This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[PATCH, i386] Committed: Fix PR target/32660
- From: Uros Bizjak <ubizjak at gmail dot com>
- To: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sat, 07 Jul 2007 11:22:49 +0200
- Subject: [PATCH, i386] Committed: Fix PR target/32660
Hello!
The PR is fixed by backporting the patch that implements missing
patterns from mainline.
Bootstrapped/regression tested on x86_64, applied to gcc-4_1 branch.
Will apply to gcc-4_2 branch in a couple of days.
2007-07-07 Uros Bizjak <ubizjak@gmail.com>
PR target/32660
Backport from mainline.
* config/i386/sse.md (*vec_extractv2di_1_sse2): New.
(*vec_extractv2di_1_sse): New.
Uros.
Index: config/i386/sse.md
===================================================================
--- config/i386/sse.md (revision 126435)
+++ config/i386/sse.md (working copy)
@@ -3483,6 +3483,35 @@
operands[1] = gen_rtx_REG (DImode, REGNO (operands[1]));
})
+(define_insn "*vec_extractv2di_1_sse2"
+ [(set (match_operand:DI 0 "nonimmediate_operand" "=m,x,x")
+ (vec_select:DI
+ (match_operand:V2DI 1 "nonimmediate_operand" "x,0,o")
+ (parallel [(const_int 1)])))]
+ "TARGET_SSE2 && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
+ "@
+ movhps\t{%1, %0|%0, %1}
+ psrldq\t{$8, %0|%0, 8}
+ movq\t{%H1, %0|%0, %H1}"
+ [(set_attr "type" "ssemov,sseishft,ssemov")
+ (set_attr "memory" "*,none,*")
+ (set_attr "mode" "V2SF,TI,TI")])
+
+;; Not sure this is ever used, but it doesn't hurt to have it. -aoliva
+(define_insn "*vec_extractv2di_1_sse"
+ [(set (match_operand:DI 0 "nonimmediate_operand" "=m,x,x")
+ (vec_select:DI
+ (match_operand:V2DI 1 "nonimmediate_operand" "x,x,o")
+ (parallel [(const_int 1)])))]
+ "!TARGET_SSE2 && TARGET_SSE
+ && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
+ "@
+ movhps\t{%1, %0|%0, %1}
+ movhlps\t{%1, %0|%0, %1}
+ movlps\t{%H1, %0|%0, %H1}"
+ [(set_attr "type" "ssemov")
+ (set_attr "mode" "V2SF,V4SF,V2SF")])
+
(define_insn "*vec_dupv4si"
[(set (match_operand:V4SI 0 "register_operand" "=Y,x")
(vec_duplicate:V4SI