binutils-gdb/cpu/or1korfpx.cpu

387 lines
12 KiB
Plaintext
Raw Normal View History

; OpenRISC 1000 architecture. -*- Scheme -*-
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
; Copyright 2000-2019 Free Software Foundation, Inc.
; Contributed by Peter Gavin, pgavin@gmail.com
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
; Modified by Andrey Bacherov, avbacherov@opencores.org
;
; This program 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 of the License, or
; (at your option) any later version.
;
; This program 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 program; if not, see <http://www.gnu.org/licenses/>
; Initial ORFPX32 instruction set
; I'm not sure how CGEN handles rounding in FP operations, except for
; in conversions to/from integers. So lf.add, lf.sub, lf.mul, and
; lf.div do not round according to the FPCSR RM field.
; NaN, overflow, and underflow are not yet handled either.
(define-normal-insn-enum insn-opcode-float-regreg
"floating point reg/reg insn opcode enums" ()
OPC_FLOAT_REGREG_ f-op-7-8
(("ADD_S" #x00)
("SUB_S" #x01)
("MUL_S" #x02)
("DIV_S" #x03)
("ITOF_S" #x04)
("FTOI_S" #x05)
("REM_S" #x06)
("MADD_S" #x07)
("SFEQ_S" #x08)
("SFNE_S" #x09)
("SFGT_S" #x0a)
("SFGE_S" #x0b)
("SFLT_S" #x0c)
("SFLE_S" #x0d)
("ADD_D" #x10)
("SUB_D" #x11)
("MUL_D" #x12)
("DIV_D" #x13)
("ITOF_D" #x14)
("FTOI_D" #x15)
("REM_D" #x16)
("MADD_D" #x17)
("SFEQ_D" #x18)
("SFNE_D" #x19)
("SFGT_D" #x1a)
("SFGE_D" #x1b)
("SFLT_D" #x1c)
("SFLE_D" #x1d)
("SFUEQ_S" #x28)
("SFUNE_S" #x29)
("SFUGT_S" #x2a)
("SFUGE_S" #x2b)
("SFULT_S" #x2c)
("SFULE_S" #x2d)
("SFUN_S" #x2e)
("SFUEQ_D" #x38)
("SFUNE_D" #x39)
("SFUGT_D" #x3a)
("SFUGE_D" #x3b)
("SFULT_D" #x3c)
("SFULE_D" #x3d)
("SFUN_D" #x3e)
("CUST1_S" #xd0)
("CUST1_D" #xe0)
)
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
; Register offset flags, if set offset is 2 otherwise offset is 1
(dnf f-rdoff-10-1 "destination register pair offset flag" ((MACH ORFPX64A32-MACHS)) 10 1)
(dnf f-raoff-9-1 "source register A pair offset flag" ((MACH ORFPX64A32-MACHS)) 9 1)
(dnf f-rboff-8-1 "source register B pair offset flag" ((MACH ORFPX64A32-MACHS)) 8 1)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dsh h-roff1 "1-bit offset flag" () (register BI))
(dnop rDSF "destination register (single floating point mode)" ((MACH ORFPX32-MACHS)) h-fsr f-r1)
(dnop rASF "source register A (single floating point mode)" ((MACH ORFPX32-MACHS)) h-fsr f-r2)
(dnop rBSF "source register B (single floating point mode)" ((MACH ORFPX32-MACHS)) h-fsr f-r3)
(dnop rDDF "or64 destination register (double floating point mode)" ((MACH ORFPX64-MACHS)) h-fdr f-r1)
(dnop rADF "or64 source register A (double floating point mode)" ((MACH ORFPX64-MACHS)) h-fdr f-r2)
(dnop rBDF "or64 source register B (double floating point mode)" ((MACH ORFPX64-MACHS)) h-fdr f-r3)
(define-pmacro (double-field-and-ops mnemonic reg offbit op-comment)
(begin
(define-multi-ifield
(name (.sym "f-r" (.downcase mnemonic) "d32"))
(comment op-comment)
(attrs (MACH ORFPX64A32-MACHS))
(mode SI)
(subfields reg offbit)
; From the multi-ifield insert the bits into subfields
(insert (sequence
()
(set (ifield reg)
(and (ifield (.sym "f-r" (.downcase mnemonic) "d32"))
(const #x1f))
)
(set (ifield offbit)
(and (sra (ifield (.sym "f-r" (.downcase mnemonic) "d32"))
(const 5))
(const 1))
)
)
)
; Extract the multi-ifield from the subfield bits
(extract
(set (ifield (.sym "f-r" (.downcase mnemonic) "d32"))
(or (ifield reg)
(sll (ifield offbit)
(const 5)))
)
)
)
(define-operand
(name (.sym "r" (.upcase mnemonic) "D32F"))
(comment (.str op-comment " (double floating point pair)"))
(attrs (MACH ORFPX64A32-MACHS))
(type h-fd32r)
(index (.sym "f-r" (.downcase mnemonic) "d32"))
(handlers (parse "regpair") (print "regpair"))
)
(define-operand
(name (.sym "r" (.upcase mnemonic) "DI"))
(comment (.str op-comment " (double integer pair)"))
(attrs (MACH ORFPX64A32-MACHS))
(type h-i64r)
(index (.sym "f-r" (.downcase mnemonic) "d32"))
(handlers (parse "regpair") (print "regpair"))
)
)
)
(double-field-and-ops D f-r1 f-rdoff-10-1 "destination register")
(double-field-and-ops A f-r2 f-raoff-9-1 "source register A")
(double-field-and-ops B f-r3 f-rboff-8-1 "source register B")
(define-pmacro (float-regreg-insn mnemonic)
(begin
(dni (.sym lf- mnemonic -s)
(.str "lf." mnemonic ".s reg/reg/reg")
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
(.str "lf." mnemonic ".s $rDSF,$rASF,$rBSF")
(+ OPC_FLOAT rDSF rASF rBSF (f-resv-10-3 0) (.sym OPC_FLOAT_REGREG_ (.upcase mnemonic) _S))
(set SF rDSF (mnemonic SF rASF rBSF))
()
)
(dni (.sym lf- mnemonic -d)
(.str "lf." mnemonic ".d reg/reg/reg")
((MACH ORFPX64-MACHS))
(.str "lf." mnemonic ".d $rDDF,$rADF,$rBDF")
(+ OPC_FLOAT rDDF rADF rBDF (f-resv-10-3 0) (.sym OPC_FLOAT_REGREG_ (.upcase mnemonic) _D))
(set DF rDDF (mnemonic DF rADF rBDF))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni (.sym lf- mnemonic -d32)
(.str "lf." mnemonic ".d regpair/regpair/regpair")
((MACH ORFPX64A32-MACHS))
(.str "lf." mnemonic ".d $rDD32F,$rAD32F,$rBD32F")
(+ OPC_FLOAT rDD32F rAD32F rBD32F (.sym OPC_FLOAT_REGREG_ (.upcase mnemonic) _D))
(set DF rDD32F (mnemonic DF rAD32F rBD32F))
()
)
)
)
(float-regreg-insn add)
(float-regreg-insn sub)
(float-regreg-insn mul)
(float-regreg-insn div)
(dni lf-rem-s
"lf.rem.s reg/reg/reg"
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
"lf.rem.s $rDSF,$rASF,$rBSF"
(+ OPC_FLOAT rDSF rASF rBSF (f-resv-10-3 0) OPC_FLOAT_REGREG_REM_S)
(set SF rDSF (rem SF rASF rBSF))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni lf-rem-d
"lf.rem.d reg/reg/reg"
((MACH ORFPX64-MACHS))
"lf.rem.d $rDDF,$rADF,$rBDF"
(+ OPC_FLOAT rDDF rADF rBDF (f-resv-10-3 0) OPC_FLOAT_REGREG_REM_D)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(set DF rDDF (rem DF rADF rBDF))
()
)
(dni lf-rem-d32
"lf.rem.d regpair/regpair/regpair"
((MACH ORFPX64A32-MACHS))
"lf.rem.d $rDD32F,$rAD32F,$rBD32F"
(+ OPC_FLOAT rDD32F rAD32F rBD32F OPC_FLOAT_REGREG_REM_D)
(set DF rDD32F (rem DF rAD32F rBD32F))
()
)
(define-pmacro (get-rounding-mode)
(case INT sys-fpcsr-rm
((0) 1) ; TIES-TO-EVEN -- I'm assuming this is what is meant by "round to nearest"
((1) 3) ; TOWARD-ZERO
((2) 4) ; TOWARD-POSITIVE
(else 5) ; TOWARD-NEGATIVE
)
)
(dni lf-itof-s
"lf.itof.s reg/reg"
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
"lf.itof.s $rDSF,$rA"
(+ OPC_FLOAT rDSF rA (f-r3 0) (f-resv-10-3 0) OPC_FLOAT_REGREG_ITOF_S)
(set SF rDSF (float SF (get-rounding-mode) (trunc SI rA)))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni lf-itof-d
"lf.itof.d reg/reg"
((MACH ORFPX64-MACHS))
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
"lf.itof.d $rDDF,$rA"
(+ OPC_FLOAT rDDF rA (f-r3 0) (f-resv-10-3 0) OPC_FLOAT_REGREG_ITOF_D)
(set DF rDDF (float DF (get-rounding-mode) rA))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni lf-itof-d32
"lf.itof.d regpair/regpair"
((MACH ORFPX64A32-MACHS))
"lf.itof.d $rDD32F,$rADI"
(+ OPC_FLOAT rDD32F rADI (f-r3 0) (f-resv-8-1 0) OPC_FLOAT_REGREG_ITOF_D)
(set DF rDD32F (float DF (get-rounding-mode) rADI))
()
)
(dni lf-ftoi-s
"lf.ftoi.s reg/reg"
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
"lf.ftoi.s $rD,$rASF"
(+ OPC_FLOAT rD rASF (f-r3 0) (f-resv-10-3 0) OPC_FLOAT_REGREG_FTOI_S)
(set WI rD (ext WI (fix SI (get-rounding-mode) rASF)))
()
)
(dni lf-ftoi-d
"lf.ftoi.d reg/reg"
((MACH ORFPX64-MACHS))
"lf.ftoi.d $rD,$rADF"
(+ OPC_FLOAT rD rADF (f-r3 0) (f-resv-10-3 0) OPC_FLOAT_REGREG_FTOI_D)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(set WI rD (fix WI (get-rounding-mode) rADF))
()
)
(dni lf-ftoi-d32
"lf.ftoi.d regpair/regpair"
((MACH ORFPX64A32-MACHS))
"lf.ftoi.d $rDDI,$rAD32F"
(+ OPC_FLOAT rDDI rAD32F (f-r3 0) (f-resv-8-1 0) OPC_FLOAT_REGREG_FTOI_D)
(set DI rDDI (fix DI (get-rounding-mode) rAD32F))
()
)
(define-pmacro (float-setflag-insn-base mnemonic rtx-mnemonic symantics)
(begin
(dni (.sym lf-sf mnemonic -s)
(.str "lf.sf" mnemonic ".s reg/reg")
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
(.str "lf.sf" mnemonic ".s $rASF,$rBSF")
(+ OPC_FLOAT (f-r1 0) rASF rBSF (f-resv-10-3 0) (.sym OPC_FLOAT_REGREG_SF (.upcase mnemonic) _S))
(symantics rtx-mnemonic SF rASF rBSF)
()
)
(dni (.sym lf-sf mnemonic -d)
(.str "lf.sf" mnemonic ".d reg/reg")
((MACH ORFPX64-MACHS))
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(.str "lf.sf" mnemonic ".d $rADF,$rBDF")
(+ OPC_FLOAT (f-r1 0) rADF rBDF (f-resv-10-3 0) (.sym OPC_FLOAT_REGREG_SF (.upcase mnemonic) _D))
(symantics rtx-mnemonic DF rADF rBDF)
()
)
(dni (.sym lf-sf mnemonic -d32)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(.str "lf.sf" mnemonic ".d regpair/regpair")
((MACH ORFPX64A32-MACHS))
(.str "lf.sf" mnemonic ".d $rAD32F,$rBD32F")
(+ OPC_FLOAT (f-r1 0) rAD32F rBD32F (f-resv-10-1 0) (.sym OPC_FLOAT_REGREG_SF (.upcase mnemonic) _D))
(symantics rtx-mnemonic DF rAD32F rBD32F)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
()
)
)
)
(define-pmacro (float-setflag-symantics mnemonic mode r1 r2)
(set BI sys-sr-f (mnemonic mode r1 r2)))
(define-pmacro (float-setflag-insn mnemonic)
(float-setflag-insn-base mnemonic mnemonic float-setflag-symantics))
(define-pmacro (float-setflag-unordered-cmp-symantics mnemonic mode r1 r2)
(set BI sys-sr-f (or (unordered mode r1 r2)
(mnemonic mode r1 r2))))
(define-pmacro (float-setflag-unordered-symantics mnemonic mode r1 r2)
(set BI sys-sr-f (unordered mode r1 r2)))
(define-pmacro (float-setflag-unordered-insn mnemonic)
(float-setflag-insn-base (.str "u" mnemonic)
mnemonic
float-setflag-unordered-cmp-symantics))
(float-setflag-insn eq)
(float-setflag-insn ne)
(float-setflag-insn ge)
(float-setflag-insn gt)
(float-setflag-insn lt)
(float-setflag-insn le)
(float-setflag-unordered-insn eq)
(float-setflag-unordered-insn ne)
(float-setflag-unordered-insn gt)
(float-setflag-unordered-insn ge)
(float-setflag-unordered-insn lt)
(float-setflag-unordered-insn le)
(float-setflag-insn-base un () float-setflag-unordered-symantics)
(dni lf-madd-s
"lf.madd.s reg/reg/reg"
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
"lf.madd.s $rDSF,$rASF,$rBSF"
(+ OPC_FLOAT rDSF rASF rBSF (f-resv-10-3 0) OPC_FLOAT_REGREG_MADD_S)
(set SF rDSF (add SF (mul SF rASF rBSF) rDSF))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni lf-madd-d
"lf.madd.d reg/reg/reg"
((MACH ORFPX64-MACHS))
"lf.madd.d $rDDF,$rADF,$rBDF"
(+ OPC_FLOAT rDDF rADF rBDF (f-resv-10-3 0) OPC_FLOAT_REGREG_MADD_D)
(set DF rDDF (add DF (mul DF rADF rBDF) rDDF))
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni lf-madd-d32
"lf.madd.d regpair/regpair/regpair"
((MACH ORFPX64A32-MACHS))
"lf.madd.d $rDD32F,$rAD32F,$rBD32F"
(+ OPC_FLOAT rDD32F rAD32F rBD32F OPC_FLOAT_REGREG_MADD_D)
(set DF rDD32F (add DF (mul DF rAD32F rBD32F) rDD32F))
()
)
(define-pmacro (float-cust-insn cust-num)
(begin
(dni (.sym "lf-cust" cust-num "-s")
(.str "lf.cust" cust-num ".s")
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
((MACH ORFPX32-MACHS))
(.str "lf.cust" cust-num ".s $rASF,$rBSF")
(+ OPC_FLOAT (f-resv-25-5 0) rASF rBSF (f-resv-10-3 0) (.sym "OPC_FLOAT_REGREG_CUST" cust-num "_S"))
(nop)
()
)
(dni (.sym "lf-cust" cust-num "-d")
(.str "lf.cust" cust-num ".d")
((MACH ORFPX64-MACHS))
(.str "lf.cust" cust-num ".d")
(+ OPC_FLOAT (f-resv-25-5 0) rADF rBDF (f-resv-10-3 0) (.sym "OPC_FLOAT_REGREG_CUST" cust-num "_D"))
(nop)
()
)
cpu/or1k: Add support for orfp64a32 spec This patch adds support for OpenRISC 64-bit FPU operations on 32-bit cores by using register pairs. The functionality has been added to OpenRISC architecture specification version 1.3 as per architecture proposal 14[0]. For supporting assembly of both 64-bit and 32-bit precision instructions we have defined CGEN_VALIDATE_INSN_SUPPORTED. This allows cgen to use 64-bit bit architecture assembly parsing on 64-bit toolchains and 32-bit architecture assembly parsing on 32-bit toolchains. Without this the assembler has issues parsing register pairs. This patch also contains a few fixes to the symantics for existing OpenRISC single and double precision FPU operations. [0] https://openrisc.io/proposals/orfpx64a32 cpu/ChangeLog: yyyy-mm-dd Andrey Bacherov <avbacherov@opencores.org> Stafford Horne <shorne@gmail.com> * or1k.cpu (ORFPX64A32-MACHS): New pmacro. (ORFPX-MACHS): Removed pmacro. * or1k.opc (or1k_cgen_insn_supported): New function. (CGEN_VALIDATE_INSN_SUPPORTED): Define macro. (parse_regpair, print_regpair): New functions. * or1kcommon.cpu (h-spr, spr-shift, spr-address, h-gpr): Reorder and add comments. (h-fdr): Update comment to indicate or64. (reg-pair-reg-lo, reg-pair-reg-hi): New pmacros for register pairs. (h-fd32r): New hardware for 64-bit fpu registers. (h-i64r): New hardware for 64-bit int registers. * or1korbis.cpu (f-resv-8-1): New field. * or1korfpx.cpu (rDSF, rASF, rBSF): Update attribute to ORFPX32-MACHS. (rDDF, rADF, rBDF): Update operand comment to indicate or64. (f-rdoff-10-1, f-raoff-9-1, f-rboff-8-1): New fields. (h-roff1): New hardware. (double-field-and-ops mnemonic): New pmacro to generate operations rDD32F, rAD32F, rBD32F, rDDI and rADI. (float-regreg-insn): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (float-setflag-insn): Update single precision generator to MACH ORFPX32-MACHS. Fix double instructions from single to double precision. Add generator for or32 64-bit instructions. (float-cust-insn cust-num): Update single precision generator to MACH ORFPX32-MACHS. Add generator for or32 64-bit instructions. (lf-rem-s, lf-itof-s, lf-ftoi-s, lf-madd-s): Update MACH to ORFPX32-MACHS. (lf-rem-d): Fix operation from mod to rem. (lf-rem-d32, lf-itof-d32, lf-ftoi-d32, lf-madd-d32): New instruction. (lf-itof-d): Fix operands from single to double. (lf-ftoi-d): Update operand mode from DI to WI.
2019-06-12 23:16:18 +02:00
(dni (.sym "lf-cust" cust-num "-d32")
(.str "lf.cust" cust-num ".d")
((MACH ORFPX64A32-MACHS))
(.str "lf.cust" cust-num ".d")
(+ OPC_FLOAT (f-resv-25-5 0) rAD32F rBD32F (f-resv-10-1 0) (.sym "OPC_FLOAT_REGREG_CUST" cust-num "_D"))
(nop)
()
)
)
)
(float-cust-insn "1")