From: Martin Möhrmann Date: Fri, 24 Aug 2018 09:02:00 +0000 (+0200) Subject: all: align cpu feature variable offset naming X-Git-Tag: go1.12beta1~1278 X-Git-Url: http://www.git.cypherpunks.su/?a=commitdiff_plain;h=05c02444eb2d8b8d3ecd949c4308d8e2323ae087;p=gostls13.git all: align cpu feature variable offset naming Add an "offset_" prefix to all cpu feature variable offset constants to signify that they are not boolean cpu feature variables. Remove _ from offset constant names. Change-Id: I6e22a79ebcbe6e2ae54c4ac8764f9260bb3223ff Reviewed-on: https://go-review.googlesource.com/131215 Run-TryBot: Martin Möhrmann TryBot-Result: Gobot Gobot Reviewed-by: Brad Fitzpatrick --- diff --git a/src/internal/bytealg/bytealg.go b/src/internal/bytealg/bytealg.go index 1ab7c30f4e..9ecd8eb004 100644 --- a/src/internal/bytealg/bytealg.go +++ b/src/internal/bytealg/bytealg.go @@ -11,11 +11,12 @@ import ( // Offsets into internal/cpu records for use in assembly. const ( - x86_HasSSE2 = unsafe.Offsetof(cpu.X86.HasSSE2) - x86_HasSSE42 = unsafe.Offsetof(cpu.X86.HasSSE42) - x86_HasAVX2 = unsafe.Offsetof(cpu.X86.HasAVX2) - x86_HasPOPCNT = unsafe.Offsetof(cpu.X86.HasPOPCNT) - s390x_HasVX = unsafe.Offsetof(cpu.S390X.HasVX) + offsetX86HasSSE2 = unsafe.Offsetof(cpu.X86.HasSSE2) + offsetX86HasSSE42 = unsafe.Offsetof(cpu.X86.HasSSE42) + offsetX86HasAVX2 = unsafe.Offsetof(cpu.X86.HasAVX2) + offsetX86HasPOPCNT = unsafe.Offsetof(cpu.X86.HasPOPCNT) + + offsetS390xHasVX = unsafe.Offsetof(cpu.S390X.HasVX) ) // MaxLen is the maximum length of the string to be searched for (argument b) in Index. diff --git a/src/internal/bytealg/compare_amd64.s b/src/internal/bytealg/compare_amd64.s index 277d77c545..05bef4aad9 100644 --- a/src/internal/bytealg/compare_amd64.s +++ b/src/internal/bytealg/compare_amd64.s @@ -47,7 +47,7 @@ TEXT cmpbody<>(SB),NOSPLIT,$0-0 CMPQ R8, $63 JBE loop - CMPB internal∕cpu·X86+const_x86_HasAVX2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasAVX2(SB), $1 JEQ big_loop_avx2 JMP big_loop loop: diff --git a/src/internal/bytealg/count_amd64.s b/src/internal/bytealg/count_amd64.s index cecba11cf9..fa864c4c76 100644 --- a/src/internal/bytealg/count_amd64.s +++ b/src/internal/bytealg/count_amd64.s @@ -6,7 +6,7 @@ #include "textflag.h" TEXT ·Count(SB),NOSPLIT,$0-40 - CMPB internal∕cpu·X86+const_x86_HasPOPCNT(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasPOPCNT(SB), $1 JEQ 2(PC) JMP ·countGeneric(SB) MOVQ b_base+0(FP), SI @@ -16,7 +16,7 @@ TEXT ·Count(SB),NOSPLIT,$0-40 JMP countbody<>(SB) TEXT ·CountString(SB),NOSPLIT,$0-32 - CMPB internal∕cpu·X86+const_x86_HasPOPCNT(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasPOPCNT(SB), $1 JEQ 2(PC) JMP ·countGenericString(SB) MOVQ s_base+0(FP), SI @@ -151,7 +151,7 @@ endofpage: RET avx2: - CMPB internal∕cpu·X86+const_x86_HasAVX2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasAVX2(SB), $1 JNE sse MOVD AX, X0 LEAQ -32(SI)(BX*1), R11 diff --git a/src/internal/bytealg/equal_386.s b/src/internal/bytealg/equal_386.s index c048b6cebc..273389284e 100644 --- a/src/internal/bytealg/equal_386.s +++ b/src/internal/bytealg/equal_386.s @@ -80,7 +80,7 @@ TEXT memeqbody<>(SB),NOSPLIT,$0-0 hugeloop: CMPL BX, $64 JB bigloop - CMPB internal∕cpu·X86+const_x86_HasSSE2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasSSE2(SB), $1 JNE bigloop MOVOU (SI), X0 MOVOU (DI), X1 diff --git a/src/internal/bytealg/equal_amd64.s b/src/internal/bytealg/equal_amd64.s index cbc62dc1d8..5263d3040d 100644 --- a/src/internal/bytealg/equal_amd64.s +++ b/src/internal/bytealg/equal_amd64.s @@ -77,7 +77,7 @@ TEXT memeqbody<>(SB),NOSPLIT,$0-0 JB small CMPQ BX, $64 JB bigloop - CMPB internal∕cpu·X86+const_x86_HasAVX2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasAVX2(SB), $1 JE hugeloop_avx2 // 64 bytes at a time using xmm registers diff --git a/src/internal/bytealg/index_amd64.s b/src/internal/bytealg/index_amd64.s index f7297c0cab..4459820801 100644 --- a/src/internal/bytealg/index_amd64.s +++ b/src/internal/bytealg/index_amd64.s @@ -233,7 +233,7 @@ success_avx2: VZEROUPPER JMP success sse42: - CMPB internal∕cpu·X86+const_x86_HasSSE42(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasSSE42(SB), $1 JNE no_sse42 CMPQ AX, $12 // PCMPESTRI is slower than normal compare, diff --git a/src/internal/bytealg/indexbyte_amd64.s b/src/internal/bytealg/indexbyte_amd64.s index 359f38904b..5bf8866476 100644 --- a/src/internal/bytealg/indexbyte_amd64.s +++ b/src/internal/bytealg/indexbyte_amd64.s @@ -139,7 +139,7 @@ endofpage: RET avx2: - CMPB internal∕cpu·X86+const_x86_HasAVX2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasAVX2(SB), $1 JNE sse MOVD AX, X0 LEAQ -32(SI)(BX*1), R11 diff --git a/src/internal/bytealg/indexbyte_s390x.s b/src/internal/bytealg/indexbyte_s390x.s index 15fd2935b4..24f5ce17fa 100644 --- a/src/internal/bytealg/indexbyte_s390x.s +++ b/src/internal/bytealg/indexbyte_s390x.s @@ -64,7 +64,7 @@ notfound: RET large: - MOVBZ internal∕cpu·S390X+const_s390x_HasVX(SB), R1 + MOVBZ internal∕cpu·S390X+const_offsetS390xHasVX(SB), R1 CMPBNE R1, $0, vectorimpl srstimpl: // no vector facility diff --git a/src/runtime/asm_386.s b/src/runtime/asm_386.s index 725271eec4..7761415ecd 100644 --- a/src/runtime/asm_386.s +++ b/src/runtime/asm_386.s @@ -881,7 +881,7 @@ TEXT runtime·stackcheck(SB), NOSPLIT, $0-0 // func cputicks() int64 TEXT runtime·cputicks(SB),NOSPLIT,$0-8 - CMPB internal∕cpu·X86+const_offset_x86_HasSSE2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasSSE2(SB), $1 JNE done CMPB runtime·lfenceBeforeRdtsc(SB), $1 JNE mfence diff --git a/src/runtime/cpuflags.go b/src/runtime/cpuflags.go index 050168c2d7..b65523766a 100644 --- a/src/runtime/cpuflags.go +++ b/src/runtime/cpuflags.go @@ -11,9 +11,9 @@ import ( // Offsets into internal/cpu records for use in assembly. const ( - offset_x86_HasAVX2 = unsafe.Offsetof(cpu.X86.HasAVX2) - offset_x86_HasERMS = unsafe.Offsetof(cpu.X86.HasERMS) - offset_x86_HasSSE2 = unsafe.Offsetof(cpu.X86.HasSSE2) + offsetX86HasAVX2 = unsafe.Offsetof(cpu.X86.HasAVX2) + offsetX86HasERMS = unsafe.Offsetof(cpu.X86.HasERMS) + offsetX86HasSSE2 = unsafe.Offsetof(cpu.X86.HasSSE2) - offset_arm_HasIDIVA = unsafe.Offsetof(cpu.ARM.HasIDIVA) + offsetARMHasIDIVA = unsafe.Offsetof(cpu.ARM.HasIDIVA) ) diff --git a/src/runtime/memclr_386.s b/src/runtime/memclr_386.s index 318f883964..65f7196312 100644 --- a/src/runtime/memclr_386.s +++ b/src/runtime/memclr_386.s @@ -29,7 +29,7 @@ tail: JBE _5through8 CMPL BX, $16 JBE _9through16 - CMPB internal∕cpu·X86+const_offset_x86_HasSSE2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasSSE2(SB), $1 JNE nosse2 PXOR X0, X0 CMPL BX, $32 diff --git a/src/runtime/memclr_amd64.s b/src/runtime/memclr_amd64.s index b64b1477f9..d79078fd00 100644 --- a/src/runtime/memclr_amd64.s +++ b/src/runtime/memclr_amd64.s @@ -38,7 +38,7 @@ tail: JBE _65through128 CMPQ BX, $256 JBE _129through256 - CMPB internal∕cpu·X86+const_offset_x86_HasAVX2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasAVX2(SB), $1 JE loop_preheader_avx2 // TODO: for really big clears, use MOVNTDQ, even without AVX2. diff --git a/src/runtime/memmove_386.s b/src/runtime/memmove_386.s index 85c622b6b6..7b54070f59 100644 --- a/src/runtime/memmove_386.s +++ b/src/runtime/memmove_386.s @@ -52,7 +52,7 @@ tail: JBE move_5through8 CMPL BX, $16 JBE move_9through16 - CMPB internal∕cpu·X86+const_offset_x86_HasSSE2(SB), $1 + CMPB internal∕cpu·X86+const_offsetX86HasSSE2(SB), $1 JNE nosse2 CMPL BX, $32 JBE move_17through32 @@ -73,7 +73,7 @@ nosse2: */ forward: // If REP MOVSB isn't fast, don't use it - CMPB internal∕cpu·X86+const_offset_x86_HasERMS(SB), $1 // enhanced REP MOVSB/STOSB + CMPB internal∕cpu·X86+const_offsetX86HasERMS(SB), $1 // enhanced REP MOVSB/STOSB JNE fwdBy4 // Check alignment diff --git a/src/runtime/memmove_amd64.s b/src/runtime/memmove_amd64.s index c5385a3d43..b4243a833b 100644 --- a/src/runtime/memmove_amd64.s +++ b/src/runtime/memmove_amd64.s @@ -84,7 +84,7 @@ forward: JLS move_256through2048 // If REP MOVSB isn't fast, don't use it - CMPB internal∕cpu·X86+const_offset_x86_HasERMS(SB), $1 // enhanced REP MOVSB/STOSB + CMPB internal∕cpu·X86+const_offsetX86HasERMS(SB), $1 // enhanced REP MOVSB/STOSB JNE fwdBy8 // Check alignment diff --git a/src/runtime/vlop_arm.s b/src/runtime/vlop_arm.s index 8df13abd98..729653488f 100644 --- a/src/runtime/vlop_arm.s +++ b/src/runtime/vlop_arm.s @@ -44,7 +44,7 @@ // the RET instruction will clobber R12 on nacl, and the compiler's register // allocator needs to know. TEXT runtime·udiv(SB),NOSPLIT|NOFRAME,$0 - MOVBU internal∕cpu·ARM+const_offset_arm_HasIDIVA(SB), Ra + MOVBU internal∕cpu·ARM+const_offsetARMHasIDIVA(SB), Ra CMP $0, Ra BNE udiv_hardware