+2011-06-11 Uros Bizjak <ubizjak@gmail.com>
+
+ * config/i386/sse.md (vec_dupv4sf): Correct mode of forced register.
+ (*vec_dupv2df): Rename from vec_dupv2df.
+ (vec_dupv2df): New expander.
+
+2011-06-11 Uros Bizjak <ubizjak@gmail.com>
+
+ * config/i386/sse.md (AVX_VEC_DUP_MODE): Rename from AVX256MODE24P.
+
+2011-06-11 Uros Bizjak <ubizjak@gmail.com>
+
+ * config/i386/i386.md: Use default value in "isa" attribute.
+ * config/i386/sse.md: Ditto.
+ * config/i386/mmx.md: Ditto.
+
2011-06-10 Wei Guozhi <carrot@google.com>
PR target/45335
2011-06-10 Jan Hubicka <jh@suse.cz>
* opts.c (default_options): Enlist OPT_finline_functions_called_once.
- * common.opt (flag_inline_functions_called_once):Do not initialize to 1.
+ * common.opt (flag_inline_functions_called_once): Do not
+ initialize to 1.
2011-06-10 Jan Hubicka <jh@suse.cz>
* ipa-cp.c (ipcp_versionable_function_p): Thunks are not versionable.
- (ipcp_initialize_node_lattices): Do not deal with aliases; Do not try to
- propagate through thunks.
+ (ipcp_initialize_node_lattices): Do not deal with aliases;
+ Do not try to propagate through thunks.
(ipcp_change_tops_to_bottom): Do not deal with aliases.
2011-06-10 Jan Hubicka <jh@suse.cz>
- * ipa-prop.c (ipa_write_node_info): Stream jump functions for indirect
- calls.
+ * ipa-prop.c (ipa_write_node_info): Stream jump functions
+ for indirect calls.
(ipa_read_node_info): Likewise.
2011-06-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
(gimple_expand_builtin_pow): Minor cleanup.
(gimple_expand_builtin_cabs): New.
(execute_cse_sincos): Add case for BUILT_IN_CABS.
-
+
2011-06-10 Jan Hubicka <jh@suse.cz>
* ipa-cp.c (ipcp_versionable_function_p): Aliases are not versionable.
2011-06-10 Richard Guenther <rguenther@suse.de>
- * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine):
- Scan stmts forward when combining, visit inserted stmts when
- a stmt was changed.
+ * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Scan stmts
+ forward when combining, visit inserted stmts when a stmt was changed.
2011-06-10 Paolo Carlini <paolo.carlini@oracle.com>
2011-06-10 Ramana Radhakrishnan <ramana.radhakrishnan@linaro.org>
Richard Earnshaw <rearnsha@arm.com>
- * config/arm/arm.c (const_ok_for_op): Check to see
- if mvn can be used.
+ * config/arm/arm.c (const_ok_for_op): Check to see if mvn can be used.
* config/arm/vfp.md (*arm_movdi_vfp): Delete.
(*thumb2_movdi_vfp): Delete.
(*arm_movdi_vfp_cortexa8): Delete.
(cgraph_for_node_thunks_and_aliases): New function.
(cgraph_for_node_and_aliases): New function.
(cgraph_make_node_local_1): Break out from ...
- (cgraph_make_node_local) ... here; use cgraph_for_node_thunks_and_aliases.
+ (cgraph_make_node_local) ... here; use
+ cgraph_for_node_thunks_and_aliases.
(cgraph_set_nothrow_flag_1): Break out from ...
- (cgraph_set_nothrow_flag) ... here; use cgraph_for_node_thunks_and_aliases.
+ (cgraph_set_nothrow_flag) ... here;
+ use cgraph_for_node_thunks_and_aliases.
(cgraph_set_const_flag_1): Break out from ...
- (cgraph_set_const_flag) ... here; use cgraph_for_node_thunks_and_aliases.
+ (cgraph_set_const_flag) ... here;
+ use cgraph_for_node_thunks_and_aliases.
(cgraph_set_pure_flag_1): Break out from ...
- (cgraph_set_pure_flag) ... here; use cgraph_for_node_thunks_and_aliases.
+ (cgraph_set_pure_flag) ... here;
+ use cgraph_for_node_thunks_and_aliases.
(cgraph_propagate_frequency_1): Break out from ...
- (cgraph_propagate_frequency) ... here; use cgraph_for_node_thunks_and_aliases.
+ (cgraph_propagate_frequency) ... here; use
+ cgraph_for_node_thunks_and_aliases.
(cgraph_used_from_object_file_p): Do not care about aliases.
- (cgraph_not_only_called_directly_p_1, cgraph_only_called_directly_p): New functions.
+ (cgraph_not_only_called_directly_p_1, cgraph_only_called_directly_p):
+ New functions.
(collect_callers_of_node_1, collect_callers_of_node): New functions.
2011-06-10 Hans-Peter Nilsson <hp@axis.com>
2011-06-10 Hans-Peter Nilsson <hp@axis.com>
PR rtl-optimization/49154
- * ira-costs.c (setup_regno_cost_classes_by_mode): If there
- already is a matching slot in the hashtable, assign it to
- classes_ptr.
+ * ira-costs.c (setup_regno_cost_classes_by_mode): If there already
+ is a matching slot in the hashtable, assign it to classes_ptr.
PR rtl-optimization/49154
* doc/tm.texi.in (Register Classes): Document rule for the narrowest
register classes.
* doc/tm.texi: Regenerate.
->>>>>>> .r174929
2011-06-09 Kaz Kojima <kkojima@gcc.gnu.org>
PR target/49307
;; Mix-n-match
(define_mode_iterator AVX256MODE2P [V8SI V8SF V4DF])
-(define_mode_iterator AVX256MODE24P [V8SI V8SF V4DI V4DF])
(define_mode_iterator FMAMODE [SF DF V4SF V2DF V8SF V4DF])
"TARGET_SSE"
{
if (!TARGET_AVX)
- operands[1] = force_reg (V4SFmode, operands[1]);
+ operands[1] = force_reg (SFmode, operands[1]);
})
(define_insn "*vec_dupv4sf_avx"
(set_attr "prefix" "orig,vex,orig,vex,maybe_vex,orig,orig,vex,maybe_vex")
(set_attr "mode" "DF,DF,V1DF,V1DF,V1DF,V2DF,V1DF,V1DF,V1DF")])
+(define_expand "vec_dupv2df"
+ [(set (match_operand:V2DF 0 "register_operand" "")
+ (vec_duplicate:V2DF
+ (match_operand:DF 1 "nonimmediate_operand" "")))]
+ "TARGET_SSE2"
+{
+ if (!TARGET_SSE3)
+ operands[1] = force_reg (DFmode, operands[1]);
+})
+
(define_insn "*vec_dupv2df_sse3"
[(set (match_operand:V2DF 0 "register_operand" "=x")
(vec_duplicate:V2DF
(set_attr "prefix" "maybe_vex")
(set_attr "mode" "DF")])
-(define_insn "vec_dupv2df"
+(define_insn "*vec_dupv2df"
[(set (match_operand:V2DF 0 "register_operand" "=x")
(vec_duplicate:V2DF
(match_operand:DF 1 "register_operand" "0")))]
(set_attr "prefix" "vex")
(set_attr "mode" "OI")])
+;; Modes handled by AVX vec_dup patterns.
+(define_mode_iterator AVX_VEC_DUP_MODE
+ [V8SI V8SF V4DI V4DF])
+
(define_insn "vec_dup<mode>"
- [(set (match_operand:AVX256MODE24P 0 "register_operand" "=x,x")
- (vec_duplicate:AVX256MODE24P
+ [(set (match_operand:AVX_VEC_DUP_MODE 0 "register_operand" "=x,x")
+ (vec_duplicate:AVX_VEC_DUP_MODE
(match_operand:<ssescalarmode> 1 "nonimmediate_operand" "m,?x")))]
"TARGET_AVX"
"@
(set_attr "mode" "V8SF")])
(define_split
- [(set (match_operand:AVX256MODE24P 0 "register_operand" "")
- (vec_duplicate:AVX256MODE24P
+ [(set (match_operand:AVX_VEC_DUP_MODE 0 "register_operand" "")
+ (vec_duplicate:AVX_VEC_DUP_MODE
(match_operand:<ssescalarmode> 1 "register_operand" "")))]
"TARGET_AVX && reload_completed"
- [(set (match_dup 2) (vec_duplicate:<ssehalfvecmode> (match_dup 1)))
- (set (match_dup 0) (vec_concat:AVX256MODE24P (match_dup 2) (match_dup 2)))]
+ [(set (match_dup 2)
+ (vec_duplicate:<ssehalfvecmode> (match_dup 1)))
+ (set (match_dup 0)
+ (vec_concat:AVX_VEC_DUP_MODE (match_dup 2) (match_dup 2)))]
"operands[2] = gen_rtx_REG (<ssehalfvecmode>mode, REGNO (operands[0]));")
(define_insn "avx_vbroadcastf128_<mode>"
projects / gcc.git / commitdiff