From: Jan Beulich <jbeulich@suse.com>
Date: Fri, 3 Feb 2023 07:22:35 +0000 (+0100)
Subject: x86-64: respect {nooptimize} when building VEX prefix
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=33740f219e4eec3ee0f428a76390b9afc8140ad6;p=binutils-gdb.git

x86-64: respect {nooptimize} when building VEX prefix

Swapping operands for commutative insns occurs outside of
optimize_encoding() and hence needs explicit checking for a request to
avoid any optimizations.
---

diff --git a/gas/config/tc-i386.c b/gas/config/tc-i386.c
index c3aa40b2a02..d5a24ae1637 100644
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -3624,7 +3624,8 @@ build_vex_prefix (const insn_template *t)
       && i.reg_operands == i.operands - i.imm_operands
       && i.tm.opcode_modifier.vex
       && i.tm.opcode_modifier.commutative
-      && (i.tm.opcode_modifier.sse2avx || optimize > 1)
+      && (i.tm.opcode_modifier.sse2avx
+	  || (optimize > 1 && !i.no_optimize))
       && i.rex == REX_B
       && i.vex.register_specifier
       && !(i.vex.register_specifier->reg_flags & RegRex))
diff --git a/gas/testsuite/gas/i386/x86-64-optimize-4.d b/gas/testsuite/gas/i386/x86-64-optimize-4.d
index d25d24d3c6f..ed01f2150c4 100644
--- a/gas/testsuite/gas/i386/x86-64-optimize-4.d
+++ b/gas/testsuite/gas/i386/x86-64-optimize-4.d
@@ -9,6 +9,8 @@ Disassembly of section .text:
 
 0+ <_start>:
  +[a-f0-9]+:	a9 7f 00 00 00       	test   \$0x7f,%eax
+ +[a-f0-9]+:	c4 c1 61 db e4       	vpand  %xmm12,%xmm3,%xmm4
+ +[a-f0-9]+:	c5 91 db e2          	vpand  %xmm2,%xmm13,%xmm4
  +[a-f0-9]+:	62 f1 7d 28 6f d1    	vmovdqa32 %ymm1,%ymm2
  +[a-f0-9]+:	62 f1 fd 28 6f d1    	vmovdqa64 %ymm1,%ymm2
  +[a-f0-9]+:	62 f1 7f 08 6f d1    	vmovdqu8 %xmm1,%xmm2
diff --git a/gas/testsuite/gas/i386/x86-64-optimize-4.s b/gas/testsuite/gas/i386/x86-64-optimize-4.s
index a3f69a3f775..97264e25d3f 100644
--- a/gas/testsuite/gas/i386/x86-64-optimize-4.s
+++ b/gas/testsuite/gas/i386/x86-64-optimize-4.s
@@ -5,6 +5,9 @@
 _start:
 	{nooptimize} testl $0x7f, %eax
 
+	{nooptimize} vpand	%xmm12, %xmm3, %xmm4
+	{nooptimize} vpand	%xmm2, %xmm13, %xmm4
+
 	{nooptimize} vmovdqa32	%ymm1, %ymm2
 	{nooptimize} vmovdqa64	%ymm1, %ymm2
 	{nooptimize} vmovdqu8	%xmm1, %xmm2