diff mbox

[PowerPC] Fix vector pack insns for little endian

Message ID 1374699910.3633.138.camel@gnopaine
State New
Headers show

Commit Message

Bill Schmidt July 24, 2013, 9:05 p.m. UTC 
In little endian mode, we need to reverse the order of operands for
vector pack operations.  This corrects 20 failures in the test suite in
LE mode.

Patch based on work by Anton Blanchard; I simplified it and generalized
it to all the vector pack insns.

Bootstrapped and tested on powerpc64-unknown-linux-gnu with no new
regressions.  Ok for trunk?

Thanks,
Bill


2013-07-24  Bill Schmidt  <wschmidt@linux.ibm.com>
	    Anton Blanchard <anton@au1.ibm.com>

	* config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
	(altivec_vpks<VI_char>ss): Likewise.
	(altivec_vpks<VI_char>us): Likewise.
	(altivec_vpku<VI_char>us): Likewise.
	(altivec_vpku<VI_char>um): Likewise.

Comments

David Edelsohn July 24, 2013, 9:36 p.m. UTC | #1 
On Wed, Jul 24, 2013 at 5:05 PM, Bill Schmidt
<wschmidt@linux.vnet.ibm.com> wrote:
> In little endian mode, we need to reverse the order of operands for
> vector pack operations.  This corrects 20 failures in the test suite in
> LE mode.
>
> Patch based on work by Anton Blanchard; I simplified it and generalized
> it to all the vector pack insns.
>
> Bootstrapped and tested on powerpc64-unknown-linux-gnu with no new
> regressions.  Ok for trunk?
>
> Thanks,
> Bill
>
>
> 2013-07-24  Bill Schmidt  <wschmidt@linux.ibm.com>
>             Anton Blanchard <anton@au1.ibm.com>
>
>         * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
>         (altivec_vpks<VI_char>ss): Likewise.
>         (altivec_vpks<VI_char>us): Likewise.
>         (altivec_vpku<VI_char>us): Likewise.
>         (altivec_vpku<VI_char>um): Likewise.

Okay.

Thanks, David
diff mbox

Patch

Index: gcc/config/rs6000/altivec.md
===================================================================
--- gcc/config/rs6000/altivec.md	(revision 201207)
+++ gcc/config/rs6000/altivec.md	(working copy)
@@ -1047,7 +1047,13 @@ 
                       (match_operand:V4SI 2 "register_operand" "v")]
 		     UNSPEC_VPKPX))]
   "TARGET_ALTIVEC"
-  "vpkpx %0,%1,%2"
+  "*
+  {
+    if (BYTES_BIG_ENDIAN)
+      return \"vpkpx %0,%1,%2\";
+    else
+      return \"vpkpx %0,%2,%1\";
+  }"
   [(set_attr "type" "vecperm")])
 
 (define_insn "altivec_vpks<VI_char>ss"
@@ -1056,7 +1062,13 @@ 
 			    (match_operand:VP 2 "register_operand" "v")]
 			   UNSPEC_VPACK_SIGN_SIGN_SAT))]
   "<VI_unit>"
-  "vpks<VI_char>ss %0,%1,%2"
+  "*
+  {
+    if (BYTES_BIG_ENDIAN)
+      return \"vpks<VI_char>ss %0,%1,%2\";
+    else
+      return \"vpks<VI_char>ss %0,%2,%1\";
+  }"
   [(set_attr "type" "vecperm")])
 
 (define_insn "altivec_vpks<VI_char>us"
@@ -1065,7 +1077,13 @@ 
 			    (match_operand:VP 2 "register_operand" "v")]
 			   UNSPEC_VPACK_SIGN_UNS_SAT))]
   "<VI_unit>"
-  "vpks<VI_char>us %0,%1,%2"
+  "*
+  {
+    if (BYTES_BIG_ENDIAN)
+      return \"vpks<VI_char>us %0,%1,%2\";
+    else
+      return \"vpks<VI_char>us %0,%2,%1\";
+  }"
   [(set_attr "type" "vecperm")])
 
 (define_insn "altivec_vpku<VI_char>us"
@@ -1074,7 +1092,13 @@ 
 			    (match_operand:VP 2 "register_operand" "v")]
 			   UNSPEC_VPACK_UNS_UNS_SAT))]
   "<VI_unit>"
-  "vpku<VI_char>us %0,%1,%2"
+  "*
+  {
+    if (BYTES_BIG_ENDIAN)
+      return \"vpku<VI_char>us %0,%1,%2\";
+    else
+      return \"vpku<VI_char>us %0,%2,%1\";
+  }"
   [(set_attr "type" "vecperm")])
 
 (define_insn "altivec_vpku<VI_char>um"
@@ -1083,7 +1107,13 @@ 
 			    (match_operand:VP 2 "register_operand" "v")]
 			   UNSPEC_VPACK_UNS_UNS_MOD))]
   "<VI_unit>"
-  "vpku<VI_char>um %0,%1,%2"
+  "*
+  {
+    if (BYTES_BIG_ENDIAN)
+      return \"vpku<VI_char>um %0,%1,%2\";
+    else
+      return \"vpku<VI_char>um %0,%2,%1\";
+  }"
   [(set_attr "type" "vecperm")])
 
 (define_insn "*altivec_vrl<VI_char>"