This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
[PATCH] x86: further refine SSE check (SSE4a, SHA, GFNI)
- From: Jan Beulich <jbeulich at suse dot com>
- To: "binutils at sourceware dot org" <binutils at sourceware dot org>
- Cc: "H.J. Lu" <hjl dot tools at gmail dot com>
- Date: Mon, 9 Dec 2019 13:39:47 +0100
- Subject: [PATCH] x86: further refine SSE check (SSE4a, SHA, GFNI)
In ("x86: extend SSE check to PCLMULQDQ, AES, and GFNI insns") I went
both a little too far and not quite far enough:
- GFNI insns also have AVX512 variants, which also shouldn't get
diagnosed,
- SSE4a insns should get diagnosed just like SSE4.x ones,
- SHA insns should get diagnosed just like PCLMULQDQ or AES ones.
gas/
2018-12-XX Jan Beulich <jbeulich@suse.com>
* config/tc-i386.c (md_assemble): Extend SSE check conditional.
* testsuite/gas/i386/sse-check.s: Add SSE4a and SHA tests.
Extend GFNI tests.
* testsuite/gas/i386/sse-check.d: Adjust expectations.
* testsuite/gas/i386/sse-check-error.l,
testsuite/gas/i386/x86-64-sse-check-error.l: Likewise.
* testsuite/gas/i386/sse-check-warn.e: Likewise.
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -4286,14 +4286,17 @@ md_assemble (char *line)
if (sse_check != check_none
&& !i.tm.opcode_modifier.noavx
&& !i.tm.cpu_flags.bitfield.cpuavx
+ && !i.tm.cpu_flags.bitfield.cpuavx512f
&& (i.tm.cpu_flags.bitfield.cpusse
|| i.tm.cpu_flags.bitfield.cpusse2
|| i.tm.cpu_flags.bitfield.cpusse3
|| i.tm.cpu_flags.bitfield.cpussse3
|| i.tm.cpu_flags.bitfield.cpusse4_1
|| i.tm.cpu_flags.bitfield.cpusse4_2
+ || i.tm.cpu_flags.bitfield.cpusse4a
|| i.tm.cpu_flags.bitfield.cpupclmul
|| i.tm.cpu_flags.bitfield.cpuaes
+ || i.tm.cpu_flags.bitfield.cpusha
|| i.tm.cpu_flags.bitfield.cpugfni))
{
(sse_check == check_warning
--- a/gas/testsuite/gas/i386/sse-check-error.l
+++ b/gas/testsuite/gas/i386/sse-check-error.l
@@ -8,6 +8,8 @@
.*:23: Error: .*
.*:26: Error: .*
.*:29: Error: .*
+.*:32: Error: .*
+.*:35: Error: .*
GAS LISTING .*
@@ -42,17 +44,33 @@ GAS LISTING .*
.* Error: SSE instruction `pcmpgtq' is used
[ ]*20[ ]+C1
[ ]*21[ ]+
-[ ]*22[ ]+\# PCMUL instruction
-[ ]*23[ ]+\?\?\?\? 660F3A44 pclmulqdq \$-1,%xmm1,%xmm2
-.* Error: SSE instruction `pclmulqdq' is used
-[ ]*23[ ]+D1FF
+[ ]*22[ ]+\# SSE4a instruction
+[ ]*23[ ]+\?\?\?\? 660F78C0 extrq \$0, \$0, %xmm0
+.* Error: SSE instruction `extrq' is used
+[ ]*23[ ]+0000
[ ]*24[ ]+
-[ ]*25[ ]+\# AES instructions
-[ ]*26[ ]+\?\?\?\? 660F38DE aesdec %xmm1,%xmm2
-.* Error: SSE instruction `aesdec' is used
-[ ]*26[ ]+D1
+[ ]*25[ ]+\# PCMUL instruction
+[ ]*26[ ]+\?\?\?\? 660F3A44 pclmulqdq \$-1,%xmm1,%xmm2
+.* Error: SSE instruction `pclmulqdq' is used
+[ ]*26[ ]+D1FF
[ ]*27[ ]+
-[ ]*28[ ]+\# GFNI instructions
-[ ]*29[ ]+\?\?\?\? 660F38CF gf2p8mulb %xmm1,%xmm2
-.* Error: SSE instruction `gf2p8mulb' is used
+[ ]*28[ ]+\# AES instructions
+[ ]*29[ ]+\?\?\?\? 660F38DE aesdec %xmm1,%xmm2
+.* Error: SSE instruction `aesdec' is used
[ ]*29[ ]+D1
+[ ]*30[ ]+
+[ ]*31[ ]+\# SHA instruction
+[ ]*32[ ]+\?\?\?\? 0F38C8C0 sha1nexte %xmm0, %xmm0
+.* Error: SSE instruction `sha1nexte' is used
+[ ]*33[ ]+
+[ ]*34[ ]+\# GFNI instructions
+[ ]*35[ ]+\?\?\?\? 660F38CF gf2p8mulb %xmm1,%xmm2
+.* Error: SSE instruction `gf2p8mulb' is used
+[ ]*35[ ]+D1
+[ ]*36[ ]+\?\?\?\? 62F27D09 vgf2p8mulb %xmm0, %xmm0, %xmm0\{%k1\}
+[ ]*36[ ]+CFC0
+[ ]*37[ ]+\?\?\?\? 62F27D48 vgf2p8mulb %zmm0, %zmm0, %zmm0
+GAS LISTING .*
+
+
+[ ]*37[ ]+CFC0
--- a/gas/testsuite/gas/i386/sse-check-warn.e
+++ b/gas/testsuite/gas/i386/sse-check-warn.e
@@ -5,6 +5,8 @@
.*:16: Warning: SSE instruction `phaddw' is used
.*:19: Warning: SSE instruction `blendvpd' is used
.*:20: Warning: SSE instruction `pcmpgtq' is used
-.*:23: Warning: SSE instruction `pclmulqdq' is used
-.*:26: Warning: SSE instruction `aesdec' is used
-.*:29: Warning: SSE instruction `gf2p8mulb' is used
+.*:23: Warning: SSE instruction `extrq' is used
+.*:26: Warning: SSE instruction `pclmulqdq' is used
+.*:29: Warning: SSE instruction `aesdec' is used
+.*:32: Warning: SSE instruction `sha1nexte' is used
+.*:35: Warning: SSE instruction `gf2p8mulb' is used
--- a/gas/testsuite/gas/i386/sse-check.d
+++ b/gas/testsuite/gas/i386/sse-check.d
@@ -13,7 +13,11 @@ Disassembly of section .text:
[ ]*[a-f0-9]+: 66 0f 38 01 ca phaddw %xmm2,%xmm1
[ ]*[a-f0-9]+: 66 0f 38 15 c1 blendvpd %xmm0,%xmm1,%xmm0
[ ]*[a-f0-9]+: 66 0f 38 37 c1 pcmpgtq %xmm1,%xmm0
+[ ]*[a-f0-9]+: 66 0f 78 c0 00 00 extrq \$0x0,\$0x0,%xmm0
[ ]*[a-f0-9]+: 66 0f 3a 44 d1 ff pclmulqdq \$0xff,%xmm1,%xmm2
[ ]*[a-f0-9]+: 66 0f 38 de d1 aesdec %xmm1,%xmm2
+[ ]*[a-f0-9]+: 0f 38 c8 c0 sha1nexte %xmm0,%xmm0
[ ]*[a-f0-9]+: 66 0f 38 cf d1 gf2p8mulb %xmm1,%xmm2
+[ ]*[a-f0-9]+: 62 f2 7d 09 cf c0 vgf2p8mulb %xmm0,%xmm0,%xmm0\{%k1\}
+[ ]*[a-f0-9]+: 62 f2 7d 48 cf c0 vgf2p8mulb %zmm0,%zmm0,%zmm0
#pass
--- a/gas/testsuite/gas/i386/sse-check.s
+++ b/gas/testsuite/gas/i386/sse-check.s
@@ -19,11 +19,19 @@ _start:
blendvpd %xmm0,%xmm1,%xmm0
pcmpgtq %xmm1,%xmm0
+# SSE4a instruction
+ extrq $0, $0, %xmm0
+
# PCMUL instruction
pclmulqdq $-1,%xmm1,%xmm2
# AES instructions
aesdec %xmm1,%xmm2
+# SHA instruction
+ sha1nexte %xmm0, %xmm0
+
# GFNI instructions
gf2p8mulb %xmm1,%xmm2
+ vgf2p8mulb %xmm0, %xmm0, %xmm0{%k1}
+ vgf2p8mulb %zmm0, %zmm0, %zmm0
--- a/gas/testsuite/gas/i386/x86-64-sse-check-error.l
+++ b/gas/testsuite/gas/i386/x86-64-sse-check-error.l
@@ -8,6 +8,8 @@
.*:23: Error: .*
.*:26: Error: .*
.*:29: Error: .*
+.*:32: Error: .*
+.*:35: Error: .*
GAS LISTING .*
@@ -42,17 +44,33 @@ GAS LISTING .*
.* Error: SSE instruction `pcmpgtq' is used
[ ]*20[ ]+C1
[ ]*21[ ]+
-[ ]*22[ ]+\# PCMUL instruction
-[ ]*23[ ]+\?\?\?\? 660F3A44 pclmulqdq \$-1,%xmm1,%xmm2
-.* Error: SSE instruction `pclmulqdq' is used
-[ ]*23[ ]+D1FF
+[ ]*22[ ]+\# SSE4a instruction
+[ ]*23[ ]+\?\?\?\? 660F78C0 extrq \$0, \$0, %xmm0
+.* Error: SSE instruction `extrq' is used
+[ ]*23[ ]+0000
[ ]*24[ ]+
-[ ]*25[ ]+\# AES instructions
-[ ]*26[ ]+\?\?\?\? 660F38DE aesdec %xmm1,%xmm2
-.* Error: SSE instruction `aesdec' is used
-[ ]*26[ ]+D1
+[ ]*25[ ]+\# PCMUL instruction
+[ ]*26[ ]+\?\?\?\? 660F3A44 pclmulqdq \$-1,%xmm1,%xmm2
+.* Error: SSE instruction `pclmulqdq' is used
+[ ]*26[ ]+D1FF
[ ]*27[ ]+
-[ ]*28[ ]+\# GFNI instructions
-[ ]*29[ ]+\?\?\?\? 660F38CF gf2p8mulb %xmm1,%xmm2
-.* Error: SSE instruction `gf2p8mulb' is used
+[ ]*28[ ]+\# AES instructions
+[ ]*29[ ]+\?\?\?\? 660F38DE aesdec %xmm1,%xmm2
+.* Error: SSE instruction `aesdec' is used
[ ]*29[ ]+D1
+[ ]*30[ ]+
+[ ]*31[ ]+\# SHA instruction
+[ ]*32[ ]+\?\?\?\? 0F38C8C0 sha1nexte %xmm0, %xmm0
+.* Error: SSE instruction `sha1nexte' is used
+[ ]*33[ ]+
+[ ]*34[ ]+\# GFNI instructions
+[ ]*35[ ]+\?\?\?\? 660F38CF gf2p8mulb %xmm1,%xmm2
+.* Error: SSE instruction `gf2p8mulb' is used
+[ ]*35[ ]+D1
+[ ]*36[ ]+\?\?\?\? 62F27D09 vgf2p8mulb %xmm0, %xmm0, %xmm0\{%k1\}
+[ ]*36[ ]+CFC0
+[ ]*37[ ]+\?\?\?\? 62F27D48 vgf2p8mulb %zmm0, %zmm0, %zmm0
+GAS LISTING .*
+
+
+[ ]*37[ ]+CFC0