld1rqb", 0xa4002000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x16), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqb", 0xa4000000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqd", 0xa5802000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x16), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqd", 0xa5800000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqh", 0xa4802000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x16), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqh", 0xa4800000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqw", 0xa5002000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x16), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rqw", 0xa5000000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ld1rsb", 0x85c08000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rsb", 0x85c0a000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rsb", 0x85c0c000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rsh", 0x85408000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x2), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rsh", 0x8540a000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x2), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rsw", 0x84c08000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x4), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rw", 0x8540c000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x4), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1rw", 0x8540e000, 0xffc0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_U6x4), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0x84000000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa5804000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa5a04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa5c04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xc4000000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xc4408000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0x84208000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa580a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa5a0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xa5c0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sb", 0xc4208000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0x84800000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0x84a00000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xa5004000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xa5204000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xc4800000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xc4a00000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xc4c08000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xc4e08000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0x84a08000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xa500a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xa520a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sh", 0xc4a08000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xa4804000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xc5000000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xc5200000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xc5408000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xc5608000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xa480a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1sw", 0xc5208000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0x85004000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0x85204000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xa5404000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xa5604000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xc5004000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xc5204000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xc540c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xc560c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0x8520c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xa540a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xa560a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld1w", 0xc520c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ld2b", 0xa420c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2b", 0xa420e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_BZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2d", 0xa5a0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2d", 0xa5a0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_DZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2h", 0xa4a0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2h", 0xa4a0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_HZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2w", 0xa520c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld2w", 0xa520e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_SZU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("ld3b", 0xa440c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3b", 0xa440e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_BZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3d", 0xa5c0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3d", 0xa5c0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_DZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3h", 0xa4c0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3h", 0xa4c0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_HZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3w", 0xa540c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld3w", 0xa540e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_SZU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("ld4b", 0xa460c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4b", 0xa460e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_BZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4d", 0xa5e0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4d", 0xa5e0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_DZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4h", 0xa4e0c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4h", 0xa4e0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_HZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4w", 0xa560c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("ld4w", 0xa560e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_SZU, F_OD(4), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0x84006000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_BZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_BZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_HZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_HZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_SZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xa4606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1b", 0xc4006000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1b", 0xc440e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1b", 0x8420e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1b", 0xc420e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1d", 0xa5e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL3), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1d", 0xa5e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1d", 0xc5806000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1d", 0xc5a06000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW3_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1d", 0xc5c0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1d", 0xc5e0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL3), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1d", 0xc5a0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x8), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0x84806000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0x84a06000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL1), OP_SVE_HZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_HZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL1), OP_SVE_SZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL1), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xa4e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1h", 0xc4806000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0xc4a06000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0xc4c0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0xc4e0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0x84a0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1h", 0xc4a0e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0x84002000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_SZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR), OP_SVE_HZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xa5c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_HZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sb", 0xc4002000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sb", 0xc440a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sb", 0x8420a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sb", 0xc420a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sh", 0x84802000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0x84a02000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0xa5006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL1), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sh", 0xa5006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sh", 0xa5206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL1), OP_SVE_SZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sh", 0xa5206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sh", 0xc4802000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0xc4a02000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0xc4c0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0xc4e0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0x84a0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sh", 0xc4a0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sw", 0xa4806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL2), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1sw", 0xa4806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1sw", 0xc5002000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sw", 0xc5202000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sw", 0xc540a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sw", 0xc560a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1sw", 0xc520a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1w", 0x85006000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0x85206000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0xa5406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL2), OP_SVE_SZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1w", 0xa5406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_SZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1w", 0xa5606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RR_LSL2), OP_SVE_DZU, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
_SVE_INSN ("ldff1w", 0xa5606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_R), OP_SVE_DZU, F_OD(1), 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldff1w", 0xc5006000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0xc5206000, 0xffa0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0xc540e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0xc560e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_DZD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0x8520e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_SZS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldff1w", 0xc520e000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_DZD, F_OD(1), 0),
|
2018-03-28 10:44:45 +02:00
|
|
|
|
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("ldnf1b", 0xa410a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1b", 0xa430a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1b", 0xa450a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1b", 0xa470a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1d", 0xa5f0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1h", 0xa4b0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1h", 0xa4d0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1h", 0xa4f0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sb", 0xa590a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sb", 0xa5b0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sb", 0xa5d0a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sh", 0xa510a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sh", 0xa530a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1sw", 0xa490a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1w", 0xa550a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnf1w", 0xa570a000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1b", 0xa400c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1b", 0xa400e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_BZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1d", 0xa580c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1d", 0xa580e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1h", 0xa480c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1h", 0xa480e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1w", 0xa500c000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldnt1w", 0xa500e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SZU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("ldr", 0x85800000, 0xffc0e010, sve_misc, 0, OP2 (SVE_Pt, SVE_ADDR_RI_S9xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("ldr", 0x85804000, 0xffc0e000, sve_misc, 0, OP2 (SVE_Zt, SVE_ADDR_RI_S9xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("lsl", 0x04208c00, 0xff20fc00, sve_size_bhs, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("lsl", 0x04209c00, 0xff20fc00, sve_shift_unpred, 0, OP3 (SVE_Zd, SVE_Zn, SVE_SHLIMM_UNPRED), OP_SVE_VVU_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("lsl", 0x04138000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lsl", 0x041b8000, 0xff3fe000, sve_size_bhs, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVD_BHS, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lsl", 0x04038000, 0xff3fe000, sve_shift_pred, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_SHLIMM_PRED), OP_SVE_VMVU_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lslr", 0x04178000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("lsr", 0x04208400, 0xff20fc00, sve_size_bhs, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("lsr", 0x04209400, 0xff20fc00, sve_shift_unpred, 0, OP3 (SVE_Zd, SVE_Zn, SVE_SHRIMM_UNPRED), OP_SVE_VVU_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("lsr", 0x04118000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lsr", 0x04198000, 0xff3fe000, sve_size_bhs, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVD_BHS, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lsr", 0x04018000, 0xff3fe000, sve_shift_pred, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_SHRIMM_PRED), OP_SVE_VMVU_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("lsrr", 0x04158000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("mad", 0x0400c000, 0xff20e000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_16, SVE_Za_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("mla", 0x04004000, 0xff20e000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("mls", 0x04006000, 0xff20e000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zn, SVE_Zm_16), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("movprfx", 0x0420bc00, 0xfffffc00, sve_misc, 0, OP2 (SVE_Zd, SVE_Zn), {}, F_SCAN, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("movprfx", 0x04102000, 0xff3ee000, sve_movprfx, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VPV_BHSD, F_SCAN, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("msb", 0x0400e000, 0xff20e000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zm_16, SVE_Za_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("mul", 0x2530c000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_SIMM8), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("mul", 0x04100000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("nand", 0x25804210, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("nands", 0x25c04210, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("neg", 0x0417a000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("nor", 0x25804200, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("nors", 0x25c04200, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("not", 0x041ea000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("orn", 0x25804010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("orns", 0x25c04010, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, 0, 0),
|
|
|
|
|
_SVE_INSN ("orr", 0x04603000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_DDD, F_HAS_ALIAS, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("orr", 0x05000000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_LIMM), OP_SVE_VVU_BHSD, F_HAS_ALIAS, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("orr", 0x04180000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("orr", 0x25804000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("orrs", 0x25c04000, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BZBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("orv", 0x04182000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("pfalse", 0x2518e400, 0xfffffff0, sve_misc, 0, OP1 (SVE_Pd), OP_SVE_B, 0, 0),
|
|
|
|
|
_SVE_INSN ("pfirst", 0x2558c000, 0xfffffe10, sve_misc, 0, OP3 (SVE_Pd, SVE_Pg4_5, SVE_Pd), OP_SVE_BUB, 0, 2),
|
|
|
|
|
_SVE_INSN ("pnext", 0x2519c400, 0xff3ffe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pg4_5, SVE_Pd), OP_SVE_VUV_BHSD, 0, 2),
|
|
|
|
|
_SVE_INSN ("prfb", 0x8400c000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RX), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0x84200000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0xc4200000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW_22), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0xc4608000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0x8400e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0x85c00000, 0xffc0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RI_S6xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfb", 0xc400e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0x84206000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW3_22), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0x8580c000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RX_LSL3), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0xc4206000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW3_22), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0xc460e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_LSL3), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0x8580e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x8), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0x85c06000, 0xffc0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RI_S6xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfd", 0xc580e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x8), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0x84202000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0x8480c000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RX_LSL1), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0xc4202000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW1_22), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0xc460a000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0x8480e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0x85c02000, 0xffc0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RI_S6xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfh", 0xc480e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0x84204000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0x8500c000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RX_LSL2), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0xc4204000, 0xffa0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_XTW2_22), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0xc460c000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0x8500e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_UUS, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0x85c04000, 0xffc0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_RI_S6xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("prfw", 0xc500e000, 0xffe0e010, sve_misc, 0, OP3 (SVE_PRFOP, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_UUD, 0, 0),
|
|
|
|
|
_SVE_INSN ("ptest", 0x2550c000, 0xffffc21f, sve_misc, 0, OP2 (SVE_Pg4_10, SVE_Pn), OP_SVE_UB, 0, 0),
|
|
|
|
|
_SVE_INSN ("ptrue", 0x2518e000, 0xff3ffc10, sve_size_bhsd, 0, OP2 (SVE_Pd, SVE_PATTERN), OP_SVE_VU_BHSD, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("ptrues", 0x2519e000, 0xff3ffc10, sve_size_bhsd, 0, OP2 (SVE_Pd, SVE_PATTERN), OP_SVE_VU_BHSD, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("punpkhi", 0x05314000, 0xfffffe10, sve_misc, 0, OP2 (SVE_Pd, SVE_Pn), OP_SVE_HB, 0, 0),
|
|
|
|
|
_SVE_INSN ("punpklo", 0x05304000, 0xfffffe10, sve_misc, 0, OP2 (SVE_Pd, SVE_Pn), OP_SVE_HB, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("rbit", 0x05278000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_BHSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("rdffr", 0x2519f000, 0xfffffff0, sve_misc, 0, OP1 (SVE_Pd), OP_SVE_B, 0, 0),
|
|
|
|
|
_SVE_INSN ("rdffr", 0x2518f000, 0xfffffe10, sve_misc, 0, OP2 (SVE_Pd, SVE_Pg4_5), OP_SVE_BZ, 0, 0),
|
|
|
|
|
_SVE_INSN ("rdffrs", 0x2558f000, 0xfffffe10, sve_misc, 0, OP2 (SVE_Pd, SVE_Pg4_5), OP_SVE_BZ, 0, 0),
|
|
|
|
|
_SVE_INSN ("rdvl", 0x04bf5000, 0xfffff800, sve_misc, 0, OP2 (Rd, SVE_SIMM6), OP_SVE_XU, 0, 0),
|
|
|
|
|
_SVE_INSN ("rev", 0x05344000, 0xff3ffe10, sve_size_bhsd, 0, OP2 (SVE_Pd, SVE_Pn), OP_SVE_VV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("rev", 0x05383800, 0xff3ffc00, sve_size_bhsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("revb", 0x05248000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("revh", 0x05a58000, 0xffbfe000, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_SD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("revw", 0x05e68000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sabd", 0x040c0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("saddv", 0x04002000, 0xff3fe000, sve_size_bhs, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_DUV_BHS, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("scvtf", 0x6552a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x6554a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x6594a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x65d0a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x6556a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x65d4a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("scvtf", 0x65d6a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sdiv", 0x04940000, 0xffbfe000, sve_size_sd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_SD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("sdivr", 0x04960000, 0xffbfe000, sve_size_sd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_SD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("sdot", 0x44800000, 0xffa0fc00, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_SD_BH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sdot", 0x44a00000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX), OP_SVE_VVV_S_B, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sdot", 0x44e00000, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX), OP_SVE_VVV_D_H, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sel", 0x0520c000, 0xff20c000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg4_10, SVE_Zn, SVE_Zm_16), OP_SVE_VUVV_BHSD, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("sel", 0x25004210, 0xfff0c210, sve_misc, 0, OP4 (SVE_Pd, SVE_Pg4_10, SVE_Pn, SVE_Pm), OP_SVE_BUBB, F_HAS_ALIAS, 0),
|
|
|
|
|
_SVE_INSN ("setffr", 0x252c9000, 0xffffffff, sve_misc, 0, OP0 (), {}, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("smax", 0x2528c000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_SIMM8), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("smax", 0x04080000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("smaxv", 0x04082000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("smin", 0x252ac000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_SIMM8), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("smin", 0x040a0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sminv", 0x040a2000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("smulh", 0x04120000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("splice", 0x052c8000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VUVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqadd", 0x04201000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqadd", 0x2524c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqdecb", 0x0430f800, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqdecb", 0x0420f800, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqdecd", 0x04e0c800, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqdecd", 0x04f0f800, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqdecd", 0x04e0f800, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqdech", 0x0460c800, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqdech", 0x0470f800, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqdech", 0x0460f800, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqdecp", 0x252a8000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqdecp", 0x252a8c00, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("sqdecp", 0x252a8800, 0xff3ffe00, sve_size_bhsd, 0, OP3 (Rd, SVE_Pg4_5, Rd), OP_SVE_XVW_BHSD, 0, 2),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqdecw", 0x04a0c800, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqdecw", 0x04b0f800, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqdecw", 0x04a0f800, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
|
|
|
|
_SVE_INSN ("sqincb", 0x0430f000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqincb", 0x0420f000, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqincd", 0x04e0c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqincd", 0x04f0f000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqincd", 0x04e0f000, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqinch", 0x0460c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqinch", 0x0470f000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqinch", 0x0460f000, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqincp", 0x25288000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqincp", 0x25288c00, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("sqincp", 0x25288800, 0xff3ffe00, sve_size_bhsd, 0, OP3 (Rd, SVE_Pg4_5, Rd), OP_SVE_XVW_BHSD, 0, 2),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqincw", 0x04a0c000, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sqincw", 0x04b0f000, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("sqincw", 0x04a0f000, 0xfff0fc00, sve_misc, 0, OP3 (Rd, Rd, SVE_PATTERN_SCALED), OP_SVE_XWU, F_OPD2_OPT | F_DEFAULT(31), 1),
|
|
|
|
|
_SVE_INSN ("sqsub", 0x04201800, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sqsub", 0x2526c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("st1b", 0xe4004000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe4008000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe400a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe4204000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe4404000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe4408000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe4604000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe400e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_BUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe420e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe440a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe440e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe460a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1b", 0xe460e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5808000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe580a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5a08000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW3_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5a0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL3), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5e04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5c0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x8), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1d", 0xe5e0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4808000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe480a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4a04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4a08000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4a0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL1), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4c04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4c08000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4e04000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4e08000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW1_14), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4a0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4c0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4c0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4e0a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x2), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1h", 0xe4e0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5008000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe500a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5208000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_14), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe520a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_LSL2), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5404000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5408000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW_14), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5604000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe5608000, 0xffe0a000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RZ_XTW2_14), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe540a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_DUD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe540e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe560a000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_ZI_U5x4), OP_SVE_SUS, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st1w", 0xe560e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("st2b", 0xe4206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2b", 0xe430e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_BUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2d", 0xe5a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2d", 0xe5b0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_DUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2h", 0xe4a06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2h", 0xe4b0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_HUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2w", 0xe5206000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st2w", 0xe530e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x2xVL), OP_SVE_SUU, F_OD(2), 0),
|
|
|
|
|
_SVE_INSN ("st3b", 0xe4406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3b", 0xe450e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_BUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3d", 0xe5c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3d", 0xe5d0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_DUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3h", 0xe4c06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3h", 0xe4d0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_HUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3w", 0xe5406000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st3w", 0xe550e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x3xVL), OP_SVE_SUU, F_OD(3), 0),
|
|
|
|
|
_SVE_INSN ("st4b", 0xe4606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4b", 0xe470e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_BUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4d", 0xe5e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4d", 0xe5f0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_DUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4h", 0xe4e06000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4h", 0xe4f0e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_HUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4w", 0xe5606000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("st4w", 0xe570e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4x4xVL), OP_SVE_SUU, F_OD(4), 0),
|
|
|
|
|
_SVE_INSN ("stnt1b", 0xe4006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX), OP_SVE_BUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1b", 0xe410e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_BUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1d", 0xe5806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL3), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1d", 0xe590e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_DUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1h", 0xe4806000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL1), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1h", 0xe490e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_HUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1w", 0xe5006000, 0xffe0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RX_LSL2), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("stnt1w", 0xe510e000, 0xfff0e000, sve_misc, 0, OP3 (SVE_ZtxN, SVE_Pg3, SVE_ADDR_RI_S4xVL), OP_SVE_SUU, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("str", 0xe5800000, 0xffc0e010, sve_misc, 0, OP2 (SVE_Pt, SVE_ADDR_RI_S9xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("str", 0xe5804000, 0xffc0e000, sve_misc, 0, OP2 (SVE_Zt, SVE_ADDR_RI_S9xVL), {}, 0, 0),
|
|
|
|
|
_SVE_INSN ("sub", 0x04200400, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sub", 0x2521c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("sub", 0x04010000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("subr", 0x2523c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("subr", 0x04030000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("sunpkhi", 0x05313800, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("sunpklo", 0x05303800, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD_BHS, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("sxtb", 0x0410a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sxth", 0x0492a000, 0xffbfe000, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_SD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("sxtw", 0x04d4a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("tbl", 0x05203000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_ZnxN, SVE_Zm_16), OP_SVE_VVV_BHSD, F_OD(1), 0),
|
|
|
|
|
_SVE_INSN ("trn1", 0x05205000, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("trn1", 0x05207000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("trn2", 0x05205400, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("trn2", 0x05207400, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uabd", 0x040d0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uaddv", 0x04012000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_DUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("ucvtf", 0x6553a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x6555a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x6595a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x65d1a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMS, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x6557a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_HMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x65d5a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_SMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("ucvtf", 0x65d7a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("udiv", 0x04950000, 0xffbfe000, sve_size_sd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_SD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("udivr", 0x04970000, 0xffbfe000, sve_size_sd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_SD, 0, C_SCAN_MOVPRFX, 2),
|
|
|
|
|
_SVE_INSNC ("udot", 0x44800400, 0xffa0fc00, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_SD_BH, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("udot", 0x44a00400, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm3_INDEX), OP_SVE_VVV_S_B, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("udot", 0x44e00400, 0xffe0fc00, sve_misc, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm4_INDEX), OP_SVE_VVV_D_H, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("umax", 0x2529c000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_UIMM8), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("umax", 0x04090000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("umaxv", 0x04092000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("umin", 0x252bc000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_UIMM8), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
|
|
|
|
_SVE_INSNC ("umin", 0x040b0000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uminv", 0x040b2000, 0xff3fe000, sve_size_bhsd, 0, OP3 (SVE_Vd, SVE_Pg3, SVE_Zn), OP_SVE_VUV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("umulh", 0x04130000, 0xff3fe000, sve_size_bhsd, 0, OP4 (SVE_Zd, SVE_Pg3, SVE_Zd, SVE_Zm_5), OP_SVE_VMVV_BHSD, 0, C_SCAN_MOVPRFX, 2),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqadd", 0x04201400, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqadd", 0x2525c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqdecb", 0x0420fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqdecb", 0x0430fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqdecd", 0x04e0cc00, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqdecd", 0x04e0fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqdecd", 0x04f0fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqdech", 0x0460cc00, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqdech", 0x0460fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqdech", 0x0470fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqdecp", 0x252b8000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqdecp", 0x252b8800, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_WV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("uqdecp", 0x252b8c00, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqdecw", 0x04a0cc00, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqdecw", 0x04a0fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqdecw", 0x04b0fc00, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqincb", 0x0420f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqincb", 0x0430f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqincd", 0x04e0c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_DU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqincd", 0x04e0f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqincd", 0x04f0f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqinch", 0x0460c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_HU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqinch", 0x0460f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqinch", 0x0470f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqincp", 0x25298000, 0xff3ffe00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Pg4_5), OP_SVE_VU_HSD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqincp", 0x25298800, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_WV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("uqincp", 0x25298c00, 0xff3ffe00, sve_size_bhsd, 0, OP2 (Rd, SVE_Pg4_5), OP_SVE_XV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqincw", 0x04a0c400, 0xfff0fc00, sve_misc, 0, OP2 (SVE_Zd, SVE_PATTERN_SCALED), OP_SVE_SU, F_OPD1_OPT | F_DEFAULT(31), C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uqincw", 0x04a0f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_WU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqincw", 0x04b0f400, 0xfff0fc00, sve_misc, 0, OP2 (Rd, SVE_PATTERN_SCALED), OP_SVE_XU, F_OPD1_OPT | F_DEFAULT(31), 0),
|
|
|
|
|
_SVE_INSN ("uqsub", 0x04201c00, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uqsub", 0x2527c000, 0xff3fc000, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zd, SVE_AIMM), OP_SVE_VVU_BHSD, 0, C_SCAN_MOVPRFX, 1),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uunpkhi", 0x05333800, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD_BHS, 0, 0),
|
|
|
|
|
_SVE_INSN ("uunpklo", 0x05323800, 0xff3ffc00, sve_size_hsd, 0, OP2 (SVE_Zd, SVE_Zn), OP_SVE_VV_HSD_BHS, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("uxtb", 0x0411a000, 0xff3fe000, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_HSD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("uxth", 0x0493a000, 0xffbfe000, sve_size_sd, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_VMV_SD, 0, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("uxtw", 0x04d5a000, 0xffffe000, sve_misc, 0, OP3 (SVE_Zd, SVE_Pg3, SVE_Zn), OP_SVE_DMD, 0, C_SCAN_MOVPRFX, 0),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("uzp1", 0x05204800, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("uzp1", 0x05206800, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("uzp2", 0x05204c00, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("uzp2", 0x05206c00, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilele", 0x25200410, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VWW_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilele", 0x25201410, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VXX_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilelo", 0x25200c00, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VWW_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilelo", 0x25201c00, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VXX_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilels", 0x25200c10, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VWW_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilels", 0x25201c10, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VXX_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilelt", 0x25200400, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VWW_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("whilelt", 0x25201400, 0xff20fc10, sve_size_bhsd, 0, OP3 (SVE_Pd, Rn, Rm), OP_SVE_VXX_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("wrffr", 0x25289000, 0xfffffe1f, sve_misc, 0, OP1 (SVE_Pn), OP_SVE_B, 0, 0),
|
|
|
|
|
_SVE_INSN ("zip1", 0x05204000, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("zip1", 0x05206000, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("zip2", 0x05204400, 0xff30fe10, sve_size_bhsd, 0, OP3 (SVE_Pd, SVE_Pn, SVE_Pm), OP_SVE_VVV_BHSD, 0, 0),
|
|
|
|
|
_SVE_INSN ("zip2", 0x05206400, 0xff20fc00, sve_size_bhsd, 0, OP3 (SVE_Zd, SVE_Zn, SVE_Zm_16), OP_SVE_VVV_BHSD, 0, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("bic", 0x05800000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_INV_LIMM), OP_SVE_VVU_BHSD, F_ALIAS | F_PSEUDO, C_SCAN_MOVPRFX, 1),
|
[AArch64][SVE 31/32] Add SVE instructions
This patch adds the SVE instruction definitions and associated OP_*
enum values.
include/
* opcode/aarch64.h (AARCH64_FEATURE_SVE): New macro.
(OP_MOV_P_P, OP_MOV_Z_P_Z, OP_MOV_Z_V, OP_MOV_Z_Z, OP_MOV_Z_Zi)
(OP_MOVM_P_P_P, OP_MOVS_P_P, OP_MOVZS_P_P_P, OP_MOVZ_P_P_P)
(OP_NOTS_P_P_P_Z, OP_NOT_P_P_P_Z): New aarch64_ops.
opcodes/
* aarch64-tbl.h (OP_SVE_B, OP_SVE_BB, OP_SVE_BBBU, OP_SVE_BMB)
(OP_SVE_BPB, OP_SVE_BUB, OP_SVE_BUBB, OP_SVE_BUU, OP_SVE_BZ)
(OP_SVE_BZB, OP_SVE_BZBB, OP_SVE_BZU, OP_SVE_DD, OP_SVE_DDD)
(OP_SVE_DMD, OP_SVE_DMH, OP_SVE_DMS, OP_SVE_DU, OP_SVE_DUD, OP_SVE_DUU)
(OP_SVE_DUV_BHS, OP_SVE_DUV_BHSD, OP_SVE_DZD, OP_SVE_DZU, OP_SVE_HB)
(OP_SVE_HMD, OP_SVE_HMS, OP_SVE_HU, OP_SVE_HUU, OP_SVE_HZU, OP_SVE_RR)
(OP_SVE_RURV_BHSD, OP_SVE_RUV_BHSD, OP_SVE_SMD, OP_SVE_SMH, OP_SVE_SMS)
(OP_SVE_SU, OP_SVE_SUS, OP_SVE_SUU, OP_SVE_SZS, OP_SVE_SZU, OP_SVE_UB)
(OP_SVE_UUD, OP_SVE_UUS, OP_SVE_VMR_BHSD, OP_SVE_VMU_SD)
(OP_SVE_VMVD_BHS, OP_SVE_VMVU_BHSD, OP_SVE_VMVU_SD, OP_SVE_VMVV_BHSD)
(OP_SVE_VMVV_SD, OP_SVE_VMV_BHSD, OP_SVE_VMV_HSD, OP_SVE_VMV_SD)
(OP_SVE_VM_SD, OP_SVE_VPU_BHSD, OP_SVE_VPV_BHSD, OP_SVE_VRR_BHSD)
(OP_SVE_VRU_BHSD, OP_SVE_VR_BHSD, OP_SVE_VUR_BHSD, OP_SVE_VUU_BHSD)
(OP_SVE_VUVV_BHSD, OP_SVE_VUVV_SD, OP_SVE_VUV_BHSD, OP_SVE_VUV_SD)
(OP_SVE_VU_BHSD, OP_SVE_VU_HSD, OP_SVE_VU_SD, OP_SVE_VVD_BHS)
(OP_SVE_VVU_BHSD, OP_SVE_VVVU_SD, OP_SVE_VVV_BHSD, OP_SVE_VVV_SD)
(OP_SVE_VV_BHSD, OP_SVE_VV_HSD_BHS, OP_SVE_VV_SD, OP_SVE_VWW_BHSD)
(OP_SVE_VXX_BHSD, OP_SVE_VZVD_BHS, OP_SVE_VZVU_BHSD, OP_SVE_VZVV_BHSD)
(OP_SVE_VZVV_SD, OP_SVE_VZV_SD, OP_SVE_V_SD, OP_SVE_WU, OP_SVE_WV_BHSD)
(OP_SVE_XU, OP_SVE_XUV_BHSD, OP_SVE_XVW_BHSD, OP_SVE_XV_BHSD)
(OP_SVE_XWU, OP_SVE_XXU): New macros.
(aarch64_feature_sve): New variable.
(SVE): New macro.
(_SVE_INSN): Likewise.
(aarch64_opcode_table): Add SVE instructions.
* aarch64-opc.h (extract_fields): Declare.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.c (do_misc_encoding): Handle the new SVE aarch64_ops.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.c (extract_fields): Make global.
(do_misc_decoding): Handle the new SVE aarch64_ops.
* aarch64-dis-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document the "sve" feature.
* config/tc-aarch64.c (REG_TYPE_R_Z_BHSDQ_VZP): New register type.
(get_reg_expected_msg): Handle it.
(parse_operands): When parsing operands of an SVE instruction,
disallow immediates that match REG_TYPE_R_Z_BHSDQ_VZP.
(aarch64_features): Add an entry for SVE.
2016-09-21 17:58:48 +02:00
|
|
|
|
_SVE_INSN ("cmple", 0x24008000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_BHSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("cmplo", 0x24000010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_BHSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("cmpls", 0x24000000, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_BHSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("cmplt", 0x24008010, 0xff20e010, sve_size_bhsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_BHSD, F_ALIAS | F_PSEUDO, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("eon", 0x05400000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_INV_LIMM), OP_SVE_VVU_BHSD, F_ALIAS | F_PSEUDO, C_SCAN_MOVPRFX, 1),
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
_SVE_INSN ("facle", 0x6500c010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_HSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("faclt", 0x6500e010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_HSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("fcmle", 0x65004000, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_HSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("fcmlt", 0x65004010, 0xff20e010, sve_size_hsd, 0, OP4 (SVE_Pd, SVE_Pg3, SVE_Zm_16, SVE_Zn), OP_SVE_VZVV_HSD, F_ALIAS | F_PSEUDO, 0),
|
|
|
|
|
_SVE_INSN ("fmov", 0x2538c000, 0xff3fffe0, sve_size_hsd, 0, OP2 (SVE_Zd, FPIMM0), OP_SVE_V_HSD, F_ALIAS | F_PSEUDO, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
_SVE_INSNC ("fmov", 0x05104000, 0xff30ffe0, sve_size_hsd, 0, OP3 (SVE_Zd, SVE_Pg4_16, FPIMM0), OP_SVE_VM_HSD, F_ALIAS | F_PSEUDO, C_SCAN_MOVPRFX, 0),
|
|
|
|
|
_SVE_INSNC ("orn", 0x05000000, 0xfffc0000, sve_limm, 0, OP3 (SVE_Zd, SVE_Zd, SVE_INV_LIMM), OP_SVE_VVU_BHSD, F_ALIAS | F_PSEUDO, C_SCAN_MOVPRFX, 1),
|
2016-04-28 10:11:03 +02:00
|
|
|
|
|
2017-06-28 12:09:01 +02:00
|
|
|
|
/* SIMD Dot Product (optional in v8.2-A). */
|
|
|
|
|
DOT_INSN ("udot", 0x2e009400, 0xbf20fc00, dotproduct, OP3 (Vd, Vn, Vm), QL_V3DOT, F_SIZEQ),
|
|
|
|
|
DOT_INSN ("sdot", 0xe009400, 0xbf20fc00, dotproduct, OP3 (Vd, Vn, Vm), QL_V3DOT, F_SIZEQ),
|
2018-05-16 13:13:42 +02:00
|
|
|
|
DOT_INSN ("udot", 0x2f00e000, 0xbf00f400, dotproduct, OP3 (Vd, Vn, Em), QL_V2DOT, F_SIZEQ),
|
|
|
|
|
DOT_INSN ("sdot", 0xf00e000, 0xbf00f400, dotproduct, OP3 (Vd, Vn, Em), QL_V2DOT, F_SIZEQ),
|
Add assembler and disassembler support for the new Armv8.4-a instructions for AArch64.
Some of these instructions have been back-ported as optional extensions to
Armv8.2-a and higher, but others are only available for Armv8.4-a.
opcodes/
* aarch64-tbl.h (sha512h, sha512h2, sha512su0, sha512su1, eor3): New.
(rax1, xar, bcax, sm3ss1, sm3tt1a, sm3tt1b, sm3tt2a, sm3tt2b): New.
(sm3partw1, sm3partw2, sm4e, sm4ekey, fmlal, fmlsl): New.
(fmlal2, fmlsl2, cfinv, rmif, setf8, setf16, stlurb): New.
(ldapurb, ldapursb, stlurh, ldapurh, ldapursh, stlur): New.
(ldapur, ldapursw, stlur): New.
* aarch64-dis-2.c: Regenerate.
gas/
* testsuite/gas/aarch64/armv8_4-a-illegal.d: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.l: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.s: New.
* testsuite/gas/aarch64/armv8_4-a.d: New.
* testsuite/gas/aarch64/armv8_4-a.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_3-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_4-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.l: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.d: New.
2017-11-16 17:07:07 +01:00
|
|
|
|
/* Crypto SHA2 (optional in ARMv8.2-a). */
|
|
|
|
|
SHA2_INSN ("sha512h", 0xce608000, 0xffe0fc00, cryptosha2, OP3 (Fd, Fn, Vm), QL_SHA512UPT, 0),
|
|
|
|
|
SHA2_INSN ("sha512h2", 0xce608400, 0xffe0fc00, cryptosha2, OP3 (Fd, Fn, Vm), QL_SHA512UPT, 0),
|
|
|
|
|
SHA2_INSN ("sha512su0", 0xcec08000, 0xfffffc00, cryptosha2, OP2 (Vd, Vn), QL_V2SAME2D, 0),
|
|
|
|
|
SHA2_INSN ("sha512su1", 0xce608800, 0xffe0fc00, cryptosha2, OP3 (Vd, Vn, Vm), QL_V3SAME2D, 0),
|
|
|
|
|
/* Crypto SHA3 (optional in ARMv8.2-a). */
|
|
|
|
|
SHA3_INSN ("eor3", 0xce000000, 0xffe08000, cryptosha3, OP4 (Vd, Vn, Vm, Va), QL_V4SAME16B, 0),
|
|
|
|
|
SHA3_INSN ("rax1", 0xce608c00, 0xffe0fc00, cryptosha3, OP3 (Vd, Vn, Vm), QL_V3SAME2D, 0),
|
|
|
|
|
SHA3_INSN ("xar", 0xce800000, 0xffe00000, cryptosha3, OP4 (Vd, Vn, Vm, IMM), QL_XAR, 0),
|
|
|
|
|
SHA3_INSN ("bcax", 0xce200000, 0xffe08000, cryptosha3, OP4 (Vd, Vn, Vm, Va), QL_V4SAME16B, 0),
|
|
|
|
|
/* Crypto SM3 (optional in ARMv8.2-a). */
|
|
|
|
|
SM4_INSN ("sm3ss1", 0xce400000, 0xffe08000, cryptosm3, OP4 (Vd, Vn, Vm, Va), QL_V4SAME4S, 0),
|
|
|
|
|
SM4_INSN ("sm3tt1a", 0xce408000, 0xffe0cc00, cryptosm3, OP3 (Vd, Vn, Em), QL_SM3TT, 0),
|
|
|
|
|
SM4_INSN ("sm3tt1b", 0xce408400, 0xffe0cc00, cryptosm3, OP3 (Vd, Vn, Em), QL_SM3TT, 0),
|
|
|
|
|
SM4_INSN ("sm3tt2a", 0xce408800, 0xffe0cc00, cryptosm3, OP3 (Vd, Vn, Em), QL_SM3TT, 0),
|
|
|
|
|
SM4_INSN ("sm3tt2b", 0xce408c00, 0xffe0cc00, cryptosm3, OP3 (Vd, Vn, Em), QL_SM3TT, 0),
|
|
|
|
|
SM4_INSN ("sm3partw1", 0xce60c000, 0xffe0fc00, cryptosm3, OP3 (Vd, Vn, Vm), QL_V3SAME4S, 0),
|
|
|
|
|
SM4_INSN ("sm3partw2", 0xce60c400, 0xffe0fc00, cryptosm3, OP3 (Vd, Vn, Vm), QL_V3SAME4S, 0),
|
|
|
|
|
/* Crypto SM4 (optional in ARMv8.2-a). */
|
|
|
|
|
SM4_INSN ("sm4e", 0xcec08400, 0xfffffc00, cryptosm4, OP2 (Vd, Vn), QL_V2SAME4S, 0),
|
|
|
|
|
SM4_INSN ("sm4ekey", 0xce60c800, 0xffe0fc00, cryptosm4, OP3 (Vd, Vn, Vm), QL_V3SAME4S, 0),
|
|
|
|
|
/* Crypto FP16 (optional in ARMv8.2-a). */
|
|
|
|
|
FP16_V8_2_INSN ("fmlal", 0xe20ec00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl", 0xea0ec00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlal2", 0x2e20cc00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl2", 0x2ea0cc00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML2S, 0),
|
|
|
|
|
|
|
|
|
|
FP16_V8_2_INSN ("fmlal", 0x4e20ec00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl", 0x4ea0ec00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlal2", 0x6e20cc00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl2", 0x6ea0cc00, 0xffa0fc00, asimdsame, OP3 (Vd, Vn, Vm), QL_V3FML4S, 0),
|
|
|
|
|
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
FP16_V8_2_INSN ("fmlal", 0xf800000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl", 0xf804000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlal2", 0x2f808000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML2S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl2", 0x2f80c000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML2S, 0),
|
Add assembler and disassembler support for the new Armv8.4-a instructions for AArch64.
Some of these instructions have been back-ported as optional extensions to
Armv8.2-a and higher, but others are only available for Armv8.4-a.
opcodes/
* aarch64-tbl.h (sha512h, sha512h2, sha512su0, sha512su1, eor3): New.
(rax1, xar, bcax, sm3ss1, sm3tt1a, sm3tt1b, sm3tt2a, sm3tt2b): New.
(sm3partw1, sm3partw2, sm4e, sm4ekey, fmlal, fmlsl): New.
(fmlal2, fmlsl2, cfinv, rmif, setf8, setf16, stlurb): New.
(ldapurb, ldapursb, stlurh, ldapurh, ldapursh, stlur): New.
(ldapur, ldapursw, stlur): New.
* aarch64-dis-2.c: Regenerate.
gas/
* testsuite/gas/aarch64/armv8_4-a-illegal.d: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.l: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.s: New.
* testsuite/gas/aarch64/armv8_4-a.d: New.
* testsuite/gas/aarch64/armv8_4-a.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_3-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_4-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.l: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.d: New.
2017-11-16 17:07:07 +01:00
|
|
|
|
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
FP16_V8_2_INSN ("fmlal", 0x4f800000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl", 0x4f804000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlal2", 0x6f808000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML4S, 0),
|
|
|
|
|
FP16_V8_2_INSN ("fmlsl2", 0x6f80c000, 0xffc0f400, asimdelem, OP3 (Vd, Vn, Em16), QL_V2FML4S, 0),
|
Add assembler and disassembler support for the new Armv8.4-a instructions for AArch64.
Some of these instructions have been back-ported as optional extensions to
Armv8.2-a and higher, but others are only available for Armv8.4-a.
opcodes/
* aarch64-tbl.h (sha512h, sha512h2, sha512su0, sha512su1, eor3): New.
(rax1, xar, bcax, sm3ss1, sm3tt1a, sm3tt1b, sm3tt2a, sm3tt2b): New.
(sm3partw1, sm3partw2, sm4e, sm4ekey, fmlal, fmlsl): New.
(fmlal2, fmlsl2, cfinv, rmif, setf8, setf16, stlurb): New.
(ldapurb, ldapursb, stlurh, ldapurh, ldapursh, stlur): New.
(ldapur, ldapursw, stlur): New.
* aarch64-dis-2.c: Regenerate.
gas/
* testsuite/gas/aarch64/armv8_4-a-illegal.d: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.l: New.
* testsuite/gas/aarch64/armv8_4-a-illegal.s: New.
* testsuite/gas/aarch64/armv8_4-a.d: New.
* testsuite/gas/aarch64/armv8_4-a.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_3-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_4-a-crypto-fp16.d: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.s: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.l: New.
* testsuite/gas/aarch64/armv8_2-a-crypto-fp16-illegal.d: New.
2017-11-16 17:07:07 +01:00
|
|
|
|
/* System extensions ARMv8.4-a. */
|
|
|
|
|
V8_4_INSN ("cfinv", 0xd500401f, 0xffffffff, ic_system, OP0 (), {}, 0),
|
|
|
|
|
V8_4_INSN ("rmif", 0xba000400, 0xffe07c10, ic_system, OP3 (Rn, IMM_2, MASK), QL_RMIF, 0),
|
|
|
|
|
V8_4_INSN ("setf8", 0x3a00080d, 0xfffffc1f, ic_system, OP1 (Rn), QL_SETF, 0),
|
|
|
|
|
V8_4_INSN ("setf16", 0x3a00480d, 0xfffffc1f, ic_system, OP1 (Rn), QL_SETF, 0),
|
|
|
|
|
/* Memory access instructions ARMv8.4-a. */
|
|
|
|
|
V8_4_INSN ("stlurb" , 0x19000000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapurb", 0x19400000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapursb", 0x19c00000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapursb", 0x19800000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLX, 0),
|
|
|
|
|
V8_4_INSN ("stlurh", 0x59000000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapurh", 0x59400000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapursh", 0x59c00000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapursh", 0x59800000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLX, 0),
|
|
|
|
|
V8_4_INSN ("stlur", 0x99000000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapur", 0x99400000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLW, 0),
|
|
|
|
|
V8_4_INSN ("ldapursw", 0x99800000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLX, 0),
|
|
|
|
|
V8_4_INSN ("stlur", 0xd9000000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLX, 0),
|
|
|
|
|
V8_4_INSN ("ldapur", 0xd9400000, 0xffe00c00, ldst_unscaled, OP2 (Rt, ADDR_OFFSET), QL_STLX, 0),
|
AArch64: Mark sve instructions that require MOVPRFX constraints
This patch series is to allow certain instructions such as the SVE MOVPRFX
instruction to apply a constraint/dependency on the instruction at PC+4.
This patch starts this off by marking which instructions impose the constraint
and which instructions must adhere to the constraint. This is done in a
generic way by extending the verifiers.
* The constraint F_SCAN indicates that an instruction opens a sequence and imposes
a constraint on an instructions following it. The length of the sequence depends
on the instruction itself and it handled in the verifier code.
* The C_SCAN_MOVPRFX flag is used to indicate which constrain the instruction is
checked against. An instruction with both F_SCAN and C_SCAN_MOVPRFX starts a
block for the C_SCAN_MOVPRFX instruction, and one with only C_SCAN_MOVPRFX must
adhere to a previous block constraint is applicable.
The SVE instructions in this list have been marked according to the SVE
specification[1].
[1] https://developer.arm.com/docs/ddi0584/latest/arm-architecture-reference-manual-supplement-the-scalable-vector-extension-sve-for-armv8-a
include/
* opcode/aarch64.h (struct aarch64_opcode): Add constraints,
extend flags field size.
(F_SCAN, C_SCAN_MOVPRFX, C_MAX_ELEM): New.
opcodes/
* aarch64-tbl.h (CORE_INSN, __FP_INSN, SIMD_INSN, CRYP_INSN, _CRC_INSN,
_LSE_INSN, _LOR_INSN, RDMA_INSN, FF16_INSN, SF16_INSN, V8_2_INSN,
_SVE_INSN, V8_3_INSN, CNUM_INSN, RCPC_INSN, SHA2_INSN, AES_INSN,
V8_4_INSN, SHA3_INSN, SM4_INSN, FP16_V8_2_INSN, DOT_INSN): Initialize
constraints.
(_SVE_INSNC): New.
(struct aarch64_opcode): (fjcvtzs, ldpsw, ldpsw, esb, psb): Initialize
constraints.
(movprfx): Change _SVE_INSN into _SVE_INSNC, add C_SCAN_MOVPRFX and
F_SCAN flags.
(msb, mul, neg, not, orr, rbit, revb, revh, revw, sabd, scvtf,
sdiv, sdivr, sdot, smax, smin, smulh, splice, sqadd, sqdecd, sqdech,
sqdecp, sqdecw, sqincd, sqinch, sqincp, sqincw, sqsub, sub, subr, sxtb,
sxth, sxtw, uabd, ucvtf, udiv, udivr, udot, umax, umin, umulh, uqadd,
uqdecd, uqdech, uqdecp, uqdecw, uqincd, uqinch, uqincp, uqincw, uqsub,
uxtb, uxth, uxtw, bic, eon, orn, mov, fmov): Change _SVE_INSN into _SVE_INSNC and add
C_SCAN_MOVPRFX and C_MAX_ELEM constraints.
2018-10-03 19:22:15 +02:00
|
|
|
|
{0, 0, 0, 0, 0, 0, {}, {}, 0, 0, 0, NULL},
|
2012-08-13 16:52:54 +02:00
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
#ifdef AARCH64_OPERANDS
|
|
|
|
|
#undef AARCH64_OPERANDS
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/* Macro-based operand decription; this will be fed into aarch64-gen for it
|
|
|
|
|
to generate the structure aarch64_operands and the function
|
|
|
|
|
aarch64_insert_operand and aarch64_extract_operand.
|
|
|
|
|
|
|
|
|
|
These inserters and extracters in the description execute the conversion
|
|
|
|
|
between the aarch64_opnd_info and value in the operand-related instruction
|
|
|
|
|
field(s). */
|
|
|
|
|
|
|
|
|
|
/* Y expects arguments (left to right) to be operand class, inserter/extractor
|
|
|
|
|
name suffix, operand name, flags, related bitfield(s) and description.
|
|
|
|
|
X only differs from Y by having the operand inserter and extractor names
|
|
|
|
|
listed separately. */
|
|
|
|
|
|
|
|
|
|
#define AARCH64_OPERANDS \
|
|
|
|
|
Y(INT_REG, regno, "Rd", 0, F(FLD_Rd), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rn", 0, F(FLD_Rn), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rm", 0, F(FLD_Rm), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rt", 0, F(FLD_Rt), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rt2", 0, F(FLD_Rt2), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rs", 0, F(FLD_Rs), "an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Ra", 0, F(FLD_Ra), "an integer register") \
|
|
|
|
|
X(INT_REG, ins_regno, ext_regrt_sysins, "Rt_SYS", 0, F(FLD_Rt), \
|
|
|
|
|
"an integer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rd_SP", OPD_F_MAYBE_SP, F(FLD_Rd), \
|
|
|
|
|
"an integer or stack pointer register") \
|
|
|
|
|
Y(INT_REG, regno, "Rn_SP", OPD_F_MAYBE_SP, F(FLD_Rn), \
|
|
|
|
|
"an integer or stack pointer register") \
|
2016-11-11 11:39:46 +01:00
|
|
|
|
Y(INT_REG, regno, "Rm_SP", OPD_F_MAYBE_SP, F(FLD_Rm), \
|
|
|
|
|
"an integer or stack pointer register") \
|
2014-09-03 15:40:41 +02:00
|
|
|
|
X(INT_REG, 0, ext_regno_pair, "PAIRREG", 0, F(), \
|
|
|
|
|
"the second reg of a pair") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(MODIFIED_REG, reg_extended, "Rm_EXT", 0, F(), \
|
|
|
|
|
"an integer register with optional extension") \
|
|
|
|
|
Y(MODIFIED_REG, reg_shifted, "Rm_SFT", 0, F(), \
|
|
|
|
|
"an integer register with optional shift") \
|
|
|
|
|
Y(FP_REG, regno, "Fd", 0, F(FLD_Rd), "a floating-point register") \
|
|
|
|
|
Y(FP_REG, regno, "Fn", 0, F(FLD_Rn), "a floating-point register") \
|
|
|
|
|
Y(FP_REG, regno, "Fm", 0, F(FLD_Rm), "a floating-point register") \
|
|
|
|
|
Y(FP_REG, regno, "Fa", 0, F(FLD_Ra), "a floating-point register") \
|
|
|
|
|
Y(FP_REG, ft, "Ft", 0, F(FLD_Rt), "a floating-point register") \
|
|
|
|
|
Y(FP_REG, regno, "Ft2", 0, F(FLD_Rt2), "a floating-point register") \
|
|
|
|
|
Y(SISD_REG, regno, "Sd", 0, F(FLD_Rd), "a SIMD scalar register") \
|
|
|
|
|
Y(SISD_REG, regno, "Sn", 0, F(FLD_Rn), "a SIMD scalar register") \
|
|
|
|
|
Y(SISD_REG, regno, "Sm", 0, F(FLD_Rm), "a SIMD scalar register") \
|
Adds the new Fields and Operand types for the new instructions in Armv8.4-a.
gas/
* config/tc-aarch64.c (process_omitted_operand):
Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2
and AARCH64_OPND_IMM_2.
(parse_operands): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_IMM_2, AARCH64_OPND_MASK
and AARCH64_OPND_ADDR_OFFSET.
include/
* opcode/aarch64.h:
(aarch64_opnd): Add AARCH64_OPND_Va, AARCH64_OPND_MASK,
AARCH64_OPND_IMM_2, AARCH64_OPND_ADDR_OFFSET
and AARCH64_OPND_SM3_IMM2.
(aarch64_insn_class): Add cryptosm3 and cryptosm4.
(arch64_feature_set): Make uint64_t.
opcodes/
* aarch64-asm.h (ins_addr_offset): New.
* aarch64-asm.c (aarch64_ins_reglane): Add cryptosm3.
(aarch64_ins_addr_offset): New.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_addr_offset): New.
* aarch64-dis.c (aarch64_ext_reglane): Add cryptosm3.
(aarch64_ext_addr_offset): New.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (aarch64_field_kind): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
* aarch64-opc.c (fields): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
(operand_general_constraint_met_p): Add AARCH64_OPND_ADDR_OFFSET.
(aarch64_print_operand): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_MASK, AARCH64_OPND_IMM_2 and AARCH64_OPND_ADDR_OFFSET.
* aarch64-opc-2.c (Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2): New.
* aarch64-tbl.h
(aarch64_opcode_table): Add Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2.
2017-11-09 16:22:30 +01:00
|
|
|
|
Y(SIMD_REG, regno, "Va", 0, F(FLD_Ra), "a SIMD vector register") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(SIMD_REG, regno, "Vd", 0, F(FLD_Rd), "a SIMD vector register") \
|
|
|
|
|
Y(SIMD_REG, regno, "Vn", 0, F(FLD_Rn), "a SIMD vector register") \
|
|
|
|
|
Y(SIMD_REG, regno, "Vm", 0, F(FLD_Rm), "a SIMD vector register") \
|
|
|
|
|
Y(FP_REG, regno, "VdD1", 0, F(FLD_Rd), \
|
|
|
|
|
"the top half of a 128-bit FP/SIMD register") \
|
|
|
|
|
Y(FP_REG, regno, "VnD1", 0, F(FLD_Rn), \
|
|
|
|
|
"the top half of a 128-bit FP/SIMD register") \
|
|
|
|
|
Y(SIMD_ELEMENT, reglane, "Ed", 0, F(FLD_Rd), \
|
|
|
|
|
"a SIMD vector element") \
|
|
|
|
|
Y(SIMD_ELEMENT, reglane, "En", 0, F(FLD_Rn), \
|
|
|
|
|
"a SIMD vector element") \
|
|
|
|
|
Y(SIMD_ELEMENT, reglane, "Em", 0, F(FLD_Rm), \
|
|
|
|
|
"a SIMD vector element") \
|
Fix AArch64 encodings for by element instructions.
Some instructions in Armv8-a place a limitation on FP16 registers that can be
used as the register from which to select an element from.
e.g. fmla restricts Rm to 4 bits when using an FP16 register. This restriction
does not apply for all instructions, e.g. fcmla does not have this restriction
as it gets an extra bit from the M field.
Unfortunately, this restriction to S_H was added for all _Em operands before,
meaning for a large number of instructions you couldn't use the full register
file.
This fixes the issue by introducing a new operand _Em16 which applies this
restriction only when paired with S_H and leaves the _Em and the other
qualifiers for _Em16 unbounded (i.e. using the full 5 bit range).
Also the patch updates all instructions that should be affected by this.
opcodes/
PR binutils/23192
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
* aarch64-dis.c (aarch64_ext_reglane): Add AARCH64_OPND_Em16 constraint.
* aarch64-opc.c (operand_general_constraint_met_p,
aarch64_print_operand): Likewise.
* aarch64-tbl.h (aarch64_opcode_table): Change Em to Em16 for smlal,
smlal2, fmla, fmls, fmul, fmulx, sqrdmlah, sqrdlsh, fmlal, fmlsl,
fmlal2, fmlsl2.
(AARCH64_OPERANDS): Add Em2.
gas/
PR binutils/23192
* config/tc-aarch64.c (process_omitted_operand, parse_operands): Add
AARCH64_OPND_Em16
* testsuite/gas/aarch64/advsimd-armv8_3.s: Expand tests to cover upper
16 registers.
* testsuite/gas/aarch64/advsimd-armv8_3.d: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.s: Likewise.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
* testsuite/gas/aarch64/sve.d: Likewise.
include/
PR binutils/23192
*opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_Em16.
2018-06-29 13:12:27 +02:00
|
|
|
|
Y(SIMD_ELEMENT, reglane, "Em16", 0, F(FLD_Rm), \
|
|
|
|
|
"a SIMD vector element limited to V0-V15") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(SIMD_REGLIST, reglist, "LVn", 0, F(FLD_Rn), \
|
|
|
|
|
"a SIMD vector register list") \
|
|
|
|
|
Y(SIMD_REGLIST, ldst_reglist, "LVt", 0, F(), \
|
|
|
|
|
"a SIMD vector register list") \
|
|
|
|
|
Y(SIMD_REGLIST, ldst_reglist_r, "LVt_AL", 0, F(), \
|
|
|
|
|
"a SIMD vector register list") \
|
|
|
|
|
Y(SIMD_REGLIST, ldst_elemlist, "LEt", 0, F(), \
|
|
|
|
|
"a SIMD vector element list") \
|
2016-12-13 13:37:18 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "CRn", 0, F(FLD_CRn), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
"a 4-bit opcode field named for historical reasons C0 - C15") \
|
2016-12-13 13:37:18 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "CRm", 0, F(FLD_CRm), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
"a 4-bit opcode field named for historical reasons C0 - C15") \
|
|
|
|
|
Y(IMMEDIATE, imm, "IDX", 0, F(FLD_imm4), \
|
|
|
|
|
"an immediate as the index of the least significant byte") \
|
Adds the new Fields and Operand types for the new instructions in Armv8.4-a.
gas/
* config/tc-aarch64.c (process_omitted_operand):
Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2
and AARCH64_OPND_IMM_2.
(parse_operands): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_IMM_2, AARCH64_OPND_MASK
and AARCH64_OPND_ADDR_OFFSET.
include/
* opcode/aarch64.h:
(aarch64_opnd): Add AARCH64_OPND_Va, AARCH64_OPND_MASK,
AARCH64_OPND_IMM_2, AARCH64_OPND_ADDR_OFFSET
and AARCH64_OPND_SM3_IMM2.
(aarch64_insn_class): Add cryptosm3 and cryptosm4.
(arch64_feature_set): Make uint64_t.
opcodes/
* aarch64-asm.h (ins_addr_offset): New.
* aarch64-asm.c (aarch64_ins_reglane): Add cryptosm3.
(aarch64_ins_addr_offset): New.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_addr_offset): New.
* aarch64-dis.c (aarch64_ext_reglane): Add cryptosm3.
(aarch64_ext_addr_offset): New.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (aarch64_field_kind): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
* aarch64-opc.c (fields): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
(operand_general_constraint_met_p): Add AARCH64_OPND_ADDR_OFFSET.
(aarch64_print_operand): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_MASK, AARCH64_OPND_IMM_2 and AARCH64_OPND_ADDR_OFFSET.
* aarch64-opc-2.c (Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2): New.
* aarch64-tbl.h
(aarch64_opcode_table): Add Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2.
2017-11-09 16:22:30 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "MASK", 0, F(FLD_imm4_2), \
|
|
|
|
|
"an immediate as the index of the least significant byte") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(IMMEDIATE, advsimd_imm_shift, "IMM_VLSL", 0, F(), \
|
|
|
|
|
"a left shift amount for an AdvSIMD register") \
|
|
|
|
|
Y(IMMEDIATE, advsimd_imm_shift, "IMM_VLSR", 0, F(), \
|
|
|
|
|
"a right shift amount for an AdvSIMD register") \
|
|
|
|
|
Y(IMMEDIATE, advsimd_imm_modified, "SIMD_IMM", 0, F(), \
|
|
|
|
|
"an immediate") \
|
|
|
|
|
Y(IMMEDIATE, advsimd_imm_modified, "SIMD_IMM_SFT", 0, F(), \
|
|
|
|
|
"an 8-bit unsigned immediate with optional shift") \
|
|
|
|
|
Y(IMMEDIATE, advsimd_imm_modified, "SIMD_FPIMM", 0, F(), \
|
|
|
|
|
"an 8-bit floating-point constant") \
|
|
|
|
|
X(IMMEDIATE, 0, ext_shll_imm, "SHLL_IMM", 0, F(), \
|
|
|
|
|
"an immediate shift amount of 8, 16 or 32") \
|
|
|
|
|
X(IMMEDIATE, 0, 0, "IMM0", 0, F(), "0") \
|
|
|
|
|
X(IMMEDIATE, 0, 0, "FPIMM0", 0, F(), "0.0") \
|
2016-09-21 17:51:24 +02:00
|
|
|
|
Y(IMMEDIATE, fpimm, "FPIMM", 0, F(FLD_imm8), \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
"an 8-bit floating-point constant") \
|
|
|
|
|
Y(IMMEDIATE, imm, "IMMR", 0, F(FLD_immr), \
|
|
|
|
|
"the right rotate amount") \
|
|
|
|
|
Y(IMMEDIATE, imm, "IMMS", 0, F(FLD_imm6), \
|
|
|
|
|
"the leftmost bit number to be moved from the source") \
|
|
|
|
|
Y(IMMEDIATE, imm, "WIDTH", 0, F(FLD_imm6), \
|
|
|
|
|
"the width of the bit-field") \
|
Adds the new Fields and Operand types for the new instructions in Armv8.4-a.
gas/
* config/tc-aarch64.c (process_omitted_operand):
Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2
and AARCH64_OPND_IMM_2.
(parse_operands): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_IMM_2, AARCH64_OPND_MASK
and AARCH64_OPND_ADDR_OFFSET.
include/
* opcode/aarch64.h:
(aarch64_opnd): Add AARCH64_OPND_Va, AARCH64_OPND_MASK,
AARCH64_OPND_IMM_2, AARCH64_OPND_ADDR_OFFSET
and AARCH64_OPND_SM3_IMM2.
(aarch64_insn_class): Add cryptosm3 and cryptosm4.
(arch64_feature_set): Make uint64_t.
opcodes/
* aarch64-asm.h (ins_addr_offset): New.
* aarch64-asm.c (aarch64_ins_reglane): Add cryptosm3.
(aarch64_ins_addr_offset): New.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_addr_offset): New.
* aarch64-dis.c (aarch64_ext_reglane): Add cryptosm3.
(aarch64_ext_addr_offset): New.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (aarch64_field_kind): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
* aarch64-opc.c (fields): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
(operand_general_constraint_met_p): Add AARCH64_OPND_ADDR_OFFSET.
(aarch64_print_operand): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_MASK, AARCH64_OPND_IMM_2 and AARCH64_OPND_ADDR_OFFSET.
* aarch64-opc-2.c (Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2): New.
* aarch64-tbl.h
(aarch64_opcode_table): Add Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2.
2017-11-09 16:22:30 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "IMM", 0, F(FLD_imm6), "an immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "IMM_2", 0, F(FLD_imm6_2), "an immediate") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "UIMM3_OP1", 0, F(FLD_op1), \
|
|
|
|
|
"a 3-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "UIMM3_OP2", 0, F(FLD_op2), \
|
|
|
|
|
"a 3-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "UIMM4", 0, F(FLD_CRm), \
|
|
|
|
|
"a 4-bit unsigned immediate") \
|
[BINUTILS, AARCH64, 2/8] Add Tag generation instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag generation instructions from
MTE. These are the following instructions added in this patch:
- IRG <Xd|SP>, <Xn|SP>{, Xm}
- ADDG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- SUBG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- GMI <Xd>, <Xn|SP>, <Xm>
where
<Xd|SP> : Is the 64-bit destination GPR or Stack pointer.
<Xn|SP> : Is the 64-bit source GPR or Stack pointer.
<uimm6> : Is the unsigned immediate, a multiple of 16
in the range 0 to 1008.
<uimm4> : Is the unsigned immediate, in the range 0 to 15.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10 as new enums.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (aarch64_field_kind): New FLD_imm4_3.
(OPD_F_SHIFT_BY_4, operand_need_shift_by_four): New.
* aarch64-opc.c (fields): Add entry for imm4_3.
(operand_general_constraint_met_p): Add cases for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_ADDG): New.
(aarch64_opcode_table): Add addg, subg, irg and gmi.
(AARCH64_OPERANDS): Define UIMM4_ADDG and UIMM10.
* aarch64-asm.c (aarch64_ins_imm): Add case for
operand_need_shift_by_four.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: New.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
* testsuite/gas/aarch64/illegal-memtag.d: Likewise.
2018-11-12 13:52:55 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "UIMM4_ADDG", 0, F(FLD_imm4_3), \
|
|
|
|
|
"a 4-bit unsigned Logical Address Tag modifier") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "UIMM7", 0, F(FLD_CRm, FLD_op2), \
|
|
|
|
|
"a 7-bit unsigned immediate") \
|
[BINUTILS, AARCH64, 2/8] Add Tag generation instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag generation instructions from
MTE. These are the following instructions added in this patch:
- IRG <Xd|SP>, <Xn|SP>{, Xm}
- ADDG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- SUBG <Xd|SP>, <Xn|SP>, #<uimm1>. #<uimm2>
- GMI <Xd>, <Xn|SP>, <Xm>
where
<Xd|SP> : Is the 64-bit destination GPR or Stack pointer.
<Xn|SP> : Is the 64-bit source GPR or Stack pointer.
<uimm6> : Is the unsigned immediate, a multiple of 16
in the range 0 to 1008.
<uimm4> : Is the unsigned immediate, in the range 0 to 15.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10 as new enums.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (aarch64_field_kind): New FLD_imm4_3.
(OPD_F_SHIFT_BY_4, operand_need_shift_by_four): New.
* aarch64-opc.c (fields): Add entry for imm4_3.
(operand_general_constraint_met_p): Add cases for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_ADDG): New.
(aarch64_opcode_table): Add addg, subg, irg and gmi.
(AARCH64_OPERANDS): Define UIMM4_ADDG and UIMM10.
* aarch64-asm.c (aarch64_ins_imm): Add case for
operand_need_shift_by_four.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_UIMM4_ADDG and AARCH64_OPND_UIMM10.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: New.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
* testsuite/gas/aarch64/illegal-memtag.d: Likewise.
2018-11-12 13:52:55 +01:00
|
|
|
|
Y(IMMEDIATE, imm, "UIMM10", OPD_F_SHIFT_BY_4, F(FLD_immr), \
|
|
|
|
|
"a 10-bit unsigned multiple of 16") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "BIT_NUM", 0, F(FLD_b5, FLD_b40), \
|
|
|
|
|
"the bit number to be tested") \
|
|
|
|
|
Y(IMMEDIATE, imm, "EXCEPTION", 0, F(FLD_imm16), \
|
|
|
|
|
"a 16-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "CCMP_IMM", 0, F(FLD_imm5), \
|
|
|
|
|
"a 5-bit unsigned immediate") \
|
[AArch64][SVE 27/32] Add SVE integer immediate operands
This patch adds the new SVE integer immediate operands. There are
three kinds:
- simple signed and unsigned ranges, but with new widths and positions.
- 13-bit logical immediates. These have the same form as in base AArch64,
but at a different bit position.
In the case of the "MOV Zn.<T>, #<limm>" alias of DUPM, the logical
immediate <limm> is not allowed to be a valid DUP immediate, since DUP
is preferred over DUPM for constants that both instructions can handle.
- a new 9-bit arithmetic immediate, of the form "<imm8>{, LSL #8}".
In some contexts the operand is signed and in others it's unsigned.
As an extension, we allow shifted immediates to be written as a single
integer, e.g. "#256" is equivalent to "#1, LSL #8". We also use the
shiftless form as the preferred disassembly, except for the special
case of "#0, LSL #8" (a redundant encoding of 0).
include/
* opcode/aarch64.h (AARCH64_OPND_SIMM5): New aarch64_opnd.
(AARCH64_OPND_SVE_AIMM, AARCH64_OPND_SVE_ASIMM)
(AARCH64_OPND_SVE_INV_LIMM, AARCH64_OPND_SVE_LIMM)
(AARCH64_OPND_SVE_LIMM_MOV, AARCH64_OPND_SVE_SHLIMM_PRED)
(AARCH64_OPND_SVE_SHLIMM_UNPRED, AARCH64_OPND_SVE_SHRIMM_PRED)
(AARCH64_OPND_SVE_SHRIMM_UNPRED, AARCH64_OPND_SVE_SIMM5)
(AARCH64_OPND_SVE_SIMM5B, AARCH64_OPND_SVE_SIMM6)
(AARCH64_OPND_SVE_SIMM8, AARCH64_OPND_SVE_UIMM3)
(AARCH64_OPND_SVE_UIMM7, AARCH64_OPND_SVE_UIMM8)
(AARCH64_OPND_SVE_UIMM8_53): Likewise.
(aarch64_sve_dupm_mov_immediate_p): Declare.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
integer immediate operands.
* aarch64-opc.h (FLD_SVE_immN, FLD_SVE_imm3, FLD_SVE_imm5)
(FLD_SVE_imm5b, FLD_SVE_imm7, FLD_SVE_imm8, FLD_SVE_imm9)
(FLD_SVE_immr, FLD_SVE_imms, FLD_SVE_tszh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries.
(operand_general_constraint_met_p): Handle the new SVE integer
immediate operands.
(aarch64_print_operand): Likewise.
(aarch64_sve_dupm_mov_immediate_p): New function.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_inv_limm, ins_sve_aimm, ins_sve_asimm)
(ins_sve_limm_mov, ins_sve_shlimm, ins_sve_shrimm): New inserters.
* aarch64-asm.c (aarch64_ins_limm_1): New function, split out from...
(aarch64_ins_limm): ...here.
(aarch64_ins_inv_limm): New function.
(aarch64_ins_sve_aimm): Likewise.
(aarch64_ins_sve_asimm): Likewise.
(aarch64_ins_sve_limm_mov): Likewise.
(aarch64_ins_sve_shlimm): Likewise.
(aarch64_ins_sve_shrimm): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_inv_limm, ext_sve_aimm, ext_sve_asimm)
(ext_sve_limm_mov, ext_sve_shlimm, ext_sve_shrimm): New extractors.
* aarch64-dis.c (decode_limm): New function, split out from...
(aarch64_ext_limm): ...here.
(aarch64_ext_inv_limm): New function.
(decode_sve_aimm): Likewise.
(aarch64_ext_sve_aimm): Likewise.
(aarch64_ext_sve_asimm): Likewise.
(aarch64_ext_sve_limm_mov): Likewise.
(aarch64_top_bit): Likewise.
(aarch64_ext_sve_shlimm): Likewise.
(aarch64_ext_sve_shrimm): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (parse_operands): Handle the new SVE integer
immediate operands.
2016-09-21 17:56:57 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "SIMM5", OPD_F_SEXT, F(FLD_imm5), \
|
|
|
|
|
"a 5-bit signed immediate") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "NZCV", 0, F(FLD_nzcv), \
|
|
|
|
|
"a flag bit specifier giving an alternative value for each flag") \
|
|
|
|
|
Y(IMMEDIATE, limm, "LIMM", 0, F(FLD_N,FLD_immr,FLD_imms), \
|
|
|
|
|
"Logical immediate") \
|
|
|
|
|
Y(IMMEDIATE, aimm, "AIMM", 0, F(FLD_shift,FLD_imm12), \
|
|
|
|
|
"a 12-bit unsigned immediate with optional left shift of 12 bits")\
|
|
|
|
|
Y(IMMEDIATE, imm_half, "HALF", 0, F(FLD_imm16), \
|
|
|
|
|
"a 16-bit immediate with optional left shift") \
|
|
|
|
|
Y(IMMEDIATE, fbits, "FBITS", 0, F(FLD_scale), \
|
|
|
|
|
"the number of bits after the binary point in the fixed-point value")\
|
|
|
|
|
X(IMMEDIATE, 0, 0, "IMM_MOV", 0, F(), "an immediate") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(IMMEDIATE, imm_rotate2, "IMM_ROT1", 0, F(FLD_rotate1), \
|
[AArch64] Add ARMv8.3 FCMLA and FCADD instructions
Add support for FCMLA and FCADD complex arithmetic SIMD instructions.
FCMLA has an indexed element variant where the index range has to be
treated specially because a complex number takes two elements and the
indexed vector size depends on the other operands.
These complex number SIMD instructions are part of ARMv8.3
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
include/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* opcode/aarch64.h (enum aarch64_opnd): Add AARCH64_OPND_IMM_ROT1,
AARCH64_OPND_IMM_ROT2, AARCH64_OPND_IMM_ROT3.
(enum aarch64_op): Add OP_FCMLA_ELEM.
opcodes/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (QL_V3SAMEHSD_ROT, QL_ELEMENT_ROT): Define.
(aarch64_feature_simd_v8_3, SIMD_V8_3): Define.
(aarch64_opcode_table): Add fcmla and fcadd.
(AARCH64_OPERANDS): Add IMM_ROT{1,2,3}.
* aarch64-asm.h (aarch64_ins_imm_rotate): Declare.
* aarch64-asm.c (aarch64_ins_imm_rotate): Define.
* aarch64-dis.h (aarch64_ext_imm_rotate): Declare.
* aarch64-dis.c (aarch64_ext_imm_rotate): Define.
* aarch64-opc.h (enum aarch64_field_kind): Add FLD_rotate{1,2,3}.
* aarch64-opc.c (fields): Add FLD_rotate{1,2,3}.
(operand_general_constraint_met_p): Rotate and index range check.
(aarch64_print_operand): Handle rotate operand.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
gas/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* config/tc-aarch64.c (parse_operands): Handle AARCH64_OPND_IMM_ROT*.
* testsuite/gas/aarch64/advsimd-armv8_3.d: New.
* testsuite/gas/aarch64/advsimd-armv8_3.s: New.
* testsuite/gas/aarch64/illegal-fcmla.s: New.
* testsuite/gas/aarch64/illegal-fcmla.l: New.
* testsuite/gas/aarch64/illegal-fcmla.d: New.
2016-11-18 11:02:16 +01:00
|
|
|
|
"a 2-bit rotation specifier for complex arithmetic operations") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(IMMEDIATE, imm_rotate2, "IMM_ROT2", 0, F(FLD_rotate2), \
|
[AArch64] Add ARMv8.3 FCMLA and FCADD instructions
Add support for FCMLA and FCADD complex arithmetic SIMD instructions.
FCMLA has an indexed element variant where the index range has to be
treated specially because a complex number takes two elements and the
indexed vector size depends on the other operands.
These complex number SIMD instructions are part of ARMv8.3
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
include/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* opcode/aarch64.h (enum aarch64_opnd): Add AARCH64_OPND_IMM_ROT1,
AARCH64_OPND_IMM_ROT2, AARCH64_OPND_IMM_ROT3.
(enum aarch64_op): Add OP_FCMLA_ELEM.
opcodes/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (QL_V3SAMEHSD_ROT, QL_ELEMENT_ROT): Define.
(aarch64_feature_simd_v8_3, SIMD_V8_3): Define.
(aarch64_opcode_table): Add fcmla and fcadd.
(AARCH64_OPERANDS): Add IMM_ROT{1,2,3}.
* aarch64-asm.h (aarch64_ins_imm_rotate): Declare.
* aarch64-asm.c (aarch64_ins_imm_rotate): Define.
* aarch64-dis.h (aarch64_ext_imm_rotate): Declare.
* aarch64-dis.c (aarch64_ext_imm_rotate): Define.
* aarch64-opc.h (enum aarch64_field_kind): Add FLD_rotate{1,2,3}.
* aarch64-opc.c (fields): Add FLD_rotate{1,2,3}.
(operand_general_constraint_met_p): Rotate and index range check.
(aarch64_print_operand): Handle rotate operand.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
gas/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* config/tc-aarch64.c (parse_operands): Handle AARCH64_OPND_IMM_ROT*.
* testsuite/gas/aarch64/advsimd-armv8_3.d: New.
* testsuite/gas/aarch64/advsimd-armv8_3.s: New.
* testsuite/gas/aarch64/illegal-fcmla.s: New.
* testsuite/gas/aarch64/illegal-fcmla.l: New.
* testsuite/gas/aarch64/illegal-fcmla.d: New.
2016-11-18 11:02:16 +01:00
|
|
|
|
"a 2-bit rotation specifier for complex arithmetic operations") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(IMMEDIATE, imm_rotate1, "IMM_ROT3", 0, F(FLD_rotate3), \
|
[AArch64] Add ARMv8.3 FCMLA and FCADD instructions
Add support for FCMLA and FCADD complex arithmetic SIMD instructions.
FCMLA has an indexed element variant where the index range has to be
treated specially because a complex number takes two elements and the
indexed vector size depends on the other operands.
These complex number SIMD instructions are part of ARMv8.3
https://community.arm.com/groups/processors/blog/2016/10/27/armv8-a-architecture-2016-additions
include/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* opcode/aarch64.h (enum aarch64_opnd): Add AARCH64_OPND_IMM_ROT1,
AARCH64_OPND_IMM_ROT2, AARCH64_OPND_IMM_ROT3.
(enum aarch64_op): Add OP_FCMLA_ELEM.
opcodes/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* aarch64-tbl.h (QL_V3SAMEHSD_ROT, QL_ELEMENT_ROT): Define.
(aarch64_feature_simd_v8_3, SIMD_V8_3): Define.
(aarch64_opcode_table): Add fcmla and fcadd.
(AARCH64_OPERANDS): Add IMM_ROT{1,2,3}.
* aarch64-asm.h (aarch64_ins_imm_rotate): Declare.
* aarch64-asm.c (aarch64_ins_imm_rotate): Define.
* aarch64-dis.h (aarch64_ext_imm_rotate): Declare.
* aarch64-dis.c (aarch64_ext_imm_rotate): Define.
* aarch64-opc.h (enum aarch64_field_kind): Add FLD_rotate{1,2,3}.
* aarch64-opc.c (fields): Add FLD_rotate{1,2,3}.
(operand_general_constraint_met_p): Rotate and index range check.
(aarch64_print_operand): Handle rotate operand.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Likewise.
* aarch64-opc-2.c: Likewise.
gas/
2016-11-18 Szabolcs Nagy <szabolcs.nagy@arm.com>
* config/tc-aarch64.c (parse_operands): Handle AARCH64_OPND_IMM_ROT*.
* testsuite/gas/aarch64/advsimd-armv8_3.d: New.
* testsuite/gas/aarch64/advsimd-armv8_3.s: New.
* testsuite/gas/aarch64/illegal-fcmla.s: New.
* testsuite/gas/aarch64/illegal-fcmla.l: New.
* testsuite/gas/aarch64/illegal-fcmla.d: New.
2016-11-18 11:02:16 +01:00
|
|
|
|
"a 1-bit rotation specifier for complex arithmetic operations") \
|
2013-11-05 21:50:18 +01:00
|
|
|
|
Y(COND, cond, "COND", 0, F(), "a condition") \
|
|
|
|
|
Y(COND, cond, "COND1", 0, F(), \
|
|
|
|
|
"one of the standard conditions, excluding AL and NV.") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
X(ADDRESS, 0, ext_imm, "ADDR_ADRP", OPD_F_SEXT, F(FLD_immhi, FLD_immlo),\
|
|
|
|
|
"21-bit PC-relative address of a 4KB page") \
|
|
|
|
|
Y(ADDRESS, imm, "ADDR_PCREL14", OPD_F_SEXT | OPD_F_SHIFT_BY_2, \
|
|
|
|
|
F(FLD_imm14), "14-bit PC-relative address") \
|
|
|
|
|
Y(ADDRESS, imm, "ADDR_PCREL19", OPD_F_SEXT | OPD_F_SHIFT_BY_2, \
|
|
|
|
|
F(FLD_imm19), "19-bit PC-relative address") \
|
|
|
|
|
Y(ADDRESS, imm, "ADDR_PCREL21", OPD_F_SEXT, F(FLD_immhi,FLD_immlo), \
|
|
|
|
|
"21-bit PC-relative address") \
|
|
|
|
|
Y(ADDRESS, imm, "ADDR_PCREL26", OPD_F_SEXT | OPD_F_SHIFT_BY_2, \
|
|
|
|
|
F(FLD_imm26), "26-bit PC-relative address") \
|
|
|
|
|
Y(ADDRESS, addr_simple, "ADDR_SIMPLE", 0, F(), \
|
|
|
|
|
"an address with base register (no offset)") \
|
|
|
|
|
Y(ADDRESS, addr_regoff, "ADDR_REGOFF", 0, F(), \
|
|
|
|
|
"an address with register offset") \
|
|
|
|
|
Y(ADDRESS, addr_simm, "ADDR_SIMM7", 0, F(FLD_imm7,FLD_index2), \
|
|
|
|
|
"an address with 7-bit signed immediate offset") \
|
|
|
|
|
Y(ADDRESS, addr_simm, "ADDR_SIMM9", 0, F(FLD_imm9,FLD_index), \
|
|
|
|
|
"an address with 9-bit signed immediate offset") \
|
|
|
|
|
Y(ADDRESS, addr_simm, "ADDR_SIMM9_2", 0, F(FLD_imm9,FLD_index), \
|
|
|
|
|
"an address with 9-bit negative or unaligned immediate offset") \
|
2016-11-18 10:49:06 +01:00
|
|
|
|
Y(ADDRESS, addr_simm10, "ADDR_SIMM10", 0, F(FLD_Rn,FLD_S_imm10,FLD_imm9,FLD_index),\
|
|
|
|
|
"an address with 10-bit scaled, signed immediate offset") \
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
Y(ADDRESS, addr_simm, "ADDR_SIMM11", 0, F(FLD_imm7,FLD_index2),\
|
|
|
|
|
"an address with 11-bit signed immediate (multiple of 16) offset")\
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(ADDRESS, addr_uimm12, "ADDR_UIMM12", 0, F(FLD_Rn,FLD_imm12), \
|
|
|
|
|
"an address with scaled, unsigned immediate offset") \
|
[BINUTILS, AARCH64, 4/8] Add Tag setting instructions in Memory Tagging Extension
This patch is part of the patch series to add support for ARMv8.5-A
Memory Tagging Extensions which is an optional extension to
ARMv8.5-A and is enabled using the +memtag command line option.
This patch add support to the Tag setting instructions from
MTE which consists of the following instructions:
- STG [<Xn|SP>, #<simm>]
- STG [<Xn|SP>, #<simm>]!
- STG [<Xn|SP>], #<simm>
- STZG [<Xn|SP>, #<simm>]
- STZG [<Xn|SP>, #<simm>]!
- STZG [<Xn|SP>], #<simm>
- ST2G [<Xn|SP>, #<simm>]
- ST2G [<Xn|SP>, #<simm>]!
- ST2G [<Xn|SP>], #<simm>
- STZ2G [<Xn|SP>, #<simm>]
- STZ2G [<Xn|SP>, #<simm>]!
- STZ2G [<Xn|SP>], #<simm>
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]
- STGP <Xt>, <Xt2>, [<Xn|SP>, #<imm>]!
- STGP <Xt>, <Xt2>, [<Xn|SP>], #<imm>
where
<Xn|SP> : Is the 64-bit GPR or Stack pointer.
<simm> : Is the optional signed immediate offset, a multiple of 16
in the range -4096 to 4080, defaulting to 0.
*** include/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (aarch64_opnd): Add AARCH64_OPND_ADDR_SIMM11
and AARCH64_OPND_ADDR_SIMM13.
(aarch64_opnd_qualifier): Add new AARCH64_OPND_QLF_imm_tag.
*** opcodes/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.c (aarch64_opnd_qualifiers): Add new data
for AARCH64_OPND_QLF_imm_tag.
(operand_general_constraint_met_p): Add case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(aarch64_print_operand): Likewise.
* aarch64-tbl.h (QL_LDST_AT, QL_STGP): New.
(aarch64_opcode_table): Add stg, stzg, st2g, stz2g and stgp
for both offset and pre/post indexed versions.
(AARCH64_OPERANDS): Define ADDR_SIMM11 and ADDR_SIMM13.
* aarch64-asm-2.c: Regenerated.
* aarch64-dis-2.c: Regenerated.
* aarch64-opc-2.c: Regenerated.
*** gas/ChangeLog ***
2018-11-12 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_operands): Add switch case for
AARCH64_OPND_ADDR_SIMM11 and AARCH64_OPND_ADDR_SIMM13.
(fix_insn): Likewise.
(warn_unpredictable_ldst): Exempt STGP.
* testsuite/gas/aarch64/armv8_5-a-memtag.s: Add tests for stg, st2g,
stzg, stz2g and stgp.
* testsuite/gas/aarch64/armv8_5-a-memtag.d: Likewise.
* testsuite/gas/aarch64/illegal-memtag.s: Likewise.
* testsuite/gas/aarch64/illegal-memtag.l: Likewise.
2018-11-12 14:09:55 +01:00
|
|
|
|
Y(ADDRESS, addr_simm, "ADDR_SIMM13", 0, F(FLD_imm9,FLD_index),\
|
|
|
|
|
"an address with 13-bit signed immediate (multiple of 16) offset")\
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(ADDRESS, addr_simple, "SIMD_ADDR_SIMPLE", 0, F(), \
|
|
|
|
|
"an address with base register (no offset)") \
|
Adds the new Fields and Operand types for the new instructions in Armv8.4-a.
gas/
* config/tc-aarch64.c (process_omitted_operand):
Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2
and AARCH64_OPND_IMM_2.
(parse_operands): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_IMM_2, AARCH64_OPND_MASK
and AARCH64_OPND_ADDR_OFFSET.
include/
* opcode/aarch64.h:
(aarch64_opnd): Add AARCH64_OPND_Va, AARCH64_OPND_MASK,
AARCH64_OPND_IMM_2, AARCH64_OPND_ADDR_OFFSET
and AARCH64_OPND_SM3_IMM2.
(aarch64_insn_class): Add cryptosm3 and cryptosm4.
(arch64_feature_set): Make uint64_t.
opcodes/
* aarch64-asm.h (ins_addr_offset): New.
* aarch64-asm.c (aarch64_ins_reglane): Add cryptosm3.
(aarch64_ins_addr_offset): New.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_addr_offset): New.
* aarch64-dis.c (aarch64_ext_reglane): Add cryptosm3.
(aarch64_ext_addr_offset): New.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (aarch64_field_kind): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
* aarch64-opc.c (fields): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
(operand_general_constraint_met_p): Add AARCH64_OPND_ADDR_OFFSET.
(aarch64_print_operand): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_MASK, AARCH64_OPND_IMM_2 and AARCH64_OPND_ADDR_OFFSET.
* aarch64-opc-2.c (Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2): New.
* aarch64-tbl.h
(aarch64_opcode_table): Add Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2.
2017-11-09 16:22:30 +01:00
|
|
|
|
Y(ADDRESS, addr_offset, "ADDR_OFFSET", 0, F(FLD_Rn,FLD_imm9,FLD_index),\
|
|
|
|
|
"an address with an optional 8-bit signed immediate offset") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(ADDRESS, simd_addr_post, "SIMD_ADDR_POST", 0, F(), \
|
|
|
|
|
"a post-indexed address with immediate or register increment") \
|
|
|
|
|
Y(SYSTEM, sysreg, "SYSREG", 0, F(), "a system register") \
|
|
|
|
|
Y(SYSTEM, pstatefield, "PSTATEFIELD", 0, F(), \
|
|
|
|
|
"a PSTATE field name") \
|
|
|
|
|
Y(SYSTEM, sysins_op, "SYSREG_AT", 0, F(), \
|
|
|
|
|
"an address translation operation specifier") \
|
|
|
|
|
Y(SYSTEM, sysins_op, "SYSREG_DC", 0, F(), \
|
|
|
|
|
"a data cache maintenance operation specifier") \
|
|
|
|
|
Y(SYSTEM, sysins_op, "SYSREG_IC", 0, F(), \
|
2015-12-11 11:11:27 +01:00
|
|
|
|
"an instruction cache maintenance operation specifier") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(SYSTEM, sysins_op, "SYSREG_TLBI", 0, F(), \
|
|
|
|
|
"a TBL invalidation operation specifier") \
|
2018-09-26 11:52:51 +02:00
|
|
|
|
Y(SYSTEM, sysins_op, "SYSREG_SR", 0, F(), \
|
|
|
|
|
"a Speculation Restriction option name (RCTX)") \
|
2012-08-13 16:52:54 +02:00
|
|
|
|
Y(SYSTEM, barrier, "BARRIER", 0, F(), \
|
|
|
|
|
"a barrier option name") \
|
|
|
|
|
Y(SYSTEM, barrier, "BARRIER_ISB", 0, F(), \
|
|
|
|
|
"the ISB option name SY or an optional 4-bit unsigned immediate") \
|
|
|
|
|
Y(SYSTEM, prfop, "PRFOP", 0, F(), \
|
2015-12-11 11:22:40 +01:00
|
|
|
|
"a prefetch operation specifier") \
|
[AArch64][SVE 25/32] Add support for SVE addressing modes
This patch adds most of the new SVE addressing modes and associated
operands. A follow-on patch adds MUL VL, since handling it separately
makes the changes easier to read.
The patch also introduces a new "operand-dependent data" field to the
operand flags, based closely on the existing one for opcode flags.
For SVE this new field needs only 2 bits, but it could be widened
in future if necessary.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_U6): New aarch64_opnd.
(AARCH64_OPND_SVE_ADDR_RI_U6x2, AARCH64_OPND_SVE_ADDR_RI_U6x4)
(AARCH64_OPND_SVE_ADDR_RI_U6x8, AARCH64_OPND_SVE_ADDR_RR)
(AARCH64_OPND_SVE_ADDR_RR_LSL1, AARCH64_OPND_SVE_ADDR_RR_LSL2)
(AARCH64_OPND_SVE_ADDR_RR_LSL3, AARCH64_OPND_SVE_ADDR_RX)
(AARCH64_OPND_SVE_ADDR_RX_LSL1, AARCH64_OPND_SVE_ADDR_RX_LSL2)
(AARCH64_OPND_SVE_ADDR_RX_LSL3, AARCH64_OPND_SVE_ADDR_RZ)
(AARCH64_OPND_SVE_ADDR_RZ_LSL1, AARCH64_OPND_SVE_ADDR_RZ_LSL2)
(AARCH64_OPND_SVE_ADDR_RZ_LSL3, AARCH64_OPND_SVE_ADDR_RZ_XTW_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW_22, AARCH64_OPND_SVE_ADDR_RZ_XTW1_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW1_22, AARCH64_OPND_SVE_ADDR_RZ_XTW2_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW2_22, AARCH64_OPND_SVE_ADDR_RZ_XTW3_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW3_22, AARCH64_OPND_SVE_ADDR_ZI_U5)
(AARCH64_OPND_SVE_ADDR_ZI_U5x2, AARCH64_OPND_SVE_ADDR_ZI_U5x4)
(AARCH64_OPND_SVE_ADDR_ZI_U5x8, AARCH64_OPND_SVE_ADDR_ZZ_LSL)
(AARCH64_OPND_SVE_ADDR_ZZ_SXTW, AARCH64_OPND_SVE_ADDR_ZZ_UXTW):
Likewise.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
address operands.
* aarch64-opc.h (FLD_SVE_imm6, FLD_SVE_msz, FLD_SVE_xs_14)
(FLD_SVE_xs_22): New aarch64_field_kinds.
(OPD_F_OD_MASK, OPD_F_OD_LSB, OPD_F_NO_ZR): New flags.
(get_operand_specific_data): New function.
* aarch64-opc.c (fields): Add entries for FLD_SVE_imm6, FLD_SVE_msz,
FLD_SVE_xs_14 and FLD_SVE_xs_22.
(operand_general_constraint_met_p): Handle the new SVE address
operands.
(sve_reg): New array.
(get_addr_sve_reg_name): New function.
(aarch64_print_operand): Handle the new SVE address operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_addr_ri_u6, ins_sve_addr_rr_lsl)
(ins_sve_addr_rz_xtw, ins_sve_addr_zi_u5, ins_sve_addr_zz_lsl)
(ins_sve_addr_zz_sxtw, ins_sve_addr_zz_uxtw): New inserters.
* aarch64-asm.c (aarch64_ins_sve_addr_ri_u6): New function.
(aarch64_ins_sve_addr_rr_lsl): Likewise.
(aarch64_ins_sve_addr_rz_xtw): Likewise.
(aarch64_ins_sve_addr_zi_u5): Likewise.
(aarch64_ins_sve_addr_zz): Likewise.
(aarch64_ins_sve_addr_zz_lsl): Likewise.
(aarch64_ins_sve_addr_zz_sxtw): Likewise.
(aarch64_ins_sve_addr_zz_uxtw): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_u6, ext_sve_addr_rr_lsl)
(ext_sve_addr_rz_xtw, ext_sve_addr_zi_u5, ext_sve_addr_zz_lsl)
(ext_sve_addr_zz_sxtw, ext_sve_addr_zz_uxtw): New extractors.
* aarch64-dis.c (aarch64_ext_sve_add_reg_imm): New function.
(aarch64_ext_sve_addr_ri_u6): Likewise.
(aarch64_ext_sve_addr_rr_lsl): Likewise.
(aarch64_ext_sve_addr_rz_xtw): Likewise.
(aarch64_ext_sve_addr_zi_u5): Likewise.
(aarch64_ext_sve_addr_zz): Likewise.
(aarch64_ext_sve_addr_zz_lsl): Likewise.
(aarch64_ext_sve_addr_zz_sxtw): Likewise.
(aarch64_ext_sve_addr_zz_uxtw): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (REG_TYPE_SVE_BASE, REG_TYPE_SVE_OFFSET): New
register types.
(get_reg_expected_msg): Handle them.
(aarch64_addr_reg_parse): New function, split out from
aarch64_reg_parse_32_64. Handle Z registers too.
(aarch64_reg_parse_32_64): Call it.
(parse_address_main): Add base_qualifier, offset_qualifier,
base_type and offset_type parameters. Handle SVE base and offset
registers.
(parse_address): Update call to parse_address_main.
(parse_sve_address): New function.
(parse_operands): Parse the new SVE address operands.
2016-09-21 17:55:49 +02:00
|
|
|
|
Y(SYSTEM, hint, "BARRIER_PSB", 0, F (), \
|
[AArch64][SVE 21/32] Add Zn and Pn registers
This patch adds the Zn and Pn registers, and associated fields and
operands.
include/
* opcode/aarch64.h (AARCH64_OPND_CLASS_SVE_REG): New
aarch64_operand_class.
(AARCH64_OPND_CLASS_PRED_REG): Likewise.
(AARCH64_OPND_SVE_Pd, AARCH64_OPND_SVE_Pg3, AARCH64_OPND_SVE_Pg4_5)
(AARCH64_OPND_SVE_Pg4_10, AARCH64_OPND_SVE_Pg4_16)
(AARCH64_OPND_SVE_Pm, AARCH64_OPND_SVE_Pn, AARCH64_OPND_SVE_Pt)
(AARCH64_OPND_SVE_Za_5, AARCH64_OPND_SVE_Za_16, AARCH64_OPND_SVE_Zd)
(AARCH64_OPND_SVE_Zm_5, AARCH64_OPND_SVE_Zm_16, AARCH64_OPND_SVE_Zn)
(AARCH64_OPND_SVE_Zn_INDEX, AARCH64_OPND_SVE_ZnxN)
(AARCH64_OPND_SVE_Zt, AARCH64_OPND_SVE_ZtxN): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for new SVE operands.
* aarch64-opc.h (FLD_SVE_Pd, FLD_SVE_Pg3, FLD_SVE_Pg4_5)
(FLD_SVE_Pg4_10, FLD_SVE_Pg4_16, FLD_SVE_Pm, FLD_SVE_Pn, FLD_SVE_Pt)
(FLD_SVE_Za_5, FLD_SVE_Za_16, FLD_SVE_Zd, FLD_SVE_Zm_5, FLD_SVE_Zm_16)
(FLD_SVE_Zn, FLD_SVE_Zt, FLD_SVE_tzsh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries here.
(operand_general_constraint_met_p): Check that SVE register lists
have the correct length. Check the ranges of SVE index registers.
Check for cases where p8-p15 are used in 3-bit predicate fields.
(aarch64_print_operand): Handle the new SVE operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_index, ins_sve_reglist): New inserters.
* aarch64-asm.c (aarch64_ins_sve_index): New function.
(aarch64_ins_sve_reglist): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_index, ext_sve_reglist): New extractors.
* aarch64-dis.c (aarch64_ext_sve_index): New function.
(aarch64_ext_sve_reglist): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (NTA_HASVARWIDTH): New macro.
(AARCH64_REG_TYPES): Add ZN and PN.
(get_reg_expected_msg): Handle them.
(parse_vector_type_for_operand): Add a reg_type parameter.
Skip the width for Zn and Pn registers.
(parse_typed_reg): Extend vector handling to Zn and Pn. Update the
call to parse_vector_type_for_operand. Set HASVARTYPE for Zn and Pn,
expecting the width to be 0.
(parse_vector_reg_list): Restrict error about [BHSD]nn operands to
REG_TYPE_VN.
(vectype_to_qualifier): Use S_[BHSD] qualifiers for NTA_HASVARWIDTH.
(parse_operands): Handle the new Zn and Pn operands.
(REGSET16): New macro, split out from...
(REGSET31): ...here.
(reg_names): Add Zn and Pn entries.
2016-09-21 17:53:54 +02:00
|
|
|
|
"the PSB option name CSYNC") \
|
[PATCH, BINUTILS, AARCH64, 7/9] Add BTI instruction
This patch is part of the patch series to add support for ARMv8.5-A
extensions.
(https://developer.arm.com/products/architecture/cpu-architecture/a-profile/docs/ddi0596/a/a64-base-instructions-alphabetic-order/bti-branch-target-identification)
The Branch Target Identification instructions (BTI) are allocated to
existing HINT space, using HINT numbers 32, 34, 36, 38, such that
bits[7:6] of the instruction identify the compatibility of the BTI
instruction to different branches.
BTI {<targets>}
where <targets> one of the following, specifying which type of
indirection is allowed:
j : Can be a target of any BR Xn isntruction.
c : Can be a target of any BLR Xn and BR {X16|X17}.
jc: Can be a target of any free branch.
A BTI instruction without any <targets> is the strictest of all and
can not be a target of nay free branch.
*** include/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* opcode/aarch64.h (AARCH64_FEATURE_BTI): New.
(AARCH64_ARCH_V8_5): Add AARCH64_FEATURE_BTI by default.
(aarch64_opnd): Add AARCH64_OPND_BTI_TARGET.
(HINT_OPD_CSYNC, HINT_OPD_C, HINT_OPD_J): New macros to
define HINT #imm values.
(HINT_OPD_JC, HINT_OPD_NULL): Likewise.
*** opcodes/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* aarch64-opc.h (HINT_OPD_NOPRINT, HINT_ENCODE): New.
(HINT_FLAG, HINT_VALUE): New macros to encode NO_PRINT flag
with the hint immediate.
* aarch64-opc.c (aarch64_hint_options): New entries for
c, j, jc and default (with HINT_OPD_F_NOPRINT flag) for BTI.
(aarch64_print_operand): Add case for AARCH64_OPND_BTI_TARGET
while checking for HINT_OPD_F_NOPRINT flag.
* aarch64-dis.c (aarch64_ext_hint): Use new HINT_VALUE to
extract value.
* aarch64-tbl.h (aarch64_feature_bti, BTI, BTI_INSN): New.
(aarch64_opcode_table): Add entry for BTI.
(AARCH64_OPERANDS): Add new description for BTI targets.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc-2.c: Regenerate.
*** gas/ChangeLog ***
2018-10-09 Sudakshina Das <sudi.das@arm.com>
* config/tc-aarch64.c (parse_bti_operand): New.
(process_omitted_operand): Add case for AARCH64_OPND_BTI_TARGET.
(parse_operands): Likewise.
* testsuite/gas/aarch64/system.d: Update for BTI.
* testsuite/gas/aarch64/bti.s: New.
* testsuite/gas/aarch64/bti.d: New.
* testsuite/gas/aarch64/illegal-bti.d: New.
* testsuite/gas/aarch64/illegal-bti.l: New.
2018-09-26 12:00:49 +02:00
|
|
|
|
Y(SYSTEM, hint, "BTI", 0, F (), \
|
|
|
|
|
"BTI targets j/c/jc") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(ADDRESS, sve_addr_ri_s4, "SVE_ADDR_RI_S4x16", \
|
|
|
|
|
4 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 4-bit signed offset, multiplied by 16") \
|
[AArch64][SVE 26/32] Add SVE MUL VL addressing modes
This patch adds support for addresses of the form:
[<base>, #<offset>, MUL VL]
This involves adding a new AARCH64_MOD_MUL_VL modifier, which is
why I split it out from the other addressing modes.
For LD2, LD3 and LD4, the offset must be a multiple of the structure
size, so for LD3 the possible values are 0, 3, 6, .... The patch
therefore extends value_aligned_p to handle non-power-of-2 alignments.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4xVL): New aarch64_opnd.
(AARCH64_OPND_SVE_ADDR_RI_S4x2xVL, AARCH64_OPND_SVE_ADDR_RI_S4x3xVL)
(AARCH64_OPND_SVE_ADDR_RI_S4x4xVL, AARCH64_OPND_SVE_ADDR_RI_S6xVL)
(AARCH64_OPND_SVE_ADDR_RI_S9xVL): Likewise.
(AARCH64_MOD_MUL_VL): New aarch64_modifier_kind.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for new MUL VL
operands.
* aarch64-opc.c (aarch64_operand_modifiers): Initialize
the AARCH64_MOD_MUL_VL entry.
(value_aligned_p): Cope with non-power-of-two alignments.
(operand_general_constraint_met_p): Handle the new MUL VL addresses.
(print_immediate_offset_address): Likewise.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_addr_ri_s4xvl, ins_sve_addr_ri_s6xvl)
(ins_sve_addr_ri_s9xvl): New inserters.
* aarch64-asm.c (aarch64_ins_sve_addr_ri_s4xvl): New function.
(aarch64_ins_sve_addr_ri_s6xvl): Likewise.
(aarch64_ins_sve_addr_ri_s9xvl): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4xvl, ext_sve_addr_ri_s6xvl)
(ext_sve_addr_ri_s9xvl): New extractors.
* aarch64-dis.c (aarch64_ext_sve_addr_reg_mul_vl): New function.
(aarch64_ext_sve_addr_ri_s4xvl): Likewise.
(aarch64_ext_sve_addr_ri_s6xvl): Likewise.
(aarch64_ext_sve_addr_ri_s9xvl): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (SHIFTED_NONE, SHIFTED_MUL_VL): New
parse_shift_modes.
(parse_shift): Handle SHIFTED_MUL_VL.
(parse_address_main): Add an imm_shift_mode parameter.
(parse_address, parse_sve_address): Update accordingly.
(parse_operands): Handle MUL VL addressing modes.
2016-09-21 17:56:15 +02:00
|
|
|
|
Y(ADDRESS, sve_addr_ri_s4xvl, "SVE_ADDR_RI_S4xVL", \
|
|
|
|
|
0 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 4-bit signed offset, multiplied by VL") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_s4xvl, "SVE_ADDR_RI_S4x2xVL", \
|
|
|
|
|
1 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 4-bit signed offset, multiplied by 2*VL") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_s4xvl, "SVE_ADDR_RI_S4x3xVL", \
|
|
|
|
|
2 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 4-bit signed offset, multiplied by 3*VL") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_s4xvl, "SVE_ADDR_RI_S4x4xVL", \
|
|
|
|
|
3 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 4-bit signed offset, multiplied by 4*VL") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_s6xvl, "SVE_ADDR_RI_S6xVL", \
|
|
|
|
|
0 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 6-bit signed offset, multiplied by VL") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_s9xvl, "SVE_ADDR_RI_S9xVL", \
|
|
|
|
|
0 << OPD_F_OD_LSB, F(FLD_Rn), \
|
|
|
|
|
"an address with a 9-bit signed offset, multiplied by VL") \
|
[AArch64][SVE 25/32] Add support for SVE addressing modes
This patch adds most of the new SVE addressing modes and associated
operands. A follow-on patch adds MUL VL, since handling it separately
makes the changes easier to read.
The patch also introduces a new "operand-dependent data" field to the
operand flags, based closely on the existing one for opcode flags.
For SVE this new field needs only 2 bits, but it could be widened
in future if necessary.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_U6): New aarch64_opnd.
(AARCH64_OPND_SVE_ADDR_RI_U6x2, AARCH64_OPND_SVE_ADDR_RI_U6x4)
(AARCH64_OPND_SVE_ADDR_RI_U6x8, AARCH64_OPND_SVE_ADDR_RR)
(AARCH64_OPND_SVE_ADDR_RR_LSL1, AARCH64_OPND_SVE_ADDR_RR_LSL2)
(AARCH64_OPND_SVE_ADDR_RR_LSL3, AARCH64_OPND_SVE_ADDR_RX)
(AARCH64_OPND_SVE_ADDR_RX_LSL1, AARCH64_OPND_SVE_ADDR_RX_LSL2)
(AARCH64_OPND_SVE_ADDR_RX_LSL3, AARCH64_OPND_SVE_ADDR_RZ)
(AARCH64_OPND_SVE_ADDR_RZ_LSL1, AARCH64_OPND_SVE_ADDR_RZ_LSL2)
(AARCH64_OPND_SVE_ADDR_RZ_LSL3, AARCH64_OPND_SVE_ADDR_RZ_XTW_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW_22, AARCH64_OPND_SVE_ADDR_RZ_XTW1_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW1_22, AARCH64_OPND_SVE_ADDR_RZ_XTW2_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW2_22, AARCH64_OPND_SVE_ADDR_RZ_XTW3_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW3_22, AARCH64_OPND_SVE_ADDR_ZI_U5)
(AARCH64_OPND_SVE_ADDR_ZI_U5x2, AARCH64_OPND_SVE_ADDR_ZI_U5x4)
(AARCH64_OPND_SVE_ADDR_ZI_U5x8, AARCH64_OPND_SVE_ADDR_ZZ_LSL)
(AARCH64_OPND_SVE_ADDR_ZZ_SXTW, AARCH64_OPND_SVE_ADDR_ZZ_UXTW):
Likewise.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
address operands.
* aarch64-opc.h (FLD_SVE_imm6, FLD_SVE_msz, FLD_SVE_xs_14)
(FLD_SVE_xs_22): New aarch64_field_kinds.
(OPD_F_OD_MASK, OPD_F_OD_LSB, OPD_F_NO_ZR): New flags.
(get_operand_specific_data): New function.
* aarch64-opc.c (fields): Add entries for FLD_SVE_imm6, FLD_SVE_msz,
FLD_SVE_xs_14 and FLD_SVE_xs_22.
(operand_general_constraint_met_p): Handle the new SVE address
operands.
(sve_reg): New array.
(get_addr_sve_reg_name): New function.
(aarch64_print_operand): Handle the new SVE address operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_addr_ri_u6, ins_sve_addr_rr_lsl)
(ins_sve_addr_rz_xtw, ins_sve_addr_zi_u5, ins_sve_addr_zz_lsl)
(ins_sve_addr_zz_sxtw, ins_sve_addr_zz_uxtw): New inserters.
* aarch64-asm.c (aarch64_ins_sve_addr_ri_u6): New function.
(aarch64_ins_sve_addr_rr_lsl): Likewise.
(aarch64_ins_sve_addr_rz_xtw): Likewise.
(aarch64_ins_sve_addr_zi_u5): Likewise.
(aarch64_ins_sve_addr_zz): Likewise.
(aarch64_ins_sve_addr_zz_lsl): Likewise.
(aarch64_ins_sve_addr_zz_sxtw): Likewise.
(aarch64_ins_sve_addr_zz_uxtw): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_u6, ext_sve_addr_rr_lsl)
(ext_sve_addr_rz_xtw, ext_sve_addr_zi_u5, ext_sve_addr_zz_lsl)
(ext_sve_addr_zz_sxtw, ext_sve_addr_zz_uxtw): New extractors.
* aarch64-dis.c (aarch64_ext_sve_add_reg_imm): New function.
(aarch64_ext_sve_addr_ri_u6): Likewise.
(aarch64_ext_sve_addr_rr_lsl): Likewise.
(aarch64_ext_sve_addr_rz_xtw): Likewise.
(aarch64_ext_sve_addr_zi_u5): Likewise.
(aarch64_ext_sve_addr_zz): Likewise.
(aarch64_ext_sve_addr_zz_lsl): Likewise.
(aarch64_ext_sve_addr_zz_sxtw): Likewise.
(aarch64_ext_sve_addr_zz_uxtw): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (REG_TYPE_SVE_BASE, REG_TYPE_SVE_OFFSET): New
register types.
(get_reg_expected_msg): Handle them.
(aarch64_addr_reg_parse): New function, split out from
aarch64_reg_parse_32_64. Handle Z registers too.
(aarch64_reg_parse_32_64): Call it.
(parse_address_main): Add base_qualifier, offset_qualifier,
base_type and offset_type parameters. Handle SVE base and offset
registers.
(parse_address): Update call to parse_address_main.
(parse_sve_address): New function.
(parse_operands): Parse the new SVE address operands.
2016-09-21 17:55:49 +02:00
|
|
|
|
Y(ADDRESS, sve_addr_ri_u6, "SVE_ADDR_RI_U6", 0 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn), "an address with a 6-bit unsigned offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_u6, "SVE_ADDR_RI_U6x2", 1 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn), \
|
|
|
|
|
"an address with a 6-bit unsigned offset, multiplied by 2") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_u6, "SVE_ADDR_RI_U6x4", 2 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn), \
|
|
|
|
|
"an address with a 6-bit unsigned offset, multiplied by 4") \
|
|
|
|
|
Y(ADDRESS, sve_addr_ri_u6, "SVE_ADDR_RI_U6x8", 3 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn), \
|
|
|
|
|
"an address with a 6-bit unsigned offset, multiplied by 8") \
|
2018-03-28 10:44:45 +02:00
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_R", 0 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_Rm), "an address with an optional scalar register offset") \
|
[AArch64][SVE 25/32] Add support for SVE addressing modes
This patch adds most of the new SVE addressing modes and associated
operands. A follow-on patch adds MUL VL, since handling it separately
makes the changes easier to read.
The patch also introduces a new "operand-dependent data" field to the
operand flags, based closely on the existing one for opcode flags.
For SVE this new field needs only 2 bits, but it could be widened
in future if necessary.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_U6): New aarch64_opnd.
(AARCH64_OPND_SVE_ADDR_RI_U6x2, AARCH64_OPND_SVE_ADDR_RI_U6x4)
(AARCH64_OPND_SVE_ADDR_RI_U6x8, AARCH64_OPND_SVE_ADDR_RR)
(AARCH64_OPND_SVE_ADDR_RR_LSL1, AARCH64_OPND_SVE_ADDR_RR_LSL2)
(AARCH64_OPND_SVE_ADDR_RR_LSL3, AARCH64_OPND_SVE_ADDR_RX)
(AARCH64_OPND_SVE_ADDR_RX_LSL1, AARCH64_OPND_SVE_ADDR_RX_LSL2)
(AARCH64_OPND_SVE_ADDR_RX_LSL3, AARCH64_OPND_SVE_ADDR_RZ)
(AARCH64_OPND_SVE_ADDR_RZ_LSL1, AARCH64_OPND_SVE_ADDR_RZ_LSL2)
(AARCH64_OPND_SVE_ADDR_RZ_LSL3, AARCH64_OPND_SVE_ADDR_RZ_XTW_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW_22, AARCH64_OPND_SVE_ADDR_RZ_XTW1_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW1_22, AARCH64_OPND_SVE_ADDR_RZ_XTW2_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW2_22, AARCH64_OPND_SVE_ADDR_RZ_XTW3_14)
(AARCH64_OPND_SVE_ADDR_RZ_XTW3_22, AARCH64_OPND_SVE_ADDR_ZI_U5)
(AARCH64_OPND_SVE_ADDR_ZI_U5x2, AARCH64_OPND_SVE_ADDR_ZI_U5x4)
(AARCH64_OPND_SVE_ADDR_ZI_U5x8, AARCH64_OPND_SVE_ADDR_ZZ_LSL)
(AARCH64_OPND_SVE_ADDR_ZZ_SXTW, AARCH64_OPND_SVE_ADDR_ZZ_UXTW):
Likewise.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
address operands.
* aarch64-opc.h (FLD_SVE_imm6, FLD_SVE_msz, FLD_SVE_xs_14)
(FLD_SVE_xs_22): New aarch64_field_kinds.
(OPD_F_OD_MASK, OPD_F_OD_LSB, OPD_F_NO_ZR): New flags.
(get_operand_specific_data): New function.
* aarch64-opc.c (fields): Add entries for FLD_SVE_imm6, FLD_SVE_msz,
FLD_SVE_xs_14 and FLD_SVE_xs_22.
(operand_general_constraint_met_p): Handle the new SVE address
operands.
(sve_reg): New array.
(get_addr_sve_reg_name): New function.
(aarch64_print_operand): Handle the new SVE address operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_addr_ri_u6, ins_sve_addr_rr_lsl)
(ins_sve_addr_rz_xtw, ins_sve_addr_zi_u5, ins_sve_addr_zz_lsl)
(ins_sve_addr_zz_sxtw, ins_sve_addr_zz_uxtw): New inserters.
* aarch64-asm.c (aarch64_ins_sve_addr_ri_u6): New function.
(aarch64_ins_sve_addr_rr_lsl): Likewise.
(aarch64_ins_sve_addr_rz_xtw): Likewise.
(aarch64_ins_sve_addr_zi_u5): Likewise.
(aarch64_ins_sve_addr_zz): Likewise.
(aarch64_ins_sve_addr_zz_lsl): Likewise.
(aarch64_ins_sve_addr_zz_sxtw): Likewise.
(aarch64_ins_sve_addr_zz_uxtw): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_u6, ext_sve_addr_rr_lsl)
(ext_sve_addr_rz_xtw, ext_sve_addr_zi_u5, ext_sve_addr_zz_lsl)
(ext_sve_addr_zz_sxtw, ext_sve_addr_zz_uxtw): New extractors.
* aarch64-dis.c (aarch64_ext_sve_add_reg_imm): New function.
(aarch64_ext_sve_addr_ri_u6): Likewise.
(aarch64_ext_sve_addr_rr_lsl): Likewise.
(aarch64_ext_sve_addr_rz_xtw): Likewise.
(aarch64_ext_sve_addr_zi_u5): Likewise.
(aarch64_ext_sve_addr_zz): Likewise.
(aarch64_ext_sve_addr_zz_lsl): Likewise.
(aarch64_ext_sve_addr_zz_sxtw): Likewise.
(aarch64_ext_sve_addr_zz_uxtw): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (REG_TYPE_SVE_BASE, REG_TYPE_SVE_OFFSET): New
register types.
(get_reg_expected_msg): Handle them.
(aarch64_addr_reg_parse): New function, split out from
aarch64_reg_parse_32_64. Handle Z registers too.
(aarch64_reg_parse_32_64): Call it.
(parse_address_main): Add base_qualifier, offset_qualifier,
base_type and offset_type parameters. Handle SVE base and offset
registers.
(parse_address): Update call to parse_address_main.
(parse_sve_address): New function.
(parse_operands): Parse the new SVE address operands.
2016-09-21 17:55:49 +02:00
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RR", 0 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_Rm), "an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RR_LSL1", 1 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_Rm), "an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RR_LSL2", 2 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_Rm), "an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RR_LSL3", 3 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_Rm), "an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RX", \
|
|
|
|
|
(0 << OPD_F_OD_LSB) | OPD_F_NO_ZR, F(FLD_Rn,FLD_Rm), \
|
|
|
|
|
"an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RX_LSL1", \
|
|
|
|
|
(1 << OPD_F_OD_LSB) | OPD_F_NO_ZR, F(FLD_Rn,FLD_Rm), \
|
|
|
|
|
"an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RX_LSL2", \
|
|
|
|
|
(2 << OPD_F_OD_LSB) | OPD_F_NO_ZR, F(FLD_Rn,FLD_Rm), \
|
|
|
|
|
"an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RX_LSL3", \
|
|
|
|
|
(3 << OPD_F_OD_LSB) | OPD_F_NO_ZR, F(FLD_Rn,FLD_Rm), \
|
|
|
|
|
"an address with a scalar register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RZ", 0 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RZ_LSL1", 1 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RZ_LSL2", 2 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rr_lsl, "SVE_ADDR_RZ_LSL3", 3 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_Rn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW_14", \
|
|
|
|
|
0 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_14), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW_22", \
|
|
|
|
|
0 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_22), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW1_14", \
|
|
|
|
|
1 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_14), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW1_22", \
|
|
|
|
|
1 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_22), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW2_14", \
|
|
|
|
|
2 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_14), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW2_22", \
|
|
|
|
|
2 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_22), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW3_14", \
|
|
|
|
|
3 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_14), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_rz_xtw, "SVE_ADDR_RZ_XTW3_22", \
|
|
|
|
|
3 << OPD_F_OD_LSB, F(FLD_Rn,FLD_SVE_Zm_16,FLD_SVE_xs_22), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zi_u5, "SVE_ADDR_ZI_U5", 0 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_SVE_Zn), "an address with a 5-bit unsigned offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zi_u5, "SVE_ADDR_ZI_U5x2", 1 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_SVE_Zn), \
|
|
|
|
|
"an address with a 5-bit unsigned offset, multiplied by 2") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zi_u5, "SVE_ADDR_ZI_U5x4", 2 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_SVE_Zn), \
|
|
|
|
|
"an address with a 5-bit unsigned offset, multiplied by 4") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zi_u5, "SVE_ADDR_ZI_U5x8", 3 << OPD_F_OD_LSB, \
|
|
|
|
|
F(FLD_SVE_Zn), \
|
|
|
|
|
"an address with a 5-bit unsigned offset, multiplied by 8") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zz_lsl, "SVE_ADDR_ZZ_LSL", 0, \
|
|
|
|
|
F(FLD_SVE_Zn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zz_sxtw, "SVE_ADDR_ZZ_SXTW", 0, \
|
|
|
|
|
F(FLD_SVE_Zn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
|
|
|
|
Y(ADDRESS, sve_addr_zz_uxtw, "SVE_ADDR_ZZ_UXTW", 0, \
|
|
|
|
|
F(FLD_SVE_Zn,FLD_SVE_Zm_16), \
|
|
|
|
|
"an address with a vector register offset") \
|
[AArch64][SVE 27/32] Add SVE integer immediate operands
This patch adds the new SVE integer immediate operands. There are
three kinds:
- simple signed and unsigned ranges, but with new widths and positions.
- 13-bit logical immediates. These have the same form as in base AArch64,
but at a different bit position.
In the case of the "MOV Zn.<T>, #<limm>" alias of DUPM, the logical
immediate <limm> is not allowed to be a valid DUP immediate, since DUP
is preferred over DUPM for constants that both instructions can handle.
- a new 9-bit arithmetic immediate, of the form "<imm8>{, LSL #8}".
In some contexts the operand is signed and in others it's unsigned.
As an extension, we allow shifted immediates to be written as a single
integer, e.g. "#256" is equivalent to "#1, LSL #8". We also use the
shiftless form as the preferred disassembly, except for the special
case of "#0, LSL #8" (a redundant encoding of 0).
include/
* opcode/aarch64.h (AARCH64_OPND_SIMM5): New aarch64_opnd.
(AARCH64_OPND_SVE_AIMM, AARCH64_OPND_SVE_ASIMM)
(AARCH64_OPND_SVE_INV_LIMM, AARCH64_OPND_SVE_LIMM)
(AARCH64_OPND_SVE_LIMM_MOV, AARCH64_OPND_SVE_SHLIMM_PRED)
(AARCH64_OPND_SVE_SHLIMM_UNPRED, AARCH64_OPND_SVE_SHRIMM_PRED)
(AARCH64_OPND_SVE_SHRIMM_UNPRED, AARCH64_OPND_SVE_SIMM5)
(AARCH64_OPND_SVE_SIMM5B, AARCH64_OPND_SVE_SIMM6)
(AARCH64_OPND_SVE_SIMM8, AARCH64_OPND_SVE_UIMM3)
(AARCH64_OPND_SVE_UIMM7, AARCH64_OPND_SVE_UIMM8)
(AARCH64_OPND_SVE_UIMM8_53): Likewise.
(aarch64_sve_dupm_mov_immediate_p): Declare.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
integer immediate operands.
* aarch64-opc.h (FLD_SVE_immN, FLD_SVE_imm3, FLD_SVE_imm5)
(FLD_SVE_imm5b, FLD_SVE_imm7, FLD_SVE_imm8, FLD_SVE_imm9)
(FLD_SVE_immr, FLD_SVE_imms, FLD_SVE_tszh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries.
(operand_general_constraint_met_p): Handle the new SVE integer
immediate operands.
(aarch64_print_operand): Likewise.
(aarch64_sve_dupm_mov_immediate_p): New function.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_inv_limm, ins_sve_aimm, ins_sve_asimm)
(ins_sve_limm_mov, ins_sve_shlimm, ins_sve_shrimm): New inserters.
* aarch64-asm.c (aarch64_ins_limm_1): New function, split out from...
(aarch64_ins_limm): ...here.
(aarch64_ins_inv_limm): New function.
(aarch64_ins_sve_aimm): Likewise.
(aarch64_ins_sve_asimm): Likewise.
(aarch64_ins_sve_limm_mov): Likewise.
(aarch64_ins_sve_shlimm): Likewise.
(aarch64_ins_sve_shrimm): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_inv_limm, ext_sve_aimm, ext_sve_asimm)
(ext_sve_limm_mov, ext_sve_shlimm, ext_sve_shrimm): New extractors.
* aarch64-dis.c (decode_limm): New function, split out from...
(aarch64_ext_limm): ...here.
(aarch64_ext_inv_limm): New function.
(decode_sve_aimm): Likewise.
(aarch64_ext_sve_aimm): Likewise.
(aarch64_ext_sve_asimm): Likewise.
(aarch64_ext_sve_limm_mov): Likewise.
(aarch64_top_bit): Likewise.
(aarch64_ext_sve_shlimm): Likewise.
(aarch64_ext_sve_shrimm): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (parse_operands): Handle the new SVE integer
immediate operands.
2016-09-21 17:56:57 +02:00
|
|
|
|
Y(IMMEDIATE, sve_aimm, "SVE_AIMM", 0, F(FLD_SVE_imm9), \
|
|
|
|
|
"a 9-bit unsigned arithmetic operand") \
|
|
|
|
|
Y(IMMEDIATE, sve_asimm, "SVE_ASIMM", 0, F(FLD_SVE_imm9), \
|
|
|
|
|
"a 9-bit signed arithmetic operand") \
|
2016-09-21 17:57:22 +02:00
|
|
|
|
Y(IMMEDIATE, fpimm, "SVE_FPIMM8", 0, F(FLD_SVE_imm8), \
|
|
|
|
|
"an 8-bit floating-point immediate") \
|
|
|
|
|
Y(IMMEDIATE, sve_float_half_one, "SVE_I1_HALF_ONE", 0, \
|
|
|
|
|
F(FLD_SVE_i1), "either 0.5 or 1.0") \
|
|
|
|
|
Y(IMMEDIATE, sve_float_half_two, "SVE_I1_HALF_TWO", 0, \
|
|
|
|
|
F(FLD_SVE_i1), "either 0.5 or 2.0") \
|
|
|
|
|
Y(IMMEDIATE, sve_float_zero_one, "SVE_I1_ZERO_ONE", 0, \
|
|
|
|
|
F(FLD_SVE_i1), "either 0.0 or 1.0") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(IMMEDIATE, imm_rotate1, "SVE_IMM_ROT1", 0, F(FLD_SVE_rot1), \
|
|
|
|
|
"a 1-bit rotation specifier for complex arithmetic operations") \
|
|
|
|
|
Y(IMMEDIATE, imm_rotate2, "SVE_IMM_ROT2", 0, F(FLD_SVE_rot2), \
|
|
|
|
|
"a 2-bit rotation specifier for complex arithmetic operations") \
|
[AArch64][SVE 27/32] Add SVE integer immediate operands
This patch adds the new SVE integer immediate operands. There are
three kinds:
- simple signed and unsigned ranges, but with new widths and positions.
- 13-bit logical immediates. These have the same form as in base AArch64,
but at a different bit position.
In the case of the "MOV Zn.<T>, #<limm>" alias of DUPM, the logical
immediate <limm> is not allowed to be a valid DUP immediate, since DUP
is preferred over DUPM for constants that both instructions can handle.
- a new 9-bit arithmetic immediate, of the form "<imm8>{, LSL #8}".
In some contexts the operand is signed and in others it's unsigned.
As an extension, we allow shifted immediates to be written as a single
integer, e.g. "#256" is equivalent to "#1, LSL #8". We also use the
shiftless form as the preferred disassembly, except for the special
case of "#0, LSL #8" (a redundant encoding of 0).
include/
* opcode/aarch64.h (AARCH64_OPND_SIMM5): New aarch64_opnd.
(AARCH64_OPND_SVE_AIMM, AARCH64_OPND_SVE_ASIMM)
(AARCH64_OPND_SVE_INV_LIMM, AARCH64_OPND_SVE_LIMM)
(AARCH64_OPND_SVE_LIMM_MOV, AARCH64_OPND_SVE_SHLIMM_PRED)
(AARCH64_OPND_SVE_SHLIMM_UNPRED, AARCH64_OPND_SVE_SHRIMM_PRED)
(AARCH64_OPND_SVE_SHRIMM_UNPRED, AARCH64_OPND_SVE_SIMM5)
(AARCH64_OPND_SVE_SIMM5B, AARCH64_OPND_SVE_SIMM6)
(AARCH64_OPND_SVE_SIMM8, AARCH64_OPND_SVE_UIMM3)
(AARCH64_OPND_SVE_UIMM7, AARCH64_OPND_SVE_UIMM8)
(AARCH64_OPND_SVE_UIMM8_53): Likewise.
(aarch64_sve_dupm_mov_immediate_p): Declare.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
integer immediate operands.
* aarch64-opc.h (FLD_SVE_immN, FLD_SVE_imm3, FLD_SVE_imm5)
(FLD_SVE_imm5b, FLD_SVE_imm7, FLD_SVE_imm8, FLD_SVE_imm9)
(FLD_SVE_immr, FLD_SVE_imms, FLD_SVE_tszh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries.
(operand_general_constraint_met_p): Handle the new SVE integer
immediate operands.
(aarch64_print_operand): Likewise.
(aarch64_sve_dupm_mov_immediate_p): New function.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_inv_limm, ins_sve_aimm, ins_sve_asimm)
(ins_sve_limm_mov, ins_sve_shlimm, ins_sve_shrimm): New inserters.
* aarch64-asm.c (aarch64_ins_limm_1): New function, split out from...
(aarch64_ins_limm): ...here.
(aarch64_ins_inv_limm): New function.
(aarch64_ins_sve_aimm): Likewise.
(aarch64_ins_sve_asimm): Likewise.
(aarch64_ins_sve_limm_mov): Likewise.
(aarch64_ins_sve_shlimm): Likewise.
(aarch64_ins_sve_shrimm): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_inv_limm, ext_sve_aimm, ext_sve_asimm)
(ext_sve_limm_mov, ext_sve_shlimm, ext_sve_shrimm): New extractors.
* aarch64-dis.c (decode_limm): New function, split out from...
(aarch64_ext_limm): ...here.
(aarch64_ext_inv_limm): New function.
(decode_sve_aimm): Likewise.
(aarch64_ext_sve_aimm): Likewise.
(aarch64_ext_sve_asimm): Likewise.
(aarch64_ext_sve_limm_mov): Likewise.
(aarch64_top_bit): Likewise.
(aarch64_ext_sve_shlimm): Likewise.
(aarch64_ext_sve_shrimm): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (parse_operands): Handle the new SVE integer
immediate operands.
2016-09-21 17:56:57 +02:00
|
|
|
|
Y(IMMEDIATE, inv_limm, "SVE_INV_LIMM", 0, \
|
|
|
|
|
F(FLD_SVE_N,FLD_SVE_immr,FLD_SVE_imms), \
|
|
|
|
|
"an inverted 13-bit logical immediate") \
|
|
|
|
|
Y(IMMEDIATE, limm, "SVE_LIMM", 0, \
|
|
|
|
|
F(FLD_SVE_N,FLD_SVE_immr,FLD_SVE_imms), \
|
|
|
|
|
"a 13-bit logical immediate") \
|
|
|
|
|
Y(IMMEDIATE, sve_limm_mov, "SVE_LIMM_MOV", 0, \
|
|
|
|
|
F(FLD_SVE_N,FLD_SVE_immr,FLD_SVE_imms), \
|
|
|
|
|
"a 13-bit logical move immediate") \
|
2016-09-21 17:54:53 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "SVE_PATTERN", 0, F(FLD_SVE_pattern), \
|
|
|
|
|
"an enumeration value such as POW2") \
|
[AArch64][SVE 24/32] Add AARCH64_OPND_SVE_PATTERN_SCALED
Some SVE instructions count the number of elements in a given vector
pattern and allow a scale factor of [1, 16] to be applied to the result.
This scale factor is written ", MUL #n", where "MUL" is a new operator.
E.g.:
UQINCD X0, POW2, MUL #2
This patch adds support for this kind of operand.
All existing operators were shifts of some kind, so there was a natural
range of [0, 63] regardless of context. This was then narrowered further
by later checks (e.g. to [0, 31] when used for 32-bit values).
In contrast, MUL doesn't really have a natural context-independent range.
Rather than pick one arbitrarily, it seemed better to make the "shift"
amount a full 64-bit value and leave the range test to the usual
operand-checking code. I've rearranged the fields of aarch64_opnd_info
so that this doesn't increase the size of the structure (although I don't
think its size is critical anyway).
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_PATTERN_SCALED): New
aarch64_opnd.
(AARCH64_MOD_MUL): New aarch64_modifier_kind.
(aarch64_opnd_info): Make shifter.amount an int64_t and
rearrange the fields.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add an entry for
AARCH64_OPND_SVE_PATTERN_SCALED.
* aarch64-opc.h (FLD_SVE_imm4): New aarch64_field_kind.
* aarch64-opc.c (fields): Add a corresponding entry.
(set_multiplier_out_of_range_error): New function.
(aarch64_operand_modifiers): Add entry for AARCH64_MOD_MUL.
(operand_general_constraint_met_p): Handle
AARCH64_OPND_SVE_PATTERN_SCALED.
(print_register_offset_address): Use PRIi64 to print the
shift amount.
(aarch64_print_operand): Likewise. Handle
AARCH64_OPND_SVE_PATTERN_SCALED.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_scale): New inserter.
* aarch64-asm.c (aarch64_ins_sve_scale): New function.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_scale): New inserter.
* aarch64-dis.c (aarch64_ext_sve_scale): New function.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (SHIFTED_MUL): New parse_shift_mode.
(parse_shift): Handle it. Reject AARCH64_MOD_MUL for all other
shift modes. Skip range tests for AARCH64_MOD_MUL.
(process_omitted_operand): Handle AARCH64_OPND_SVE_PATTERN_SCALED.
(parse_operands): Likewise.
2016-09-21 17:55:22 +02:00
|
|
|
|
Y(IMMEDIATE, sve_scale, "SVE_PATTERN_SCALED", 0, \
|
|
|
|
|
F(FLD_SVE_pattern), "an enumeration value such as POW2") \
|
2016-09-21 17:54:53 +02:00
|
|
|
|
Y(IMMEDIATE, imm, "SVE_PRFOP", 0, F(FLD_SVE_prfop), \
|
|
|
|
|
"an enumeration value such as PLDL1KEEP") \
|
[AArch64][SVE 21/32] Add Zn and Pn registers
This patch adds the Zn and Pn registers, and associated fields and
operands.
include/
* opcode/aarch64.h (AARCH64_OPND_CLASS_SVE_REG): New
aarch64_operand_class.
(AARCH64_OPND_CLASS_PRED_REG): Likewise.
(AARCH64_OPND_SVE_Pd, AARCH64_OPND_SVE_Pg3, AARCH64_OPND_SVE_Pg4_5)
(AARCH64_OPND_SVE_Pg4_10, AARCH64_OPND_SVE_Pg4_16)
(AARCH64_OPND_SVE_Pm, AARCH64_OPND_SVE_Pn, AARCH64_OPND_SVE_Pt)
(AARCH64_OPND_SVE_Za_5, AARCH64_OPND_SVE_Za_16, AARCH64_OPND_SVE_Zd)
(AARCH64_OPND_SVE_Zm_5, AARCH64_OPND_SVE_Zm_16, AARCH64_OPND_SVE_Zn)
(AARCH64_OPND_SVE_Zn_INDEX, AARCH64_OPND_SVE_ZnxN)
(AARCH64_OPND_SVE_Zt, AARCH64_OPND_SVE_ZtxN): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for new SVE operands.
* aarch64-opc.h (FLD_SVE_Pd, FLD_SVE_Pg3, FLD_SVE_Pg4_5)
(FLD_SVE_Pg4_10, FLD_SVE_Pg4_16, FLD_SVE_Pm, FLD_SVE_Pn, FLD_SVE_Pt)
(FLD_SVE_Za_5, FLD_SVE_Za_16, FLD_SVE_Zd, FLD_SVE_Zm_5, FLD_SVE_Zm_16)
(FLD_SVE_Zn, FLD_SVE_Zt, FLD_SVE_tzsh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries here.
(operand_general_constraint_met_p): Check that SVE register lists
have the correct length. Check the ranges of SVE index registers.
Check for cases where p8-p15 are used in 3-bit predicate fields.
(aarch64_print_operand): Handle the new SVE operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_index, ins_sve_reglist): New inserters.
* aarch64-asm.c (aarch64_ins_sve_index): New function.
(aarch64_ins_sve_reglist): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_index, ext_sve_reglist): New extractors.
* aarch64-dis.c (aarch64_ext_sve_index): New function.
(aarch64_ext_sve_reglist): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (NTA_HASVARWIDTH): New macro.
(AARCH64_REG_TYPES): Add ZN and PN.
(get_reg_expected_msg): Handle them.
(parse_vector_type_for_operand): Add a reg_type parameter.
Skip the width for Zn and Pn registers.
(parse_typed_reg): Extend vector handling to Zn and Pn. Update the
call to parse_vector_type_for_operand. Set HASVARTYPE for Zn and Pn,
expecting the width to be 0.
(parse_vector_reg_list): Restrict error about [BHSD]nn operands to
REG_TYPE_VN.
(vectype_to_qualifier): Use S_[BHSD] qualifiers for NTA_HASVARWIDTH.
(parse_operands): Handle the new Zn and Pn operands.
(REGSET16): New macro, split out from...
(REGSET31): ...here.
(reg_names): Add Zn and Pn entries.
2016-09-21 17:53:54 +02:00
|
|
|
|
Y(PRED_REG, regno, "SVE_Pd", 0, F(FLD_SVE_Pd), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pg3", 0, F(FLD_SVE_Pg3), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pg4_5", 0, F(FLD_SVE_Pg4_5), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pg4_10", 0, F(FLD_SVE_Pg4_10), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pg4_16", 0, F(FLD_SVE_Pg4_16), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pm", 0, F(FLD_SVE_Pm), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pn", 0, F(FLD_SVE_Pn), \
|
|
|
|
|
"an SVE predicate register") \
|
|
|
|
|
Y(PRED_REG, regno, "SVE_Pt", 0, F(FLD_SVE_Pt), \
|
|
|
|
|
"an SVE predicate register") \
|
2016-09-21 17:57:43 +02:00
|
|
|
|
Y(INT_REG, regno, "SVE_Rm", 0, F(FLD_SVE_Rm), \
|
|
|
|
|
"an integer register or zero") \
|
|
|
|
|
Y(INT_REG, regno, "SVE_Rn_SP", OPD_F_MAYBE_SP, F(FLD_SVE_Rn), \
|
|
|
|
|
"an integer register or SP") \
|
[AArch64][SVE 27/32] Add SVE integer immediate operands
This patch adds the new SVE integer immediate operands. There are
three kinds:
- simple signed and unsigned ranges, but with new widths and positions.
- 13-bit logical immediates. These have the same form as in base AArch64,
but at a different bit position.
In the case of the "MOV Zn.<T>, #<limm>" alias of DUPM, the logical
immediate <limm> is not allowed to be a valid DUP immediate, since DUP
is preferred over DUPM for constants that both instructions can handle.
- a new 9-bit arithmetic immediate, of the form "<imm8>{, LSL #8}".
In some contexts the operand is signed and in others it's unsigned.
As an extension, we allow shifted immediates to be written as a single
integer, e.g. "#256" is equivalent to "#1, LSL #8". We also use the
shiftless form as the preferred disassembly, except for the special
case of "#0, LSL #8" (a redundant encoding of 0).
include/
* opcode/aarch64.h (AARCH64_OPND_SIMM5): New aarch64_opnd.
(AARCH64_OPND_SVE_AIMM, AARCH64_OPND_SVE_ASIMM)
(AARCH64_OPND_SVE_INV_LIMM, AARCH64_OPND_SVE_LIMM)
(AARCH64_OPND_SVE_LIMM_MOV, AARCH64_OPND_SVE_SHLIMM_PRED)
(AARCH64_OPND_SVE_SHLIMM_UNPRED, AARCH64_OPND_SVE_SHRIMM_PRED)
(AARCH64_OPND_SVE_SHRIMM_UNPRED, AARCH64_OPND_SVE_SIMM5)
(AARCH64_OPND_SVE_SIMM5B, AARCH64_OPND_SVE_SIMM6)
(AARCH64_OPND_SVE_SIMM8, AARCH64_OPND_SVE_UIMM3)
(AARCH64_OPND_SVE_UIMM7, AARCH64_OPND_SVE_UIMM8)
(AARCH64_OPND_SVE_UIMM8_53): Likewise.
(aarch64_sve_dupm_mov_immediate_p): Declare.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for the new SVE
integer immediate operands.
* aarch64-opc.h (FLD_SVE_immN, FLD_SVE_imm3, FLD_SVE_imm5)
(FLD_SVE_imm5b, FLD_SVE_imm7, FLD_SVE_imm8, FLD_SVE_imm9)
(FLD_SVE_immr, FLD_SVE_imms, FLD_SVE_tszh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries.
(operand_general_constraint_met_p): Handle the new SVE integer
immediate operands.
(aarch64_print_operand): Likewise.
(aarch64_sve_dupm_mov_immediate_p): New function.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_inv_limm, ins_sve_aimm, ins_sve_asimm)
(ins_sve_limm_mov, ins_sve_shlimm, ins_sve_shrimm): New inserters.
* aarch64-asm.c (aarch64_ins_limm_1): New function, split out from...
(aarch64_ins_limm): ...here.
(aarch64_ins_inv_limm): New function.
(aarch64_ins_sve_aimm): Likewise.
(aarch64_ins_sve_asimm): Likewise.
(aarch64_ins_sve_limm_mov): Likewise.
(aarch64_ins_sve_shlimm): Likewise.
(aarch64_ins_sve_shrimm): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_inv_limm, ext_sve_aimm, ext_sve_asimm)
(ext_sve_limm_mov, ext_sve_shlimm, ext_sve_shrimm): New extractors.
* aarch64-dis.c (decode_limm): New function, split out from...
(aarch64_ext_limm): ...here.
(aarch64_ext_inv_limm): New function.
(decode_sve_aimm): Likewise.
(aarch64_ext_sve_aimm): Likewise.
(aarch64_ext_sve_asimm): Likewise.
(aarch64_ext_sve_limm_mov): Likewise.
(aarch64_top_bit): Likewise.
(aarch64_ext_sve_shlimm): Likewise.
(aarch64_ext_sve_shrimm): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (parse_operands): Handle the new SVE integer
immediate operands.
2016-09-21 17:56:57 +02:00
|
|
|
|
Y(IMMEDIATE, sve_shlimm, "SVE_SHLIMM_PRED", 0, \
|
|
|
|
|
F(FLD_SVE_tszh,FLD_SVE_imm5), "a shift-left immediate operand") \
|
|
|
|
|
Y(IMMEDIATE, sve_shlimm, "SVE_SHLIMM_UNPRED", 0, \
|
|
|
|
|
F(FLD_SVE_tszh,FLD_imm5), "a shift-left immediate operand") \
|
|
|
|
|
Y(IMMEDIATE, sve_shrimm, "SVE_SHRIMM_PRED", 0, \
|
|
|
|
|
F(FLD_SVE_tszh,FLD_SVE_imm5), "a shift-right immediate operand") \
|
|
|
|
|
Y(IMMEDIATE, sve_shrimm, "SVE_SHRIMM_UNPRED", 0, \
|
|
|
|
|
F(FLD_SVE_tszh,FLD_imm5), "a shift-right immediate operand") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_SIMM5", OPD_F_SEXT, F(FLD_SVE_imm5), \
|
|
|
|
|
"a 5-bit signed immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_SIMM5B", OPD_F_SEXT, F(FLD_SVE_imm5b), \
|
|
|
|
|
"a 5-bit signed immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_SIMM6", OPD_F_SEXT, F(FLD_SVE_imms), \
|
|
|
|
|
"a 6-bit signed immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_SIMM8", OPD_F_SEXT, F(FLD_SVE_imm8), \
|
|
|
|
|
"an 8-bit signed immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_UIMM3", 0, F(FLD_SVE_imm3), \
|
|
|
|
|
"a 3-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_UIMM7", 0, F(FLD_SVE_imm7), \
|
|
|
|
|
"a 7-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_UIMM8", 0, F(FLD_SVE_imm8), \
|
|
|
|
|
"an 8-bit unsigned immediate") \
|
|
|
|
|
Y(IMMEDIATE, imm, "SVE_UIMM8_53", 0, F(FLD_imm5,FLD_imm3), \
|
|
|
|
|
"an 8-bit unsigned immediate") \
|
2016-09-21 17:57:43 +02:00
|
|
|
|
Y(SIMD_REG, regno, "SVE_VZn", 0, F(FLD_SVE_Zn), "a SIMD register") \
|
|
|
|
|
Y(SIMD_REG, regno, "SVE_Vd", 0, F(FLD_SVE_Vd), "a SIMD register") \
|
|
|
|
|
Y(SIMD_REG, regno, "SVE_Vm", 0, F(FLD_SVE_Vm), "a SIMD register") \
|
|
|
|
|
Y(SIMD_REG, regno, "SVE_Vn", 0, F(FLD_SVE_Vn), "a SIMD register") \
|
[AArch64][SVE 21/32] Add Zn and Pn registers
This patch adds the Zn and Pn registers, and associated fields and
operands.
include/
* opcode/aarch64.h (AARCH64_OPND_CLASS_SVE_REG): New
aarch64_operand_class.
(AARCH64_OPND_CLASS_PRED_REG): Likewise.
(AARCH64_OPND_SVE_Pd, AARCH64_OPND_SVE_Pg3, AARCH64_OPND_SVE_Pg4_5)
(AARCH64_OPND_SVE_Pg4_10, AARCH64_OPND_SVE_Pg4_16)
(AARCH64_OPND_SVE_Pm, AARCH64_OPND_SVE_Pn, AARCH64_OPND_SVE_Pt)
(AARCH64_OPND_SVE_Za_5, AARCH64_OPND_SVE_Za_16, AARCH64_OPND_SVE_Zd)
(AARCH64_OPND_SVE_Zm_5, AARCH64_OPND_SVE_Zm_16, AARCH64_OPND_SVE_Zn)
(AARCH64_OPND_SVE_Zn_INDEX, AARCH64_OPND_SVE_ZnxN)
(AARCH64_OPND_SVE_Zt, AARCH64_OPND_SVE_ZtxN): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for new SVE operands.
* aarch64-opc.h (FLD_SVE_Pd, FLD_SVE_Pg3, FLD_SVE_Pg4_5)
(FLD_SVE_Pg4_10, FLD_SVE_Pg4_16, FLD_SVE_Pm, FLD_SVE_Pn, FLD_SVE_Pt)
(FLD_SVE_Za_5, FLD_SVE_Za_16, FLD_SVE_Zd, FLD_SVE_Zm_5, FLD_SVE_Zm_16)
(FLD_SVE_Zn, FLD_SVE_Zt, FLD_SVE_tzsh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries here.
(operand_general_constraint_met_p): Check that SVE register lists
have the correct length. Check the ranges of SVE index registers.
Check for cases where p8-p15 are used in 3-bit predicate fields.
(aarch64_print_operand): Handle the new SVE operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_index, ins_sve_reglist): New inserters.
* aarch64-asm.c (aarch64_ins_sve_index): New function.
(aarch64_ins_sve_reglist): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_index, ext_sve_reglist): New extractors.
* aarch64-dis.c (aarch64_ext_sve_index): New function.
(aarch64_ext_sve_reglist): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (NTA_HASVARWIDTH): New macro.
(AARCH64_REG_TYPES): Add ZN and PN.
(get_reg_expected_msg): Handle them.
(parse_vector_type_for_operand): Add a reg_type parameter.
Skip the width for Zn and Pn registers.
(parse_typed_reg): Extend vector handling to Zn and Pn. Update the
call to parse_vector_type_for_operand. Set HASVARTYPE for Zn and Pn,
expecting the width to be 0.
(parse_vector_reg_list): Restrict error about [BHSD]nn operands to
REG_TYPE_VN.
(vectype_to_qualifier): Use S_[BHSD] qualifiers for NTA_HASVARWIDTH.
(parse_operands): Handle the new Zn and Pn operands.
(REGSET16): New macro, split out from...
(REGSET31): ...here.
(reg_names): Add Zn and Pn entries.
2016-09-21 17:53:54 +02:00
|
|
|
|
Y(SVE_REG, regno, "SVE_Za_5", 0, F(FLD_SVE_Za_5), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, regno, "SVE_Za_16", 0, F(FLD_SVE_Za_16), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, regno, "SVE_Zd", 0, F(FLD_SVE_Zd), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, regno, "SVE_Zm_5", 0, F(FLD_SVE_Zm_5), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, regno, "SVE_Zm_16", 0, F(FLD_SVE_Zm_16), \
|
|
|
|
|
"an SVE vector register") \
|
[AArch64] Additional SVE instructions
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
2017-02-24 19:29:00 +01:00
|
|
|
|
Y(SVE_REG, sve_quad_index, "SVE_Zm3_INDEX", \
|
|
|
|
|
3 << OPD_F_OD_LSB, F(FLD_SVE_Zm_16), \
|
|
|
|
|
"an indexed SVE vector register") \
|
|
|
|
|
Y(SVE_REG, sve_quad_index, "SVE_Zm3_22_INDEX", \
|
|
|
|
|
3 << OPD_F_OD_LSB, F(FLD_SVE_i3h, FLD_SVE_Zm_16), \
|
|
|
|
|
"an indexed SVE vector register") \
|
|
|
|
|
Y(SVE_REG, sve_quad_index, "SVE_Zm4_INDEX", \
|
|
|
|
|
4 << OPD_F_OD_LSB, F(FLD_SVE_Zm_16), \
|
|
|
|
|
"an indexed SVE vector register") \
|
[AArch64][SVE 21/32] Add Zn and Pn registers
This patch adds the Zn and Pn registers, and associated fields and
operands.
include/
* opcode/aarch64.h (AARCH64_OPND_CLASS_SVE_REG): New
aarch64_operand_class.
(AARCH64_OPND_CLASS_PRED_REG): Likewise.
(AARCH64_OPND_SVE_Pd, AARCH64_OPND_SVE_Pg3, AARCH64_OPND_SVE_Pg4_5)
(AARCH64_OPND_SVE_Pg4_10, AARCH64_OPND_SVE_Pg4_16)
(AARCH64_OPND_SVE_Pm, AARCH64_OPND_SVE_Pn, AARCH64_OPND_SVE_Pt)
(AARCH64_OPND_SVE_Za_5, AARCH64_OPND_SVE_Za_16, AARCH64_OPND_SVE_Zd)
(AARCH64_OPND_SVE_Zm_5, AARCH64_OPND_SVE_Zm_16, AARCH64_OPND_SVE_Zn)
(AARCH64_OPND_SVE_Zn_INDEX, AARCH64_OPND_SVE_ZnxN)
(AARCH64_OPND_SVE_Zt, AARCH64_OPND_SVE_ZtxN): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (AARCH64_OPERANDS): Add entries for new SVE operands.
* aarch64-opc.h (FLD_SVE_Pd, FLD_SVE_Pg3, FLD_SVE_Pg4_5)
(FLD_SVE_Pg4_10, FLD_SVE_Pg4_16, FLD_SVE_Pm, FLD_SVE_Pn, FLD_SVE_Pt)
(FLD_SVE_Za_5, FLD_SVE_Za_16, FLD_SVE_Zd, FLD_SVE_Zm_5, FLD_SVE_Zm_16)
(FLD_SVE_Zn, FLD_SVE_Zt, FLD_SVE_tzsh): New aarch64_field_kinds.
* aarch64-opc.c (fields): Add corresponding entries here.
(operand_general_constraint_met_p): Check that SVE register lists
have the correct length. Check the ranges of SVE index registers.
Check for cases where p8-p15 are used in 3-bit predicate fields.
(aarch64_print_operand): Handle the new SVE operands.
* aarch64-opc-2.c: Regenerate.
* aarch64-asm.h (ins_sve_index, ins_sve_reglist): New inserters.
* aarch64-asm.c (aarch64_ins_sve_index): New function.
(aarch64_ins_sve_reglist): Likewise.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_index, ext_sve_reglist): New extractors.
* aarch64-dis.c (aarch64_ext_sve_index): New function.
(aarch64_ext_sve_reglist): Likewise.
* aarch64-dis-2.c: Regenerate.
gas/
* config/tc-aarch64.c (NTA_HASVARWIDTH): New macro.
(AARCH64_REG_TYPES): Add ZN and PN.
(get_reg_expected_msg): Handle them.
(parse_vector_type_for_operand): Add a reg_type parameter.
Skip the width for Zn and Pn registers.
(parse_typed_reg): Extend vector handling to Zn and Pn. Update the
call to parse_vector_type_for_operand. Set HASVARTYPE for Zn and Pn,
expecting the width to be 0.
(parse_vector_reg_list): Restrict error about [BHSD]nn operands to
REG_TYPE_VN.
(vectype_to_qualifier): Use S_[BHSD] qualifiers for NTA_HASVARWIDTH.
(parse_operands): Handle the new Zn and Pn operands.
(REGSET16): New macro, split out from...
(REGSET31): ...here.
(reg_names): Add Zn and Pn entries.
2016-09-21 17:53:54 +02:00
|
|
|
|
Y(SVE_REG, regno, "SVE_Zn", 0, F(FLD_SVE_Zn), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, sve_index, "SVE_Zn_INDEX", 0, F(FLD_SVE_Zn), \
|
|
|
|
|
"an indexed SVE vector register") \
|
|
|
|
|
Y(SVE_REG, sve_reglist, "SVE_ZnxN", 0, F(FLD_SVE_Zn), \
|
|
|
|
|
"a list of SVE vector registers") \
|
|
|
|
|
Y(SVE_REG, regno, "SVE_Zt", 0, F(FLD_SVE_Zt), \
|
|
|
|
|
"an SVE vector register") \
|
|
|
|
|
Y(SVE_REG, sve_reglist, "SVE_ZtxN", 0, F(FLD_SVE_Zt), \
|
Adds the new Fields and Operand types for the new instructions in Armv8.4-a.
gas/
* config/tc-aarch64.c (process_omitted_operand):
Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2
and AARCH64_OPND_IMM_2.
(parse_operands): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_IMM_2, AARCH64_OPND_MASK
and AARCH64_OPND_ADDR_OFFSET.
include/
* opcode/aarch64.h:
(aarch64_opnd): Add AARCH64_OPND_Va, AARCH64_OPND_MASK,
AARCH64_OPND_IMM_2, AARCH64_OPND_ADDR_OFFSET
and AARCH64_OPND_SM3_IMM2.
(aarch64_insn_class): Add cryptosm3 and cryptosm4.
(arch64_feature_set): Make uint64_t.
opcodes/
* aarch64-asm.h (ins_addr_offset): New.
* aarch64-asm.c (aarch64_ins_reglane): Add cryptosm3.
(aarch64_ins_addr_offset): New.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_addr_offset): New.
* aarch64-dis.c (aarch64_ext_reglane): Add cryptosm3.
(aarch64_ext_addr_offset): New.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (aarch64_field_kind): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
* aarch64-opc.c (fields): Add FLD_imm6_2,
FLD_imm4_2 and FLD_SM3_imm2.
(operand_general_constraint_met_p): Add AARCH64_OPND_ADDR_OFFSET.
(aarch64_print_operand): Add AARCH64_OPND_Va, AARCH64_OPND_SM3_IMM2,
AARCH64_OPND_MASK, AARCH64_OPND_IMM_2 and AARCH64_OPND_ADDR_OFFSET.
* aarch64-opc-2.c (Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2): New.
* aarch64-tbl.h
(aarch64_opcode_table): Add Va, MASK, IMM_2, ADDR_OFFSET, SM3_IMM2.
2017-11-09 16:22:30 +01:00
|
|
|
|
"a list of SVE vector registers") \
|
|
|
|
|
Y(SIMD_ELEMENT, reglane, "SM3_IMM2", 0, F(FLD_SM3_imm2), \
|
|
|
|
|
"an indexed SM3 vector immediate")
|