x86: also optimize KXOR{D,Q} and KANDN{D,Q}

These can be converted to 2-byte VEX encoding when both source registers
are the same, by using KXORW / KANDNW as replacement.

6 files changed, 43 insertions, 1 deletions

diff --git a/gas/ChangeLog b/gas/ChangeLog
index 4472821..9682f66 100644
--- a/gas/ChangeLog
+++ b/gas/ChangeLog
@@ -1,5 +1,14 @@
 2018-07-31  Jan Beulich  <jbeulich@suse.com>
 
+	* config/tc-i386.c (optimize_encoding): Also handle kandnd,
+	kandnq, kxord, and kxorq.
+	* testsuite/gas/i386/optimize-1.s: Add kandn and kxor tests.
+	* testsuite/gas/i386/optimize-1.d,
+	testsuite/gas/i386/optimize-4.d,
+	testsuite/gas/i386/optimize-5.d: Adjust expectations.
+
+2018-07-31  Jan Beulich  <jbeulich@suse.com>
+
 	* config/tc-i386.c (check_VecOperands): Convert masking handling
 	to switch(), to deal with DYNAMIC_MASKING.
 
diff --git a/gas/config/tc-i386.c b/gas/config/tc-i386.c
index 8761e97..e07056f 100644
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -3942,7 +3942,11 @@ optimize_encoding (void)
 		|| i.tm.base_opcode == 0x66f8
 		|| i.tm.base_opcode == 0x66f9
 		|| i.tm.base_opcode == 0x66fa
-		|| i.tm.base_opcode == 0x66fb)
+		|| i.tm.base_opcode == 0x66fb
+		|| i.tm.base_opcode == 0x42
+		|| i.tm.base_opcode == 0x6642
+		|| i.tm.base_opcode == 0x47
+		|| i.tm.base_opcode == 0x6647)
 	       && i.tm.extension_opcode == None))
     {
       /* Optimize: -O2:
@@ -3973,6 +3977,12 @@ optimize_encoding (void)
 	     EVEX VOP %ymmM, %ymmM, %ymmN
 	       -> VEX vpxor %xmmM, %xmmM, %xmmN (M and N < 16)
 	       -> EVEX VOP %xmmM, %xmmM, %xmmN (M || N >= 16)
+	   VOP, one of kxord and kxorq:
+	     VEX VOP %kM, %kM, %kN
+	       -> VEX kxorw %kM, %kM, %kN
+	   VOP, one of kandnd and kandnq:
+	     VEX VOP %kM, %kM, %kN
+	       -> VEX kandnw %kM, %kM, %kN
        */
       if (is_evex_encoding (&i.tm))
 	{
@@ -3985,6 +3995,11 @@ optimize_encoding (void)
 	      i.tm.opcode_modifier.evex = 0;
 	    }
 	}
+      else if (i.tm.operand_types[0].bitfield.regmask)
+	{
+	  i.tm.base_opcode &= 0xff;
+	  i.tm.opcode_modifier.vexw = VEXW0;
+	}
       else
 	i.tm.opcode_modifier.vex = VEX128;
 
diff --git a/gas/testsuite/gas/i386/optimize-1.d b/gas/testsuite/gas/i386/optimize-1.d
index 33ce43c..4358c19 100644
--- a/gas/testsuite/gas/i386/optimize-1.d
+++ b/gas/testsuite/gas/i386/optimize-1.d
@@ -58,4 +58,8 @@ Disassembly of section .text:
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
 #pass
diff --git a/gas/testsuite/gas/i386/optimize-1.s b/gas/testsuite/gas/i386/optimize-1.s
index 21b9594..f61a176 100644
--- a/gas/testsuite/gas/i386/optimize-1.s
+++ b/gas/testsuite/gas/i386/optimize-1.s
@@ -66,3 +66,9 @@ _start:
 	vpsubq %ymm1, %ymm1, %ymm5{z}{%k7}
 	vpsubq %zmm1, %zmm1, %zmm5
 	vpsubq %ymm1, %ymm1, %ymm5
+
+	kxord %k1, %k1, %k5
+	kxorq %k1, %k1, %k5
+
+	kandnd %k1, %k1, %k5
+	kandnq %k1, %k1, %k5
diff --git a/gas/testsuite/gas/i386/optimize-4.d b/gas/testsuite/gas/i386/optimize-4.d
index c9f50f5..9f99dad 100644
--- a/gas/testsuite/gas/i386/optimize-4.d
+++ b/gas/testsuite/gas/i386/optimize-4.d
@@ -58,6 +58,10 @@ Disassembly of section .text:
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
  +[a-f0-9]+:	62 f1 f5 08 55 e9    	vandnpd %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	62 f1 f5 08 55 e9    	vandnpd %xmm1,%xmm1,%xmm5
 #pass
diff --git a/gas/testsuite/gas/i386/optimize-5.d b/gas/testsuite/gas/i386/optimize-5.d
index ba4601e..cfd0df0 100644
--- a/gas/testsuite/gas/i386/optimize-5.d
+++ b/gas/testsuite/gas/i386/optimize-5.d
@@ -58,6 +58,10 @@ Disassembly of section .text:
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	c5 f1 fb e9          	vpsubq %xmm1,%xmm1,%xmm5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 47 e9          	kxorw  %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
+ +[a-f0-9]+:	c5 f4 42 e9          	kandnw %k1,%k1,%k5
  +[a-f0-9]+:	62 f1 f5 08 55 e9    	vandnpd %xmm1,%xmm1,%xmm5
  +[a-f0-9]+:	62 f1 f5 08 55 e9    	vandnpd %xmm1,%xmm1,%xmm5
 #pass