This patch corrects the disassembly masks for by element dot product
instructions. The bit 10 was wrong and supposed to be 1.
This caused incorrect disassembly of instructions in the unallocated space to
disassemble as dot product instructions.
No encoding errors can arrise from this issue.
opcodes/
PR binutils/23109
* aarch64-tbl.h (aarch64_opcode_table): Correct sdot and udot.
* aarch64-dis-2.c: Regenerate.
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).
This patch adds a new platform option "notes" that can be used to indicate if
disassembly notes should be placed in the disassembly as comments.
These notes can contain information about a failing constraint such as reading
from a write-only register. The disassembly will not be blocked because of this
but -M notes will emit a comment saying that the operation is not allowed.
For assembly this patch adds a new non-fatal status for errors. This is
essentially a warning. The reason for not creating an actual warning type is
that this causes the interaction between the ordering of warnings and errors to
be problematic. Currently the error buffer is almost always filled because of
the way operands are matched during assembly. An earlier template may have put
an error there that would only be displayed if no other template matches or
generates a higher priority error. But by definition a warning is lower
priority than a warning, so the error (which is incorrect if another template
matched) will supersede the warning. By treating warnings as errors and only
later relaxing the severity this relationship keeps working and the existing
reporting infrastructure can be re-used.
binutils/
PR binutils/21446
* doc/binutils.texi (-M): Document AArch64 options.
* NEWS: Document notes and warnings.
gas/
PR binutils/21446
* config/tc-aarch64.c (print_operands): Indicate no notes.
(output_operand_error_record): Support non-fatal errors.
(output_operand_error_report, warn_unpredictable_ldst, md_assemble):
Likewise.
include/
PR binutils/21446
* opcode/aarch64.h (aarch64_operand_error): Add non_fatal.
(aarch64_print_operand): Support notes.
opcodes/
PR binutils/21446
* aarch64-dis.c (no_notes: New.
(parse_aarch64_dis_option): Support notes.
(aarch64_decode_insn, print_operands): Likewise.
(print_aarch64_disassembler_options): Document notes.
* aarch64-opc.c (aarch64_print_operand): Support notes.
This patch if the first patch in a series to add the ability to add constraints
to system registers that an instruction must adhere to in order for the register
to be usable with that instruction.
These constraints can also be used to disambiguate between registers with the
same encoding during disassembly.
This patch adds a new flags entry in the sysreg structures and ensures it is
filled in and read out during assembly/disassembly. It also adds the ability for
the assemble and disassemble functions to be able to gracefully fail and re-use
the existing error reporting infrastructure.
The return type of these functions are changed to a boolean to denote success or
failure and the error structure is passed around to them. This requires
aarch64-gen changes so a lot of the changes here are just mechanical.
gas/
PR binutils/21446
* config/tc-aarch64.c (parse_sys_reg): Return register flags.
(parse_operands): Fill in register flags.
gdb/
PR binutils/21446
* aarch64-tdep.c (aarch64_analyze_prologue,
aarch64_software_single_step, aarch64_displaced_step_copy_insn):
Indicate not interested in errors.
include/
PR binutils/21446
* opcode/aarch64.h (aarch64_opnd_info): Change sysreg to struct.
(aarch64_decode_insn): Accept error struct.
opcodes/
PR binutils/21446
* aarch64-asm.h (aarch64_insert_operand, aarch64_##x): Return boolean
and take error struct.
* aarch64-asm.c (aarch64_ext_regno, aarch64_ins_reglane,
aarch64_ins_reglist, aarch64_ins_ldst_reglist,
aarch64_ins_ldst_reglist_r, aarch64_ins_ldst_elemlist,
aarch64_ins_advsimd_imm_shift, aarch64_ins_imm, aarch64_ins_imm_half,
aarch64_ins_advsimd_imm_modified, aarch64_ins_fpimm,
aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2, aarch64_ins_fbits,
aarch64_ins_aimm, aarch64_ins_limm_1, aarch64_ins_limm,
aarch64_ins_inv_limm, aarch64_ins_ft, aarch64_ins_addr_simple,
aarch64_ins_addr_regoff, aarch64_ins_addr_offset, aarch64_ins_addr_simm,
aarch64_ins_addr_simm10, aarch64_ins_addr_uimm12,
aarch64_ins_simd_addr_post, aarch64_ins_cond, aarch64_ins_sysreg,
aarch64_ins_pstatefield, aarch64_ins_sysins_op, aarch64_ins_barrier,
aarch64_ins_prfop, aarch64_ins_hint, aarch64_ins_reg_extended,
aarch64_ins_reg_shifted, aarch64_ins_sve_addr_ri_s4xvl,
aarch64_ins_sve_addr_ri_s6xvl, aarch64_ins_sve_addr_ri_s9xvl,
aarch64_ins_sve_addr_ri_s4, aarch64_ins_sve_addr_ri_u6,
aarch64_ins_sve_addr_rr_lsl, aarch64_ins_sve_addr_rz_xtw,
aarch64_ins_sve_addr_zi_u5, aarch64_ext_sve_addr_zz,
aarch64_ins_sve_addr_zz_lsl, aarch64_ins_sve_addr_zz_sxtw,
aarch64_ins_sve_addr_zz_uxtw, aarch64_ins_sve_aimm,
aarch64_ins_sve_asimm, aarch64_ins_sve_index, aarch64_ins_sve_limm_mov,
aarch64_ins_sve_quad_index, aarch64_ins_sve_reglist,
aarch64_ins_sve_scale, aarch64_ins_sve_shlimm, aarch64_ins_sve_shrimm,
aarch64_ins_sve_float_half_one, aarch64_ins_sve_float_half_two,
aarch64_ins_sve_float_zero_one, aarch64_opcode_encode): Likewise.
* aarch64-dis.h (aarch64_extract_operand, aarch64_##x): Likewise.
* aarch64-dis.c (aarch64_ext_regno, aarch64_ext_reglane,
aarch64_ext_reglist, aarch64_ext_ldst_reglist,
aarch64_ext_ldst_reglist_r, aarch64_ext_ldst_elemlist,
aarch64_ext_advsimd_imm_shift, aarch64_ext_imm, aarch64_ext_imm_half,
aarch64_ext_advsimd_imm_modified, aarch64_ext_fpimm,
aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2, aarch64_ext_fbits,
aarch64_ext_aimm, aarch64_ext_limm_1, aarch64_ext_limm, decode_limm,
aarch64_ext_inv_limm, aarch64_ext_ft, aarch64_ext_addr_simple,
aarch64_ext_addr_regoff, aarch64_ext_addr_offset, aarch64_ext_addr_simm,
aarch64_ext_addr_simm10, aarch64_ext_addr_uimm12,
aarch64_ext_simd_addr_post, aarch64_ext_cond, aarch64_ext_sysreg,
aarch64_ext_pstatefield, aarch64_ext_sysins_op, aarch64_ext_barrier,
aarch64_ext_prfop, aarch64_ext_hint, aarch64_ext_reg_extended,
aarch64_ext_reg_shifted, aarch64_ext_sve_addr_ri_s4xvl,
aarch64_ext_sve_addr_ri_s6xvl, aarch64_ext_sve_addr_ri_s9xvl,
aarch64_ext_sve_addr_ri_s4, aarch64_ext_sve_addr_ri_u6,
aarch64_ext_sve_addr_rr_lsl, aarch64_ext_sve_addr_rz_xtw,
aarch64_ext_sve_addr_zi_u5, aarch64_ext_sve_addr_zz,
aarch64_ext_sve_addr_zz_lsl, aarch64_ext_sve_addr_zz_sxtw,
aarch64_ext_sve_addr_zz_uxtw, aarch64_ext_sve_aimm,
aarch64_ext_sve_asimm, aarch64_ext_sve_index, aarch64_ext_sve_limm_mov,
aarch64_ext_sve_quad_index, aarch64_ext_sve_reglist,
aarch64_ext_sve_scale, aarch64_ext_sve_shlimm, aarch64_ext_sve_shrimm,
aarch64_ext_sve_float_half_one, aarch64_ext_sve_float_half_two,
aarch64_ext_sve_float_zero_one, aarch64_opcode_decode): Likewise.
(determine_disassembling_preference, aarch64_decode_insn,
print_insn_aarch64_word, print_insn_data): Take errors struct.
(print_insn_aarch64): Use errors.
* aarch64-asm-2.c: Regenerate.
* aarch64-dis-2.c: Regenerate.
* aarch64-gen.c (print_operand_inserter): Use errors and change type to
boolean in aarch64_insert_operan.
(print_operand_extractor): Likewise.
* aarch64-opc.c (aarch64_print_operand): Use sysreg struct.
bfd * targets.c: Wrap nfp_elf64_vec in BFD64 ifdef.
include * opcode/nfp.h: Use uint64_t instead of bfd_vma.
opcodes * nfp-dis.c: Use uint64_t for instruction variables, not bfd_vma.
Defining SPE2_OPCD_SEGS as 13 discounts the possibility that we'd
ever look up spe2_opcd_indices[14..16], which I think is possible.
Extend that array to size 16+1, using the macros we use to index the
array. Similarly use the index macros for PPC_OPCD_SEGS and
VLE_OPCD_SEGS.
* ppc-dis.c (PPC_OPCD_SEGS): Define using PPC_OP.
(VLE_OPCD_SEGS, SPE2_OPCD_SEGS): Similarly, using macros used to
partition opcode space for index lookup.
This patch replaces AddrPrefixOp0 with AddrPrefixOpReg to indicate that
the size of register operand is controlled by the address size prefix.
This will be used by Intel MOVDIRI and MOVDIR64B instructions later.
gas/
* config/tc-i386.c (process_suffix): Check addrprefixopreg
instead of addrprefixop0.
opcodes/
* i386-gen.c (opcode_modifiers): Replace AddrPrefixOp0 with
AddrPrefixOpReg.
* i386-opc.h (AddrPrefixOp0): Renamed to ...
(AddrPrefixOpReg): This.
(i386_opcode_modifier): Rename addrprefixop0 to addrprefixopreg.
* i386-opc.tbl: Replace AddrPrefixOp0 with AddrPrefixOpReg.
This patch fixes an issue where the memory for the opcode structure is not zero'd before
the first exit branch. So there is one failure mode for which uninitialized memory
is returned.
This causes weird failures when the return code is not checked before inst is used.
opcodes/
* aarch64-dis.c (aarch64_opcode_decode): Moved memory clear code.
There's little point carrying up to three templates per insn flavor
when the sole difference is operand size and the dependency on AVX512VL
being enabled. Instead the need for AVX512VL can be derived from an
operand allowing for ZMMword as well as one or both or XMMword and
YMMword (irrespective of whether this is a register or memory operand).
Without further abstraction to deal with the different Disp8MemShift
values between the templates, only a limited set (mostly ones only
allowing for non-memory operands) can be folded, which is being done
here.
Also drop IgnoreSize wherever possible from anything that's being
touched anyway.
All of AVX, LWP, MPX, and PKU require XSAVE, and hence it as well as
XRSTOR should be enabled when enabling these ISA extensions. Leverage
these implications to shorten some of the cpu_flag_init[] entries.
It's not clear to me why they had been introduced - the respective
comments in opcodes/i386-gen.c are certainly wrong: ymm<N> registers
are very well supported (and necessary) with just AVX512F.
Neither 287 wrt 8087 nor 387 wrt 287 are proper supersets - in each case
some insns get removed from the ISA (they become NOPs, but code intended
for newer co-processors should not use them).
Furthermore with .no87, ST should not be recognized as a register name.
Rn is supposed to have a 5 bit range but instead was given 4 bits
causing these instructions to disassemble as unknown instructions.
opcodes/
* aarch64-tbl.h (sqrdmlah, sqrdmlsh): Fix masks.
gas/
* testsuite/gas/aarch64/rdma.s: Test for larger register numbers.
* testsuite/gas/aarch64/rdma.d: Update results.
* testsuite/gas/aarch64/rdma-directive.d: Likewise.
Since only the first 32 bits of input operand are used for tpause and
umwait, the REX.W bit is skipped. Both 32-bit registers and 64-bit
registers are allowed.
gas/
* testsuite/gas/i386/x86-64-waitpkg.s: Add 32-bit registers
tests for tpause and umwait.
* testsuite/gas/i386/x86-64-waitpkg-intel.d: Updated.
* testsuite/gas/i386/x86-64-waitpkg.d: Likewise.
opcodes/
* i386-dis.c (prefix_table): Replace Em with Edq on tpause and
umwait.
* i386-opc.tbl: Allow 32-bit registers for tpause and umwait in
64-bit mode.
* i386-tbl.h: Regenerated.
"vex" has many fields to control how to decode an instruction. Clear
all fields in "vex" before decoding an instruction to avoid using values
left from the previous instruction.
gas/
PR binutils/23025
* testsuite/gas/i386/prefix.s: Add tests for vcvtpd2dq with
VEX and EVEX prefixes.
* testsuite/gas/i386/prefix.d: Updated.
opcodes/
PR binutils/23025
* i386-dis.c (get_valid_dis386): Don't set vex.prefix nor vex.w
to 0.
(print_insn): Clear vex instead of vex.evex.
It again can be inferred from other information.
The vpopcntd templates all need to have Dword added to their memory
operands; the lack thereof was actually a bug preventing certain Intel
syntax code to assemble, so test cases get extended.
In the course of folding their patterns (possible now that the pointless
and partly even bogus VecESize are no longer in the way) I've noticed
that vcvt*2usi, other than their vcvt*2si counterparts, don't allow for
any suffixes. As that is supposedly intentional, make the disassembler
consistently omit suffixes for all to-scalar-int conversion insns.
A few of them were missed in commit 7ac2002247 ("x86: derive DispN
from BaseIndex") and also couldn't be removed by subsequent commits
touching certain templates anyway.
The wrong placement of the Load attribute in the templates prevented
this from working. The disassembler also didn't handle this consistently
with other similar dual-encoding insns.
While many templates allowing multiple suitably matching XMM/YMM/ZMM
operand sizes can be folded, a few need to be split in order to not
wrongly accept "xmmword ptr" operands when only XMM registers are
permitted (and memory operands are more narrow). Add a test case
validating this.
Alan Modra
Nick Clifton
Tamar Christina
Francois H. Theron
H.J. Lu
Jim Wilson
Peter Bergner
Igor Tsimbalist
Jan Beulich