re PR target/47840 (incorrect _mm256_insert_epi{32,64} implementations)

	PR target/47840
	* config/i386/avxintrin.h (_mm256_insert_epi32): Use _mm_insert_epi32.
	(_mm256_insert_epi64): Use _mm_insert_epi64.

From-SVN: r170371

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 47bca58aad14d59fa584e231a5664f836fa45363..89113ffbfd7b5d852e57dee9bb0cea39296a65a7 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2011-02-21  Uros Bizjak  <ubizjak@gmail.com>
+
+       PR target/47840
+       * config/i386/avxintrin.h (_mm256_insert_epi32): Use _mm_insert_epi32.
+       (_mm256_insert_epi64): Use _mm_insert_epi64.
+
 2011-02-21  Anatoly Sokolov  <aesok@post.ru>
 
        * config/stormy16/stormy16.h (GO_IF_MODE_DEPENDENT_ADDRESS): Remove.
@@ -15,7 +21,7 @@
        Do not stream DECL_INITIAL for TRANSLATION_UNIT_DECLs.
        (lto_input_ts_block_tree_pointers): Hook a BLOCK into the
        TUs context.
-       * lto-streamer-out.c (lto_output_ts_decl_common_tree_pointers): 
+       * lto-streamer-out.c (lto_output_ts_decl_common_tree_pointers):
        Do not stream DECL_INITIAL for TRANSLATION_UNIT_DECLs.
 
 2011-02-20  Richard Guenther  <rguenther@suse.de>
@@ -23,7 +29,7 @@
        PR lto/47822
        * tree.c (free_lang_data_in_decl): Clean builtins from
        the TU decl BLOCK_VARS.
- 
+
 2011-02-19  Alexandre Oliva  <aoliva@redhat.com>
 
        PR debug/47620
@@ -131,7 +137,7 @@
 2011-02-17  Iain Sandoe  <iains@gcc.gnu.org>
 
        * config/darwin-c.c (darwin_cpp_builtins): Define __OBJC2__ for
-       objc_abi == 2.  
+       objc_abi == 2.
        * config/darwin.c (output_objc_section_asm_op): Added support for
        ABI v1 and v2.
        (is_objc_metadata): New.
@@ -187,12 +193,10 @@
        (copy_arguments_for_versioning): Drop redundant get_var_ann.
        * ipa-inline.c (compute_inline_parameters): Do not compute
        disregard_inline_limits here.
-       are not available.
        (compute_inlinable_for_current, pass_inlinable): New.
        (pass_inline_parameters): Require PROP_referenced_vars.
        * cgraphunit.c (cgraph_process_new_functions): Don't run
-       compute_inline_parameters explicitly unless function is in
-       SSA form.
+       compute_inline_parameters explicitly unless function is in SSA form.
        (cgraph_analyze_function): Set .disregard_inline_limits.
        * tree-sra.c (convert_callers): Compute inliner parameters
        only for functions already in SSA form.

diff --git a/gcc/config/i386/avxintrin.h b/gcc/config/i386/avxintrin.h
index 70bfce1a27462287ebddb038e9117050c1522da1..8055cc607224e8e69d9ac1b235f5d3dbd6be6216 100644 (file)
--- a/gcc/config/i386/avxintrin.h
+++ b/gcc/config/i386/avxintrin.h
@@ -737,7 +737,7 @@ extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __arti
 _mm256_insert_epi32 (__m256i __X, int __D, int const __N)
 {
   __m128i __Y = _mm256_extractf128_si256 (__X, __N >> 2);
-  __Y = _mm_insert_epi16 (__Y, __D, __N % 4);
+  __Y = _mm_insert_epi32 (__Y, __D, __N % 4);
   return _mm256_insertf128_si256 (__X, __Y, __N >> 2);
 }
 
@@ -762,7 +762,7 @@ extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __arti
 _mm256_insert_epi64 (__m256i __X, int __D, int const __N)
 {
   __m128i __Y = _mm256_extractf128_si256 (__X, __N >> 1);
-  __Y = _mm_insert_epi16 (__Y, __D, __N % 2);
+  __Y = _mm_insert_epi64 (__Y, __D, __N % 2);
   return _mm256_insertf128_si256 (__X, __Y, __N >> 1);
 }
 #endif