re PR target/48708 (Invalid V2DI vector set insn generated)

	PR target/48708
	* config/i386/i386.c (ix86_expand_vector_set) <V2DImode>: Generate
	vec_extract and vec_concat for non-SSE4_1 targets.

testsuite/ChangeLog:

	PR target/48708
	* gcc.target/i386/pr48708.c: New test.

From-SVN: r172821

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index a414107f799fe69a4fa6e2322b4ba9094f58bd66..2f4316cc0bc224ab62ace48ceab55ee5d92c96e5 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2011-04-21  Uros Bizjak  <ubizjak@gmail.com>
+
+       PR target/48708
+       * config/i386/i386.c (ix86_expand_vector_set) <V2DImode>: Generate
+       vec_extract and vec_concat for non-SSE4_1 targets.
+
 2011-04-21  Richard Guenther  <rguenther@suse.de>
 
        * tree-ssa-alias.c (ref_maybe_used_by_stmt_p): Handle
@@ -16,8 +22,7 @@
        * config/bfin/bfin.h (bfin_cc_rtx, bfin_rets_rtx): Use rtx, not
        struct rtx_def *.
        * config/cris/cris-protos.h (STDIO_INCLUDED): Don't define.
-       * config/h8300/h8300.h (struct cum_arg): Use rtx, not struct
-       rtx_def *.
+       * config/h8300/h8300.h (struct cum_arg): Use rtx, not struct rtx_def *.
        * config/iq2000/iq2000.h (struct iq2000_args): Use rtx, not struct
        rtx_def *.
        * config/m32c/m32c-protos.h (MM, UINT): Don't define.  Expand
@@ -26,18 +31,15 @@
        rtx, not struct rtx_def *.
        * config/pa/pa-protos.h (return_addr_rtx): Use rtx, not struct
        rtx_def *.
-       * config/pa/pa.h (hppa_pic_save_rtx): Use rtx, not struct rtx_def
-       *.
-       * config/pdp11/pdp11.h (cc0_reg_rtx): Use rtx, not struct rtx_def
-       *.
+       * config/pa/pa.h (hppa_pic_save_rtx): Use rtx, not struct rtx_def *.
+       * config/pdp11/pdp11.h (cc0_reg_rtx): Use rtx, not struct rtx_def *.
        * config/sh/sh-protos.h (sfunc_uses_reg, get_fpscr_rtx): Use rtx,
        not struct rtx_def *.
        * config/sh/sh.h (sh_compare_op0, sh_compare_op1): Use rtx, not
        struct rtx_def *.
        * config/spu/spu-protos.h (spu_float_const): Use rtx, not struct
        rtx_def *.
-       * config/spu/spu.c (spu_float_const): Use rtx, not struct rtx_def
-       *.
+       * config/spu/spu.c (spu_float_const): Use rtx, not struct rtx_def *.
 
 2011-04-21  Richard Sandiford  <richard.sandiford@linaro.org>
 
@@ -183,8 +185,8 @@
        * config/mn10300/mn10300-protos.h (mn10300_legitimate_constant_p):
        Delete.
        * config/mn10300/mn10300.h (LEGITIMATE_CONSTANT_P): Likewise.
-       * config/mn10300/mn10300.c (mn10300_legitimate_constant_p):
-       Make static.  Add a mode argument.
+       * config/mn10300/mn10300.c (mn10300_legitimate_constant_p): Make
+       static.  Add a mode argument.
        (TARGET_LEGITIMATE_CONSTANT_P): Define.
 
        * config/moxie/moxie.h (LEGITIMATE_CONSTANT_P): Delete.
@@ -284,8 +286,7 @@
        (mips_secondary_reload_class): Likewise.
        * config/pa/pa.c (TARGET_CANNOT_FORCE_CONST_MEM): Redefine to...
        (pa_cannot_force_const_mem): ...this new function.
-       * config/rs6000/rs6000.c (TARGET_CANNOT_FORCE_CONST_MEM): Reefine
-       to...
+       * config/rs6000/rs6000.c (TARGET_CANNOT_FORCE_CONST_MEM): Reefine to...
        (rs6000_cannot_force_const_mem): ...this new function.
        * config/s390/s390.c (s390_cannot_force_const_mem): Add a mode
        argument.
@@ -386,8 +387,7 @@
 2011-04-20  Chung-Lin Tang  <cltang@codesourcery.com>
 
        * config/arm/arm.c (arm_legitimize_reload_address): For NEON
-       quad-word modes, reduce to 9-bit index range when above 1016
-       limit.
+       quad-word modes, reduce to 9-bit index range when above 1016 limit.
 
 2011-04-20  Andrew Stubbs  <ams@codesourcery.com>
 
@@ -396,8 +396,8 @@
 
 2011-04-20  Kai Tietz  <ktietz@redhat.com>
 
-       * fold-const.c (fold_binary_loc): Add handling for
-       (X & ~Y) | (~X & Y) and (X && !Y) | (!X && Y) optimization to (X ^ Y).
+       * fold-const.c (fold_binary_loc): Add handling for (X & ~Y) | (~X & Y)
+       and (X && !Y) | (!X && Y) optimization to (X ^ Y).
 
 2011-04-20  Andrew Stubbs  <ams@codesourcery.com>
 
@@ -445,8 +445,7 @@
 
        PR middle-end/48695
        * tree-ssa-alias.c (aliasing_component_refs_p): Compute base
-       objects and types here.  Adjust for their offset before
-       comparing.
+       objects and types here.  Adjust for their offset before comparing.
 
 2011-04-20  Richard Sandiford  <richard.sandiford@linaro.org>
 

diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index b55c67bb56490e3daecabef0ee078310f2eb3355..558eff119367211f1ce21671d26214e7be478bfb 100644 (file)
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -31484,10 +31484,19 @@ ix86_expand_vector_set (bool mmx_ok, rtx target, rtx val, int elt)
       break;
 
     case V2DImode:
-      use_vec_merge = TARGET_SSE4_1;
+      use_vec_merge = TARGET_SSE4_1 && TARGET_64BIT;
       if (use_vec_merge)
        break;
 
+      tmp = gen_reg_rtx (GET_MODE_INNER (mode));
+      ix86_expand_vector_extract (false, tmp, target, 1 - elt);
+      if (elt == 0)
+       tmp = gen_rtx_VEC_CONCAT (mode, tmp, val);
+      else
+       tmp = gen_rtx_VEC_CONCAT (mode, val, tmp);
+      emit_insn (gen_rtx_SET (VOIDmode, target, tmp));
+      return;
+
     case V2DFmode:
       {
        rtx op0, op1;

diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 5917c5ba638cac869e5e43e24a35ccdac9237d8a..531984693452d17d6599dd0d4ffe9c1bd365ab1d 100644 (file)
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2011-04-21  Uros Bizjak  <ubizjak@gmail.com>
+
+       PR target/48708
+       * gcc.target/i386/pr48708.c: New test.
+
 2011-04-21  Richard Sandiford  <richard.sandiford@linaro.org>
 
        * gcc.dg/vect/vect-119.c: New test.

diff --git a/gcc/testsuite/gcc.target/i386/pr48708.c b/gcc/testsuite/gcc.target/i386/pr48708.c
new file mode 100644 (file)
index 0000000..355c2b2
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr48708.c
@@ -0,0 +1,15 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -msse2" } */
+
+#include <emmintrin.h>
+
+typedef long long T __attribute__((may_alias));
+struct S { __m128i d; };
+
+__m128i
+foo (long long *x, struct S *y, __m128i *z)
+{
+  struct S s = *y;
+  ((T *) &s.d)[0] = *x;
+  return _mm_cmpeq_epi16 (s.d, *z);
+}