]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
projects / thirdparty / gcc.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
regex_scanner.tcc: (_Scanner<>::_M_scan_normal...
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
01e127b1
JL		 12013-10-18 Jeff Law <law@redhat.com>
2
3 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
4
5 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
6 (register_jump_thread): Add "registered_jump_thread" debug counter support.
7 * dbgcnt.def (registered_jump_thread): New debug counter.
8
f41b8848
AM		 92013-10-18 Andrew MacLeod <amacleod@redhat.com>
10
11 * config/rs6000/rs6000.c: Include cgraph.h.
12
adc7a812
TJ		 132013-10-18 Teresa Johnson <tejohnson@google.com>
14
15 * tree-ssa-tail-merge.c (replace_block_by): Update edge
16 weights during merging.
17
377d9792
AM		 182013-10-18 Andrew MacLeod <amacleod@redhat.com>
19
20 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
21 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
22 * tree-ssa-operands.h: Remove prototype.
23 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
24 * gimple.c (virtual_operand_p): Relocate from gimple.c.
25 * gimple.h: Add prototype.
26 * gimple-ssa.h: Include tree-ssa-operands.h.
27 * tree-dump.c: Add tree-cfg.h to include list.
28 * tree-ssa-alias.c: Add ipa-reference.h to include list.
29 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
30 * config/i386/i386.c: Don't include tree-flow.h.
31 * config/rs6000/rs6000.c: Likewise.
32
8ad8afaf
JBG		 332013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
34
35 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
36
c0e081a9
RB		 372013-10-18 Richard Biener <rguenther@suse.de>
38
39 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
40 or TYPE_UNSIGNED of integral types.
41 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
42 NULL_TREE for zero-precision integral types.
43
054fceb5
JG		 442013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
45
46 * config/aarch64/arm_neon.h
47 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
48
70e8b89b
ST		 492013-10-17 Sriraman Tallam <tmsriram@google.com>
50
51 * opth-gen.awk: Define target_flags_explicit.
52
3170766c 532013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5845f602
MM		 54
55 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
56 fields to the reg_addr array that describes the valid addressing
57 mode for any register, general purpose registers, floating point
58 registers, and Altivec registers.
59 (FIRST_RELOAD_REG_CLASS): Likewise.
60 (LAST_RELOAD_REG_CLASS): Likewise.
61 (struct reload_reg_map_type): Likewise.
62 (reload_reg_map_type): Likewise.
63 (RELOAD_REG_VALID): Likewise.
64 (RELOAD_REG_MULTIPLE): Likewise.
65 (RELOAD_REG_INDEXED): Likewise.
66 (RELOAD_REG_OFFSET): Likewise.
67 (RELOAD_REG_PRE_INCDEC): Likewise.
68 (RELOAD_REG_PRE_MODIFY): Likewise.
69 (reg_addr): Likewise.
70 (mode_supports_pre_incdec_p): New helper functions to say whether
71 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
72 (mode_supports_pre_modify_p): Likewise.
73 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
74 print the valid address mode bits for each mode.
75 (rs6000_debug_print_mode): Likewise.
76 (rs6000_debug_reg_global): Likewise.
77 (rs6000_setup_reg_addr_masks): New function to set up the address
78 mask bits for each type.
79 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
80 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
81 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
82 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
83 PRE_MODIFY are supported.
3170766c
MM		 84 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
85 registers, instead of {src,dest}_av_p.
5845f602
MM		 86 (rs6000_print_options_internal): Tweak the debug output slightly.
87
82e33890
UB		 882013-10-17 Uros Bizjak <ubizjak@gmail.com>
89
90 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
91 isa attribute.
92
0645c1a2
AM		 932013-10-17 Andrew MacLeod <amacleod@redhat.com>
94
95 * tree-flow.h (struct omp_region): Move to omp-low.c.
96 Remove omp_ prototypes and variables.
97 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
98 (copy_var_decl): Relocate prototype from tree-flow.h.
99 * gimple.c (copy_var_decl): Relocate from omp-low.c.
100 * tree.h: Move prototype to omp-low.h.
101 * omp-low.h: New File. Relocate prototypes here.
102 * omp-low.c (struct omp_region): Make local here.
103 (root_omp_region): Make static.
104 (copy_var_decl) Move to gimple.c.
105 (new_omp_region): Make static.
106 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
107 * tree-cfg.c: Include omp-low.h.
108 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
109 * gimplify.c: Include omp-low.h.
110 * tree-parloops.c: Likewise.
111
6f1abb55
UB		 1122013-10-17 Uros Bizjak <ubizjak@gmail.com>
113
114 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
115 operands are in memory, prefer to force non-matched operand 1 to
116 the register.
117
1ddbbbc0
MM		 1182013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
119
120 PR target/58673
121 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
122 restrict TImode addresses to single indirect registers if both
123 -mquad-memory and -mvsx-timode are used.
124 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
125 we should emit load/store quad. Remove using %y for quad memory
126 addresses.
127
128 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
129 constraints to allow load/store quad on machines where TImode is
130 not allowed in VSX registers. Use 'n' instead of 'F' constraint
131 for TImode to load integer constants.
132
f541a481
KT		 1332013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
134
135 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
136
51bb310d
MS		 1372013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
138
139 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
140 handling of STACK_REG.
141
f4bed77b
RB		 1422013-10-17 Richard Biener <rguenther@suse.de>
143
144 PR tree-optimization/58143
145 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
146 New function.
147 (rewrite_to_defined_overflow): Likewise.
148 (move_computations_dom_walker::before_dom): Rewrite stmts
149 with undefined signed overflow that are not always executed
150 into unsigned arithmetic.
151
d358fb9d
MM		 1522013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
153
154 PR target/57756
155 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
156 explicit isa flag to be an options variable, instead of using
157 global_options_set. Remove define from rs6000.h.
158 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
159
160 * config/rs6000/rs6000.c (rs6000_option_override_internal):
161 Initialize rs6000_isa_flags_explicit.
162 (rs6000_function_specific_save): Add gcc_options* parameter, so
163 that the powerpc builds after the 2013-10-15 changes.
164 (rs6000_function_specific_restore): Likewise.
165
c71aed47
DD		 1662013-10-16 DJ Delorie <dj@redhat.com>
167
168 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
169 op is a REG before checking REGNO.
170 (rl78_alloc_physical_registers): Verify pattern is a SET before
171 checking SET_SRC.
172
90d3bd51
BS		 1732013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
174
175 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
176 endianness.
177 (vec_unpacks_lo_v4sf): Likewise.
178 (vec_unpacks_float_hi_v4si): Likewise.
179 (vec_unpacks_float_lo_v4si): Likewise.
180 (vec_unpacku_float_hi_v4si): Likewise.
181 (vec_unpacku_float_lo_v4si): Likewise.
182
de75c876
BS		 1832013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
184
185 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
186 (vsx_concat_v2sf): Likewise.
187
21216998
JG		 1882013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
189
190 * config/aarch64/aarch64.md
191 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
192
1ef88893 1932013-10-16 Andrew MacLeod <amacleod@redhat.com>
7ab4168e 194
1ef88893 195 PR tree-optimization/58697
7ab4168e 196 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1ef88893
AM		 197 estimated_loop_iterations_int.
198 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7ab4168e 199 (get_max_loop_iterations_int): New. HWINT version of
1ef88893
AM		 200 get_max_loop_iterations.
201 * cfgloop.h: Add prototypes.
202 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
7ab4168e 203 * loop-unroll.c (decide_peel_once_rolling): Call
1ef88893
AM		 204 get_estimated_loop_iterations_int.
205 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
206 * tree-ssa-loop-niter.h: Tweak prototypes.
207
c22df64f
DM		 2082013-10-16 David Malcolm <dmalcolm@redhat.com>
209
210 * gengtype-parse.c (struct_field_seq): Ignore access-control
211 keywords ("public:" etc).
212
f24bb080
MS		 2132013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
214
215 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
216 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
217
c281bf26
YR		 2182013-10-16 Yvan Roux <yvan.roux@linaro.org>
219
220 * config/arm/arm.opt (mlra): New option.
221 * config/arm/arm.c (arm_lra_p): New function.
222 (TARGET_LRA_P): Define.
223
5806f481
PM		 2242013-10-16 Paulo Matos <pmatos@broadcom.com>
225
226 * tree-core.h (tree_code_name): Remove.
227 * tree.h (get_tree_code_name): New prototype.
228 * tree.c (tree_code_name): Make static.
229 (get_tree_code_name): New function.
230 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
231 tree_class_check_failed, tree_range_check_failed,
232 tree_not_class_check_failed, omp_clause_check_failed,
233 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
234 wrapper get_tree_code_name instead of calling tree_code_name directly.
235 * tree-vrp.c (dump_asserts_for): Likewise.
236 * tree-dump.c (dequeue_and_dump): Likewise.
237 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
238 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
239 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
240 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
241 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
242 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
243 dump_gimple_omp_for): Likewise.
244 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
245 * tree-ssa-pre.c (print_pre_expr): Likewise.
246 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
247 * print-tree.c (print_node_brief, print_node): Likewise.
248 * gimple.c (gimple_check_failed): Likewise.
249 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
250 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
251 * config/mep/mep.c (mep_validate_vliw): Likewise.
252 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
253 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
254
82c9df56
GG		 2552013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
256
257 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
258 for AMD bdver3.
259
b82d0df9
HPN		 2602013-10-16 Hans-Peter Nilsson <hp@axis.com>
261
262 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
263 (MULTILIB_MATCHES): Add multilib for -march=v8.
264
bf7b5747
ST		 2652013-10-15 Sriraman Tallam <tmsriram@google.com>
266
267 PR target/57756
268 * optc-save-gen.awk: Add extra parameter to the save and restore
269 target calls.
270 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
271 * tree.c (build_optimization_node): New parameter. Add extra parameter
272 to call to cl_optimization_save.
273 (build_target_option_node): New parameter. Add extra parameter
274 to call to cl_target_option_save.
275 * tree.h (build_optimization_node): New parameter.
276 (build_target_option_node): New parameter.
277 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
278 build_optimization_node and build_target_option_node.
279 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
280 (handle_pragma_push_options): Ditto.
281 * toplev.c (process_options): Ditto.
282 * opts.c (init_options_struct): Check for opts_set non-null.
283 * target.def (target_option.save): New parameter.
284 (target_option.restore): New parameter.
285 * tm.texi: Generate.
286 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
287 (ix86_pragma_target_parse): Ditto.
288 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
289 parameters.
290 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
291 to build_optimization_node and build_target_option_node.
292 (rs6000_valid_attribute_p): Ditto.
293 (rs6000_pragma_target_parse): Ditto.
294 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
295 data.
296 * config/i386/i386.h:
297 TARGET_64BIT_P: New Macro
298 TARGET_MMX_P: New Macro.
299 TARGET_3DNOW_P: New Macro.
300 TARGET_3DNOW_A_P: New Macro.
301 TARGET_SSE_P: New Macro.
302 TARGET_SSE2_P: New Macro.
303 TARGET_SSE3_P: New Macro.
304 TARGET_SSSE3_P: New Macro.
305 TARGET_SSE4_1_P: New Macro.
306 TARGET_SSE4_2_P: New Macro.
307 TARGET_AVX_P: New Macro.
308 TARGET_AVX2_P: New Macro.
309 TARGET_AVX512F_P: New Macro.
310 TARGET_AVX512PF_P: New Macro.
311 TARGET_AVX512ER_P: New Macro.
312 TARGET_AVX512CD_P: New Macro.
313 TARGET_FMA_P: New Macro.
314 TARGET_SSE4A_P: New Macro.
315 TARGET_FMA4_P: New Macro.
316 TARGET_XOP_P: New Macro.
317 TARGET_LWP_P: New Macro.
318 TARGET_ABM_P: New Macro.
319 TARGET_BMI_P: New Macro.
320 TARGET_BMI2_P: New Macro.
321 TARGET_LZCNT_P: New Macro.
322 TARGET_TBM_P: New Macro.
323 TARGET_POPCNT_P: New Macro.
324 TARGET_SAHF_P: New Macro.
325 TARGET_MOVBE_P: New Macro.
326 TARGET_CRC32_P: New Macro.
327 TARGET_AES_P: New Macro.
328 TARGET_PCLMUL_P: New Macro.
329 TARGET_CMPXCHG16B_P: New Macro.
330 TARGET_FSGSBASE_P: New Macro.
331 TARGET_RDRND_P: New Macro.
332 TARGET_F16C_P: New Macro.
333 TARGET_RTM_P: New Macro.
334 TARGET_HLE_P: New Macro.
335 TARGET_RDSEED_P: New Macro.
336 TARGET_PRFCHW_P: New Macro.
337 TARGET_ADX_P: New Macro.
338 TARGET_FXSR_P: New Macro.
339 TARGET_XSAVE_P: New Macro.
340 TARGET_XSAVEOPT_P: New Macro.
341 TARGET_LP64_P: New Macro.
342 TARGET_X32_P: New Macro.
343 TARGET_FPMATH_DEFAULT_P: New Macro.
344 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
345 * config/i386/i386.c (ix86_option_override_internal): New parameters.
346 opts and opts_set.
347 Change ix86_tune_string to access opts->x_ix86_tune_string.
348 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
349 Change ix86_arch_string to access opts->x_ix86_arch_string.
350 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
351 Change ix86_pmode to access opts->x_ix86_pmode.
352 Change ix86_abi to access opts->x_ix86_abi.
353 Change ix86_cmodel to access opts->x_ix86_cmodel.
354 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
355 Change ix86_isa_flags_explicit to access
356 opts->x_ix86_isa_flags_explicit.
357 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
358 Change ix86_regparm to access opts->x_ix86_regparm.
359 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
360 Change ix86_preferred_stack_boundary_arg to access
361 opts->x_ix86_preferred_stack_boundary_arg.
362 Change ix86_force_align_arg_pointer to access
363 opts->x_ix86_force_align_arg_pointer.
364 Change ix86_incoming_stack_boundar_arg to access
365 opts->x_ix86_incoming_stack_boundar_arg.
366 Change ix86_fpmath to access opts->x_ix86_fpmath.
367 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
368 Change ix86_recip_name to access opts->x_ix86_recip_name.
369 Change ix86_stack_protector_guard to access
370 opts->x_ix86_stack_protector_guard.
371 Change ix86_tune_memcpy_strategy to access
372 opts->x_ix86_tune_memcpy_strategy.
373 Change ix86_tune_memset_strategy to access
374 opts->x_ix86_tune_memset_strategy.
375 Change global_options to access opts.
376 Change global_options_set to access opts_set.
6f1abb55
UB		 377 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
378 Change TARGET_MMX to TARGET_MMX_P (opts->...).
379 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
380 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
381 Change TARGET_SSE to TARGET_SSE_P (opts->...).
382 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
383 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
384 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
385 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
386 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
387 Change TARGET_AVX to TARGET_AVX_P (opts->...).
388 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
389 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
390 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
391 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
392 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
393 Change TARGET_FMA to TARGET_FMA_P (opts->...).
394 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
395 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
396 Change TARGET_XOP to TARGET_XOP_P (opts->...).
397 Change TARGET_LWP to TARGET_LWP_P (opts->...).
398 Change TARGET_ABM to TARGET_ABM_P (opts->...).
399 Change TARGET_BMI to TARGET_BMI_P (opts->...).
400 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
401 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
402 Change TARGET_TBM to TARGET_TBM_P (opts->...).
403 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
404 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
405 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
406 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
407 Change TARGET_AES to TARGET_AES_P (opts->...).
408 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
409 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
410 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
411 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
412 Change TARGET_F16C to TARGET_F16C_P (opts->...).
413 Change TARGET_RTM to TARGET_RTM_P (opts->...).
414 Change TARGET_HLE to TARGET_HLE_P (opts->...).
415 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
416 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
417 Change TARGET_ADX to TARGET_ADX_P (opts->...).
418 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
419 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
420 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
421 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
422 Change TARGET_X32 to TARGET_X32_P (opts->...).
423 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
424 Change TARGET_FLOAT_RETURNS_IN_80387 to
425 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
bf7b5747
ST		 426 (ix86_function_specific_save): New parameter. Use opts-> fields
427 to replace global fields.
428 (ix86_function_specific_restore): Ditto.
429 (ix86_valid_target_attribute_inner_p): New parameters.
430 Fix recursive call.
431 Fix call to ix86_handle_option and set_option.
432 (ix86_valid_target_attribute_tree): New parameters.
433 Change global_options to access opts.
434 Change global_options_set to access opts_set.
435 Fix call to ix86_valid_target_attribute_inner_p.
436 Change ix86_tune_string to access opts->x_ix86_tune_string.
437 Change ix86_arch_string to access opts->x_ix86_arch_string.
438 Change ix86_fpmath to access opts->x_ix86_fpmath
439 Fix call to ix86_option_override_internal.
440 Fix call to ix86_add_new_builtins.
441 Fix calls to build_optimization_node and build_target_option_node.
442 (ix86_valid_target_attribute_p): Remove access to global_options.
443 Use new gcc_options structure func_options.
444 Fix call to ix86_valid_target_attribute_tree.
445 Fix call to build_optimization_node.
446 (get_builtin_code_for_version): Fix call to
447 ix86_valid_target_attribute_tree.
448
459260ec
DM		 4492013-10-15 David Malcolm <dmalcolm@redhat.com>
450
451 * Makefile.in (PICFLAG): New.
452 (enable_host_shared): New.
453 (INTERNAL_CFLAGS): Use PICFLAG.
454 (LIBIBERTY): Use pic build of libiberty.a if configured with
455 --enable-host-shared.
456 * configure.ac: Add --enable-host-shared, setting up new
457 PICFLAG variable.
458 * configure: Regenerate.
6f1abb55 459 * doc/install.texi (--enable-shared): Add note contrasting it with ...
459260ec
DM		 460 (--enable-host-shared): New option.
461
16022041
RS		 4622013-10-15 Richard Biener <rguenther@suse.de>
463
464 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
465 for built-in functions.
466
b114bfb4
ZC		 4672013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
468
469 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
470 (optimize_range_tests_1): New function,
471 extracted from optimize_range_tests.
472 (optimize_range_tests_xor): Similarly.
473 (optimize_range_tests_diff): New function.
474 (optimize_range_tests): Use optimize_range_tests_1.
475
69b8f2f9
CH		 4762013-10-15 Cong Hou <congh@google.com>
477
7ab4168e
MS		 478 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
479 requirement of the reduction pattern so that one operand of the
69b8f2f9
CH		 480 reduction operation can come from outside of the loop.
481
8ff5a4e4
JG		 4822013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
483
484 * config/arm/neon-schedgen.ml: Remove.
485 * config/arm/cortex-a9-neon.md: Remove comment regarding
486 neon-schedgen.ml.
487
52b69ad2
JG		 4882013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
489
490 * config/arm/types: Remove old neon types.
491
e7c6a556
JG		 4922013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
493
494 * config/arm/cortex-a7.md
495 (cortex_a7_neon_type): New.
496 (cortex_a7_neon_mul): Update for new types.
497 (cortex_a7_neon_mla): Likewise.
498 (cortex_a7_neon): Likewise.
499
795dd8d5
JG		 5002013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
501
502 * config/arm/cortex-a15-neon.md
503 (cortex_a15_neon_type): New,
504
505 (cortex_a15_neon_int_1): Remove.
506 (cortex_a15_neon_int_2): Likewise.
507 (cortex_a15_neon_int_3): Likewise.
508 (cortex_a15_neon_int_4): Likewise.
509 (cortex_a15_neon_int_5): Likewise.
510 (cortex_a15_neon_vqneg_vqabs): Likewise.
511 (cortex_a15_neon_vmov): Likewise.
512 (cortex_a15_neon_vaba): Likewise.
513 (cortex_a15_neon_vaba_qqq): Likewise.
514 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
515 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
516 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
517 Likewise.
518 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
519 (cortex_a15_neon_mla_qqq_8_16): Likewise.
520 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
521 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
522 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
523 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
524 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
525 (cortex_a15_neon_shift_1): Likewise.
526 (cortex_a15_neon_shift_2): Likewise.
527 (cortex_a15_neon_shift_3): Likewise.
528 (cortex_a15_neon_vshl_ddd): Likewise.
529 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
530 (cortex_a15_neon_vsra_vrsra): Likewise.
531 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
532 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
533 (cortex_a15_neon_bp_3cycle): Likewise.
534 (cortex_a15_neon_ldm_2): Likewise.
535 (cortex_a15_neon_stm_2): Likewise.
536 (cortex_a15_neon_mcr): Likewise.
537 (cortex_a15_neon_mrc): Likewise.
538 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
539 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
540 (cortex_a15_neon_fp_vmul_ddd): Likewise.
541 (cortex_a15_neon_fp_vmul_qqd): Likewise.
542 (cortex_a15_neon_fp_vmla_ddd): Likewise.
543 (cortex_a15_neon_fp_vmla_qqq): Likewise.
544 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
545 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
546 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
547 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
548 (cortex_a15_neon_bp_simple): Likewise.
549 (cortex_a15_neon_bp_2cycle): Likewise.
550 (cortex_a15_neon_bp_3cycle): Likewise.
551 (cortex_a15_neon_vld1_1_2_regs): Likewise.
552 (cortex_a15_neon_vld1_3_4_regs): Likewise.
553 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
554 (cortex_a15_neon_vld2_4_regs): Likewise.
555 (cortex_a15_neon_vld3_vld4): Likewise.
556 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
557 (cortex_a15_neon_vst1_3_4_regs): Likewise.
558 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
559 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
560 (cortex_a15_neon_vst3_vst4): Rename to...
561 (cortex_a15_neon_vst4): This, update for new attributes.
562 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
563 (cortex_a15_neon_vld3_vld4_lane): Likewise.
564 (cortex_a15_neon_vst1_vst2_lane): Likewise.
565 (cortex_a15_neon_vst3_vst4_lane): Likewise.
566 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
567 (cortex_a15_neon_ldm_2): Likewise.
568 (cortex_a15_neon_stm_2): Likewise.
569 (cortex_a15_neon_mcr): Likewise.
570 (cortex_a15_neon_mcr_2_mcrr): Likewise.
571 (cortex_a15_neon_mrc): Likewise.
572 (cortex_a15_neon_mrrc): Likewise.
573
574 (cortex_a15_neon_abd): New.
575 (cortex_a15_neon_abd_q): Likewise.
576 (cortex_a15_neon_aba): Likewise.
577 (cortex_a15_neon_aba_q): Likewise.
578 (cortex_a15_neon_acc): Likewise.
579 (cortex_a15_neon_acc_q): Likewise.
580 (cortex_a15_neon_arith_basic): Likewise.
581 (cortex_a15_neon_arith_complex): Likewise.
582 (cortex_a15_neon_multiply): Likewise.
583 (cortex_a15_neon_multiply_q): Likewise.
584 (cortex_a15_neon_mla): Likewise.
585 (cortex_a15_neon_mla_q): Likewise.
586 (cortex_a15_neon_sat_mla_long): Likewise.
587 (cortex_a15_neon_shift_acc): Likewise.
588 (cortex_a15_neon_shift_imm_basic): Likewise.
589 (cortex_a15_neon_shift_imm_complex): Likewise.
590 (cortex_a15_neon_shift_reg_basic): Likewise.
591 (cortex_a15_neon_shift_reg_basic_q): Likewise.
592 (cortex_a15_neon_shift_reg_complex): Likewise.
593 (cortex_a15_neon_shift_reg_complex_q): Likewise.
594 (cortex_a15_neon_fp_negabs): Likewise
595 (cortex_a15_neon_fp_arith): Likewise
596 (cortex_a15_neon_fp_arith_q): Likewise
597 (cortex_a15_neon_fp_cvt_int): Likewise
598 (cortex_a15_neon_fp_cvt_int_q): Likewise
599 (cortex_a15_neon_fp_cvt_16): Likewise
600 (cortex_a15_neon_fp_mul): Likewise
601 (cortex_a15_neon_fp_mul_q): Likewise
602 (cortex_a15_neon_fp_mla): Likewise
603 (cortex_a15_neon_fp_mla_q): Likewise
604 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
605 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
606 (cortex_a15_neon_bitops): Likewise.
607 (cortex_a15_neon_bitops_q): Likewise.
608 (cortex_a15_neon_from_gp): Likewise.
609 (cortex_a15_neon_from_gp_q): Likewise.
610 (cortex_a15_neon_tbl3_tbl4): Likewise.
611 (cortex_a15_neon_zip_q): Likewise.
612 (cortex_a15_neon_to_gp): Likewise.
613 (cortex_a15_neon_load_a): Likewise.
614 (cortex_a15_neon_load_b): Likewise.
615 (cortex_a15_neon_load_c): Likewise.
616 (cortex_a15_neon_load_d): Likewise.
617 (cortex_a15_neon_load_e): Likewise.
618 (cortex_a15_neon_load_f): Likewise.
619 (cortex_a15_neon_store_a): Likewise.
620 (cortex_a15_neon_store_b): Likewise.
621 (cortex_a15_neon_store_c): Likewise.
622 (cortex_a15_neon_store_d): Likewise.
623 (cortex_a15_neon_store_e): Likewise.
624 (cortex_a15_neon_store_f): Likewise.
625 (cortex_a15_neon_store_g): Likewise.
626 (cortex_a15_neon_store_h): Likewise.
627 (cortex_a15_vfp_to_from_gp): Likewise.
628
bc5e57e3
JG		 6292013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
630
631 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
632
633 (cortex_a9_neon_vshl_ddd): Remove.
634 (cortex_a9_neon_vst3_vst4): Likewise.
635 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
636
637 (cortex_a9_neon_bit_ops_q): New.
638
639 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
640 (cortex_a9_neon_int_2): Likewise.
641 (cortex_a9_neon_int_3): Likewise.
642 (cortex_a9_neon_int_4): Likewise.
643 (cortex_a9_neon_int_5): Likewise.
644 (cortex_a9_neon_vqneg_vqabs): Likewise.
645 (cortex_a9_neon_vmov): Likewise.
646 (cortex_a9_neon_vaba): Likewise.
647 (cortex_a9_neon_vaba_qqq): Likewise.
648 (cortex_a9_neon_shift_1): Likewise.
649 (cortex_a9_neon_shift_2): Likewise.
650 (cortex_a9_neon_shift_3): Likewise.
651 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
652 (cortex_a9_neon_vsra_vrsra): Likewise.
653 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
654 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
655 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
656 Likewise.
657 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
658 (cortex_a9_neon_mla_qqq_8_16): Likewise.
659 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
660 Likewise.
661 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
662 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
663 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
664 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
665 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
666 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
667 (cortex_a9_neon_fp_vsum): Likewise.
668 (cortex_a9_neon_fp_vmul_ddd): Likewise.
669 (cortex_a9_neon_fp_vmul_qqd): Likewise.
670 (cortex_a9_neon_fp_vmla_ddd): Likewise.
671 (cortex_a9_neon_fp_vmla_qqq): Likewise.
672 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
673 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
674 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
675 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
676 (cortex_a9_neon_bp_simple): Likewise.
677 (cortex_a9_neon_bp_2cycle): Likewise.
678 (cortex_a9_neon_bp_3cycle): Likewise.
679 (cortex_a9_neon_ldr): Likewise.
680 (cortex_a9_neon_str): Likewise.
681 (cortex_a9_neon_vld1_1_2_regs): Likewise.
682 (cortex_a9_neon_vld1_3_4_regs): Likewise.
683 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
684 (cortex_a9_neon_vld2_4_regs): Likewise.
685 (cortex_a9_neon_vld3_vld4): Likewise.
686 (cortex_a9_neon_vld1_vld2_lane): Likewise.
687 (cortex_a9_neon_vld3_vld4_lane): Likewise.
688 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
689 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
690 (cortex_a9_neon_vst1_3_4_regs): Likewise.
691 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
692 (cortex_a9_neon_vst1_vst2_lane): Likewise.
693 (cortex_a9_neon_vst3_vst4_lane): Likewise.
694 (cortex_a9_neon_mcr): Likewise.
695 (cortex_a9_neon_mcr_2_mcrr): Likewise.
696 (cortex_a9_neon_mrc): Likewise.
697 (cortex_a9_neon_mrrc): Likewise.
698
9deda8cb
JG		 6992013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
700
701 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
702
703 (cortex_a8_neon_vshl_ddd): Remove.
704 (cortex_a8_neon_vst3_vst4): Likewise.
705 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
706
707 (cortex_a8_neon_bit_ops_q): New.
708
709 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
710 (cortex_a8_neon_int_2): Likewise..
711 (cortex_a8_neon_int_3): Likewise.
712 (cortex_a8_neon_int_5): Likewise.
713 (cortex_a8_neon_vqneg_vqabs): Likewise.
714 (cortex_a8_neon_int_4): Likewise.
715 (cortex_a8_neon_vaba): Likewise.
716 (cortex_a8_neon_vaba_qqq): Likewise.
717 (cortex_a8_neon_shift_1): Likewise.
718 (cortex_a8_neon_shift_2): Likewise.
719 (cortex_a8_neon_shift_3): Likewise.
720 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
721 (cortex_a8_neon_vsra_vrsra): Likewise.
722 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
723 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
724 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
725 Likewise.
726 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
727 (cortex_a8_neon_mla_qqq_8_16): Likewise.
728 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
729 Likewise.
730 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
731 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
732 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
733 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
734 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
735 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
736 (cortex_a8_neon_fp_vsum): Likewise.
737 (cortex_a8_neon_fp_vmul_ddd): Likewise.
738 (cortex_a8_neon_fp_vmul_qqd): Likewise.
739 (cortex_a8_neon_fp_vmla_ddd): Likewise.
740 (cortex_a8_neon_fp_vmla_qqq): Likewise.
741 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
742 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
743 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
744 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
745 (cortex_a8_neon_bp_simple): Likewise.
746 (cortex_a8_neon_bp_2cycle): Likewise.
747 (cortex_a8_neon_bp_3cycle): Likewise.
748 (cortex_a8_neon_ldr): Likewise.
749 (cortex_a8_neon_str): Likewise.
750 (cortex_a8_neon_vld1_1_2_regs): Likewise.
751 (cortex_a8_neon_vld1_3_4_regs): Likewise.
752 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
753 (cortex_a8_neon_vld2_4_regs): Likewise.
754 (cortex_a8_neon_vld3_vld4): Likewise.
755 (cortex_a8_neon_vld1_vld2_lane): Likewise.
756 (cortex_a8_neon_vld3_vld4_lane): Likewise.
757 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
758 (cortex_a8_neon_vst1_3_4_regs): Likewise.
759 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
760 (cortex_a8_neon_vst1_vst2_lane): Likewise.
761 (cortex_a8_neon_vst3_vst4_lane): Likewise.
762 (cortex_a8_neon_mcr): Likewise.
763 (cortex_a8_neon_mcr_2_mcrr): Likewise.
764 (cortex_a8_neon_mrc): Likewise.
765 (cortex_a8_neon_mrrc): Likewise.
766
0f686aa9
JG		 7672013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
768
769 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
770 (fp): New.
771 * config/aarch64/aarch64-simd.md (neon_type): Remove.
772 (aarch64_simd_dup<mode>): Add "type" attribute.
773 (aarch64_dup_lane<mode>): Likewise.
774 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
775 (*aarch64_simd_mov<mode>): Likewise.
776 (aarch64_simd_mov_from_<mode>low): Likewise.
777 (aarch64_simd_mov_from_<mode>high): Likewise.
778 (orn<mode>3): Likewise.
779 (bic<mode>3): Likewise.
780 (add<mode>3): Likewise.
781 (sub<mode>3): Likewise.
782 (mul<mode>3): Likewise.
783 (*aarch64_mul3_elt<mode>): Likewise.
784 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
785 (*aarch64_mul3_elt_to_128df): Likewise.
786 (*aarch64_mul3_elt_to_64v2df): Likewise.
787 (neg<mode>2): Likewise.
788 (abs<mode>2): Likewise.
789 (abd<mode>_3): Likewise.
790 (aba<mode>_3): Likewise.
791 (fabd<mode>_3): Likewise.
792 (*fabd_scalar<mode>3): Likewise.
793 (and<mode>3): Likewise.
794 (ior<mode>3): Likewise.
795 (xor<mode>3): Likewise.
796 (one_cmpl<mode>2): Likewise.
797 (aarch64_simd_vec_set<mode>): Likewise.
798 (aarch64_simd_lshr<mode>): Likewise.
799 (aarch64_simd_ashr<mode>): Likewise.
800 (aarch64_simd_imm_shl<mode>): Likewise.
801 (aarch64_simd_reg_sshl<mode): Likewise.
802 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
803 (aarch64_simd_reg_shl<mode>_signed): Likewise.
804 (aarch64_simd_vec_setv2di): Likewise.
805 (aarch64_simd_vec_set<mode>): Likewise.
806 (aarch64_mla<mode>): Likewise.
807 (*aarch64_mla_elt<mode>): Likewise.
808 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
809 (aarch64_mls<mode>): Likewise.
810 (*aarch64_mls_elt<mode>): Likewise.
811 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
812 (<su><maxmin><mode>3): Likewise.
813 (move_lo_quad_<mode>): Likewise.
814 (aarch64_simd_move_hi_quad_<mode>): Likewise.
815 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
816 (vec_pack_trunc_<mode>): Likewise.
817 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
818 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
819 (*aarch64_<su>mlal_lo<mode>): Likewise.
820 (*aarch64_<su>mlal_hi<mode>): Likewise.
821 (*aarch64_<su>mlsl_lo<mode>): Likewise.
822 (*aarch64_<su>mlsl_hi<mode>): Likewise.
823 (*aarch64_<su>mlal<mode>): Likewise.
824 (*aarch64_<su>mlsl<mode>): Likewise.
825 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
826 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
827 (add<mode>3): Likewise.
828 (sub<mode>3): Likewise.
829 (mul<mode>3): Likewise.
830 (div<mode>3): Likewise.
831 (neg<mode>2): Likewise.
832 (abs<mode>2): Likewise.
833 (fma<mode>4): Likewise.
834 (*aarch64_fma4_elt<mode>): Likewise.
835 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
836 (*aarch64_fma4_elt_to_128df): Likewise.
837 (*aarch64_fma4_elt_to_64v2df): Likewise.
838 (fnma<mode>4): Likewise.
839 (*aarch64_fnma4_elt<mode>): Likewise.
840 (*aarch64_fnma4_elt_<vswap_width_name><mode>
841 (*aarch64_fnma4_elt_to_128df): Likewise.
842 (*aarch64_fnma4_elt_to_64v2df): Likewise.
843 (<frint_pattern><mode>2): Likewise.
844 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
845 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
846 (vec_unpacks_lo_v4sf): Likewise.
847 (aarch64_float_extend_lo_v2df): Likewise.
848 (vec_unpacks_hi_v4sf): Likewise.
849 (aarch64_float_truncate_lo_v2sf): Likewise.
850 (aarch64_float_truncate_hi_v4sf): Likewise.
851 (aarch64_vmls<mode>): Likewise.
852 (<su><maxmin><mode>3): Likewise.
853 (<maxmin_uns><mode>3): Likewise.
854 (reduc_<sur>plus_<mode>): Likewise.
855 (reduc_<sur>plus_v2di): Likewise.
856 (reduc_<sur>plus_v2si): Likewise.
857 (reduc_<sur>plus_<mode>): Likewise.
858 (aarch64_addpv4sf): Likewise.
859 (clz<mode>2): Likewise.
860 (reduc_<maxmin_uns>_<mode>): Likewise.
861 (reduc_<maxmin_uns>_v2di): Likewise.
862 (reduc_<maxmin_uns>_v2si): Likewise.
863 (reduc_<maxmin_uns>_<mode>): Likewise.
864 (reduc_<maxmin_uns>_v4sf): Likewise.
865 (aarch64_simd_bsl<mode>_internal): Likewise.
866 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
867 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
868 (aarch64_get_lane<mode>): Likewise.
869 (*aarch64_combinez<mode>): Likewise.
870 (aarch64_combine<mode>): Likewise.
871 (aarch64_simd_combine<mode>): Likewise.
872 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
873 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
874 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
875 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
876 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
877 (aarch64_<sur>h<addsub><mode>): Likewise.
878 (aarch64_<sur><addsub>hn<mode>): Likewise.
879 (aarch64_<sur><addsub>hn2<mode>): Likewise.
880 (aarch64_pmul<mode>): Likewise.
881 (aarch64_<su_optab><optab><mode>): Likewise.
882 (aarch64_<sur>qadd<mode>): Likewise.
883 (aarch64_sqmovun<mode>): Likewise.
884 (aarch64_<sur>qmovn<mode>): Likewise.
885 (aarch64_s<optab><mode>): Likewise.
886 (aarch64_sq<r>dmulh<mode>): Likewise.
887 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
888 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
889 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
890 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
891 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
892 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
893 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
894 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
895 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
896 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
897 (aarch64_sqdmull<mode>): Likewise.
898 (aarch64_sqdmull_lane<mode>_internal): Likewise.
899 (aarch64_sqdmull_n<mode>): Likewise.
900 (aarch64_sqdmull2<mode>_internal): Likewise.
901 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
902 (aarch64_sqdmull2_n<mode>_internal): Likewise.
903 (aarch64_<sur>shl<mode>): Likewise.
904 (aarch64_<sur>q<r>shl<mode>
905 (aarch64_<sur>shll_n<mode>): Likewise.
906 (aarch64_<sur>shll2_n<mode>): Likewise.
907 (aarch64_<sur>shr_n<mode>): Likewise.
908 (aarch64_<sur>sra_n<mode>): Likewise.
909 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
910 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
911 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
912 (aarch64_cm<optab><mode>): Likewise.
913 (aarch64_cm<optab>di): Likewise.
914 (aarch64_cm<optab><mode>): Likewise.
915 (aarch64_cm<optab>di): Likewise.
916 (aarch64_cmtst<mode>): Likewise.
917 (aarch64_cmtstdi): Likewise.
918 (aarch64_cm<optab><mode>): Likewise.
919 (*aarch64_fac<optab><mode>): Likewise.
920 (aarch64_addp<mode>): Likewise.
921 (aarch64_addpdi): Likewise.
922 (sqrt<mode>2): Likewise.
923 (vec_load_lanesoi<mode>): Likewise.
924 (vec_store_lanesoi<mode>): Likewise.
925 (vec_load_lanesci<mode>): Likewise.
926 (vec_store_lanesci<mode>): Likewise.
927 (vec_load_lanesxi<mode>): Likewise.
928 (vec_store_lanesxi<mode>): Likewise.
929 (*aarch64_mov<mode>): Likewise.
930 (aarch64_ld2<mode>_dreg): Likewise.
931 (aarch64_ld2<mode>_dreg): Likewise.
932 (aarch64_ld3<mode>_dreg): Likewise.
933 (aarch64_ld3<mode>_dreg): Likewise.
934 (aarch64_ld4<mode>_dreg): Likewise.
935 (aarch64_ld4<mode>_dreg): Likewise.
936 (aarch64_tbl1<mode>): Likewise.
937 (aarch64_tbl2v16qi): Likewise.
938 (aarch64_combinev16qi): Likewise.
939 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
940 (aarch64_st2<mode>_dreg): Likewise.
941 (aarch64_st2<mode>_dreg): Likewise.
942 (aarch64_st3<mode>_dreg): Likewise.
943 (aarch64_st3<mode>_dreg): Likewise.
944 (aarch64_st4<mode>_dreg): Likewise.
945 (aarch64_st4<mode>_dreg): Likewise.
946 (*aarch64_simd_ld1r<mode>): Likewise.
947 (aarch64_frecpe<mode>): Likewise.
948 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
949 (aarch64_frecps<mode>): Likewise.
950
f7379e5e
JG		 9512013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
952
953 * config/arm/iterators.md (V_elem_ch): New.
954 (q): Likewise.
955 (VQH_type): Likewise.
956 * config/arm/arm.md (is_neon_type): New.
957 (conds): Use is_neon_type.
958 (anddi3_insn): Update type attribute.
959 (xordi3_insn): Likewise.
960 (one_cmpldi2): Likewise.
961 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
962 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
963 (*movmisalign<mode>_neon_store): Likewise.
964 (*movmisalign<mode>_neon_load): Likewise.
965 (vec_set<mode>_internal): Likewise.
966 (vec_set<mode>_internal): Likewise.
967 (vec_setv2di_internal): Likewise.
968 (vec_extract<mode>): Likewise.
969 (vec_extract<mode>): Likewise.
970 (vec_extractv2di): Likewise.
971 (*add<mode>3_neon): Likewise.
972 (adddi3_neon): Likewise.
973 (*sub<mode>3_neon): Likewise.
974 (subdi3_neon): Likewise.
975 (fma<VCVTF:mode>4): Likewise.
976 (fma<VCVTF:mode>4_intrinsic): Likewise.
977 (*fmsub<VCVTF:mode>4): Likewise.
978 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
979 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
980 (ior<mode>3): Likewise.
981 (and<mode>3): Likewise.
982 (orn<mode>3_neon): Likewise.
983 (orndi3_neon): Likewise.
984 (bic<mode>3_neon): Likewise.
985 (bicdi3_neon): Likewise.
986 (xor<mode>3): Likewise.
987 (one_cmpl<mode>2): Likewise.
988 (abs<mode>2): Likewise.
989 (neg<mode>2): Likewise.
990 (negdi2_neon): Likewise.
991 (*umin<mode>3_neon): Likewise.
992 (*umax<mode>3_neon): Likewise.
993 (*smin<mode>3_neon): Likewise.
994 (*smax<mode>3_neon): Likewise.
995 (vashl<mode>3): Likewise.
996 (vashr<mode>3_imm): Likewise.
997 (vlshr<mode>3_imm): Likewise.
998 (ashl<mode>3_signed): Likewise.
999 (ashl<mode>3_unsigned): Likewise.
1000 (neon_load_count): Likewise.
1001 (ashldi3_neon_noclobber): Likewise.
1002 (ashldi3_neon): Likewise.
1003 (signed_shift_di3_neon): Likewise.
1004 (unsigned_shift_di3_neon): Likewise.
1005 (ashrdi3_neon_imm_noclobber): Likewise.
1006 (lshrdi3_neon_imm_noclobber): Likewise.
1007 (<shift>di3_neon): Likewise.
1008 (widen_ssum<mode>3): Likewise.
1009 (widen_usum<mode>3): Likewise.
1010 (quad_halves_<code>v4si): Likewise.
1011 (quad_halves_<code>v4sf): Likewise.
1012 (quad_halves_<code>v8hi): Likewise.
1013 (quad_halves_<code>v16qi): Likewise.
1014 (reduc_splus_v2di): Likewise.
1015 (neon_vpadd_internal<mode>): Likewise.
1016 (neon_vpsmin<mode>): Likewise.
1017 (neon_vpsmax<mode>): Likewise.
1018 (neon_vpumin<mode>): Likewise.
1019 (neon_vpumax<mode>): Likewise.
1020 (*ss_add<mode>_neon): Likewise.
1021 (*us_add<mode>_neon): Likewise.
1022 (*ss_sub<mode>_neon): Likewise.
1023 (*us_sub<mode>_neon): Likewise.
1024 (neon_vadd<mode>_unspec): Likewise.
1025 (neon_vaddl<mode>): Likewise.
1026 (neon_vaddw<mode>): Likewise.
1027 (neon_vhadd<mode>): Likewise.
1028 (neon_vqadd<mode>): Likewise.
1029 (neon_vaddhn<mode>): Likewise.
1030 (neon_vmul<mode>): Likewise.
1031 (neon_vfms<VCVTF:mode>): Likewise.
1032 (neon_vmlal<mode>): Likewise.
1033 (neon_vmls<mode>): Likewise.
1034 (neon_vmlsl<mode>): Likewise.
1035 (neon_vqdmulh<mode>): Likewise.
1036 (neon_vqdmlal<mode>): Likewise.
1037 (neon_vqdmlsl<mode>): Likewise.
1038 (neon_vmull<mode>): Likewise.
1039 (neon_vqdmull<mode>): Likewise.
1040 (neon_vsub<mode>_unspec): Likewise.
1041 (neon_vsubl<mode>): Likewise.
1042 (neon_vsubw<mode>): Likewise.
1043 (neon_vqsub<mode>): Likewise.
1044 (neon_vhsub<mode>): Likewise.
1045 (neon_vsubhn<mode>): Likewise.
1046 (neon_vceq<mode>): Likewise.
1047 (neon_vcge<mode>): Likewise.
1048 (neon_vcgeu<mode>): Likewise.
1049 (neon_vcgt<mode>): Likewise.
1050 (neon_vcgtu<mode>): Likewise.
1051 (neon_vcle<mode>): Likewise.
1052 (neon_vclt<mode>): Likewise.
1053 (neon_vcage<mode>): Likewise.
1054 (neon_vcagt<mode>): Likewise.
1055 (neon_vtst<mode>): Likewise.
1056 (neon_vabd<mode>): Likewise.
1057 (neon_vabdl<mode>): Likewise.
1058 (neon_vaba<mode>): Likewise.
1059 (neon_vabal<mode>): Likewise.
1060 (neon_vmax<mode>): Likewise.
1061 (neon_vmin<mode>): Likewise.
1062 (neon_vpaddl<mode>): Likewise.
1063 (neon_vpadal<mode>): Likewise.
1064 (neon_vpmax<mode>): Likewise.
1065 (neon_vpmin<mode>): Likewise.
1066 (neon_vrecps<mode>): Likewise.
1067 (neon_vrsqrts<mode>): Likewise.
1068 (neon_vqabs<mode>): Likewise.
1069 (neon_vqneg<mode>): Likewise.
1070 (neon_vcls<mode>): Likewise.
1071 (clz<mode>2): Likewise.
1072 (popcount<mode>2): Likewise.
1073 (neon_vrecpe<mode>): Likewise.
1074 (neon_vrsqrte<mode>): Likewise.
1075 (neon_vget_lane<mode>_sext_internal): Likewise.
1076 (neon_vget_lane<mode>_zext_internal): Likewise.
1077 (neon_vdup_n<mode>): Likewise.
1078 (neon_vdup_n<mode>): Likewise.
1079 (neon_vdup_nv2di): Likewise.
1080 (neon_vdup_lane<mode>_interal): Likewise.
1081 (*neon_vswp<mode>): Likewise.
1082 (neon_vcombine<mode>): Likewise.
1083 (float<mode><V_cvtto>2): Likewise.
1084 (floatuns<mode><V_cvtto>2): Likewise.
1085 (fix_trunc<mode><V_cvtto>2): Likewise.
1086 (fixuns_trunc<mode><V_cvtto>2
1087 (neon_vcvt<mode>): Likewise.
1088 (neon_vcvt<mode>): Likewise.
1089 (neon_vcvtv4sfv4hf): Likewise.
1090 (neon_vcvtv4hfv4sf): Likewise.
1091 (neon_vcvt_n<mode>): Likewise.
1092 (neon_vcvt_n<mode>): Likewise.
1093 (neon_vmovn<mode>): Likewise.
1094 (neon_vqmovn<mode>): Likewise.
1095 (neon_vqmovun<mode>): Likewise.
1096 (neon_vmovl<mode>): Likewise.
1097 (neon_vmul_lane<mode>): Likewise.
1098 (neon_vmul_lane<mode>): Likewise.
1099 (neon_vmull_lane<mode>): Likewise.
1100 (neon_vqdmull_lane<mode>): Likewise.
1101 (neon_vqdmulh_lane<mode>): Likewise.
1102 (neon_vqdmulh_lane<mode>): Likewise.
1103 (neon_vmla_lane<mode>): Likewise.
1104 (neon_vmla_lane<mode>): Likewise.
1105 (neon_vmlal_lane<mode>): Likewise.
1106 (neon_vqdmlal_lane<mode>): Likewise.
1107 (neon_vmls_lane<mode>): Likewise.
1108 (neon_vmls_lane<mode>): Likewise.
1109 (neon_vmlsl_lane<mode>): Likewise.
1110 (neon_vqdmlsl_lane<mode>): Likewise.
1111 (neon_vext<mode>): Likewise.
1112 (neon_vrev64<mode>): Likewise.
1113 (neon_vrev32<mode>): Likewise.
1114 (neon_vrev16<mode>): Likewise.
1115 (neon_vbsl<mode>_internal): Likewise.
1116 (neon_vshl<mode>): Likewise.
1117 (neon_vqshl<mode>): Likewise.
1118 (neon_vshr_n<mode>): Likewise.
1119 (neon_vshrn_n<mode>): Likewise.
1120 (neon_vqshrn_n<mode>): Likewise.
1121 (neon_vqshrun_n<mode>): Likewise.
1122 (neon_vshl_n<mode>): Likewise.
1123 (neon_vqshl_n<mode>): Likewise.
1124 (neon_vqshlu_n<mode>): Likewise.
1125 (neon_vshll_n<mode>): Likewise.
1126 (neon_vsra_n<mode>): Likewise.
1127 (neon_vsri_n<mode>): Likewise.
1128 (neon_vsli_n<mode>): Likewise.
1129 (neon_vtbl1v8qi): Likewise.
1130 (neon_vtbl2v8qi): Likewise.
1131 (neon_vtbl3v8qi): Likewise.
1132 (neon_vtbl4v8qi): Likewise.
1133 (neon_vtbl1v16qi): Likewise.
1134 (neon_vtbl2v16qi): Likewise.
1135 (neon_vcombinev16qi): Likewise.
1136 (neon_vtbx1v8qi): Likewise.
1137 (neon_vtbx2v8qi): Likewise.
1138 (neon_vtbx3v8qi): Likewise.
1139 (neon_vtbx4v8qi): Likewise.
1140 (*neon_vtrn<mode>_insn): Likewise.
1141 (*neon_vzip<mode>_insn): Likewise.
1142 (*neon_vuzp<mode>_insn): Likewise.
1143 (neon_vld1<mode>): Likewise.
1144 (neon_vld1_lane<mode>): Likewise.
1145 (neon_vld1_lane<mode>): Likewise.
1146 (neon_vld1_dup<mode>): Likewise.
1147 (neon_vld1_dup<mode>): Likewise.
1148 (neon_vld1_dupv2di): Likewise.
1149 (neon_vst1<mode>): Likewise.
1150 (neon_vst1_lane<mode>): Likewise.
1151 (neon_vst1_lane<mode>): Likewise.
1152 (neon_vld2<mode>): Likewise.
1153 (neon_vld2<mode>): Likewise.
1154 (neon_vld2_lane<mode>): Likewise.
1155 (neon_vld2_lane<mode>): Likewise.
1156 (neon_vld2_dup<mode>): Likewise.
1157 (neon_vst2<mode>): Likewise.
1158 (neon_vst2<mode>): Likewise.
1159 (neon_vst2_lane<mode>): Likewise.
1160 (neon_vst2_lane<mode>): Likewise.
1161 (neon_vld3<mode>): Likewise.
1162 (neon_vld3qa<mode>): Likewise.
1163 (neon_vld3qb<mode>): Likewise.
1164 (neon_vld3_lane<mode>): Likewise.
1165 (neon_vld3_lane<mode>): Likewise.
1166 (neon_vld3_dup<mode>): Likewise.
1167 (neon_vst3<mode>): Likewise.
1168 (neon_vst3qa<mode>): Likewise.
1169 (neon_vst3qb<mode>): Likewise.
1170 (neon_vst3_lane<mode>): Likewise.
1171 (neon_vst3_lane<mode>): Likewise.
1172 (neon_vld4<mode>): Likewise.
1173 (neon_vld4qa<mode>): Likewise.
1174 (neon_vld4qb<mode>): Likewise.
1175 (neon_vld4_lane<mode>): Likewise.
1176 (neon_vld4_lane<mode>): Likewise.
1177 (neon_vld4_dup<mode>): Likewise.
1178 (neon_vst4<mode>): Likewise.
1179 (neon_vst4qa<mode>): Likewise.
1180 (neon_vst4qb<mode>): Likewise.
1181 (neon_vst4_lane<mode>): Likewise.
1182 (neon_vst4_lane<mode>): Likewise.
1183 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1184 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1185 (neon_vec_<US>mult_lo_<mode>): Likewise.
1186 (neon_vec_<US>mult_hi_<mode>): Likewise.
1187 (neon_vec_<US>shiftl_<mode>): Likewise.
1188 (neon_unpack<US>_<mode>): Likewise.
1189 (neon_vec_<US>mult_<mode>): Likewise.
1190 (vec_pack_trunc_<mode>): Likewise.
1191 (neon_vec_pack_trunc_<mode>): Likewise.
1192 (neon_vabd<mode>_2): Likewise.
1193 (neon_vabd<mode>_3): Likewise.
1194
a9e66678
JG		 11952013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1196
1197 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1198 (load_pair): Update type attribute.
1199 (store_pair): Update type attribute.
1200 * config/aarch64/iterators.md (q): New.
1201
20445ed6
JG		 12022013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1203
1204 * config/arm/types.md: Add new types for Neon insns.
1205
afb4ac68
AI		 12062013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1207 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1208 Sergey Lega <sergey.s.lega@intel.com>
1209 Anna Tikhonova <anna.tikhonova@intel.com>
1210 Ilya Tocar <ilya.tocar@intel.com>
1211 Andrey Turetskiy <andrey.turetskiy@intel.com>
1212 Ilya Verbin <ilya.verbin@intel.com>
1213 Kirill Yukhin <kirill.yukhin@intel.com>
1214 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1215
1216 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1217 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1218 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1219 (rcp14<mode>): New.
1220 (srcp14<mode>): Ditto.
1221 (rsqrt14<mode>): Ditto.
1222 (rsqrt14<mode>): Ditto.
1223 (avx512f_vmscalef<mode>): Ditto.
1224 (avx512f_scalef<mode>): Ditto.
1225 (avx512f_getexp<mode>): Ditto.
1226 (avx512f_sgetexp<mode>): Ditto.
1227 (avx512f_fixupimm<mode>): Ditto.
1228 (avx512f_sfixupimm<mode>): Ditto.
1229 (avx512f_rndscale<mode>): Ditto.
1230 (*avx512er_exp2<mode>): Ditto.
1231 (*avx512er_rcp28<mode>): Ditto.
1232 (avx512er_rsqrt28<mode>): Ditto.
1233 (avx512f_getmant<mode>): Ditto.
1234 (avx512f_getmant<mode>): Ditto.
1235 (avx512f_rndscale<mode>): Fix formatting.
1236
a1b499a3
JL		 12372013-10-15 Martin Jambor <mjambor@suse.cz>
1238
1239 * ipa-utils.h (ipa_edge_within_scc): Declare.
1240 * ipa-cp.c (edge_within_scc): Moved...
1241 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 1242
2e2206fa
AI		 12432013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1244 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1245 Sergey Lega <sergey.s.lega@intel.com>
1246 Anna Tikhonova <anna.tikhonova@intel.com>
1247 Ilya Tocar <ilya.tocar@intel.com>
1248 Andrey Turetskiy <andrey.turetskiy@intel.com>
1249 Ilya Verbin <ilya.verbin@intel.com>
1250 Kirill Yukhin <kirill.yukhin@intel.com>
1251 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1252
1253 * config/i386/predicates.md (const_8_to_15_operand): New.
1254 (const_16_to_31_operand): Ditto.
1255 * config/i386/sse.md (V8FI): New.
1256 (V16FI): Ditto.
1257 (reduc_splus_v8df): Ditto.
1258 (reduc_splus_v16sf): Ditto.
1259 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1260 (vec_extract_hi_<mode>): Ditto.
1261 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1262 (vec_set_lo_<mode>): Ditto.
1263 (vec_set_hi_<mode>): Ditto.
1264 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1265 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1266 (avx512f_pshufd_1): Ditto.
1267 (avx512f_broadcast<mode>): Ditto.
1268 (avx512f_broadcast<mode>): Ditto.
1269 (define_split): Split vec_extract_lo into move.
1270 (ssequartermode): Ditto.
1271 (ssedoublemode): Extened with wider modes.
1272 (vec_extract_lo_<mode>): Ditto.
1273
0fe65b75
AI		 12742013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1275 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1276 Sergey Lega <sergey.s.lega@intel.com>
1277 Anna Tikhonova <anna.tikhonova@intel.com>
1278 Ilya Tocar <ilya.tocar@intel.com>
1279 Andrey Turetskiy <andrey.turetskiy@intel.com>
1280 Ilya Verbin <ilya.verbin@intel.com>
1281 Kirill Yukhin <kirill.yukhin@intel.com>
1282 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1283
1284 * config/i386/predicates.md (register_or_constm1_operand): New.
1285 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1286 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1287 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1288 UNSPEC_SCATTER_PREFETCH
1289 (VI48_512): New.
1290 (avx512f_ucmp<mode>3): Ditto.
1291 (avx512f_vternlog<mode>): Ditto.
1292 (avx512f_align<mode>): Ditto.
1293 (<shift_insn><mode>3): Ditto.
1294 (avx512f_<rotate>v<mode>): Ditto.
1295 (avx512f_<rotate><mode>): Ditto.
1296 (avx512f_eq<mode>3): Ditto.
1297 (avx512f_eq<mode>3_1): Ditto.
1298 (avx512f_gt<mode>3): Ditto.
1299 (avx512f_testm<mode>3): Ditto.
1300 (avx512f_testnm<mode>3): Ditto.
1301 (avx512pf_gatherpf<mode>): Ditto.
1302 (*avx512pf_gatherpf<mode>_mask): Ditto.
1303 (*avx512pf_gatherpf<mode>): Ditto.
1304 (avx512pf_scatterpf<mode>): Ditto.
1305 (*avx512pf_scatterpf<mode>_mask): Ditto.
1306 (*avx512pf_scatterpf<mode>): Ditto.
1307 (avx512f_vec_dup_gpr<mode>): Ditto.
1308 (clz<mode>2): Ditto.
1309 (conflict<mode>): Ditto.
1310 (REDUC_SMINMAX_MODE): Extened with wider modes.
1311 (reduc_<code>_<mode>): Ditto.
1312 (vlshr<mode>3): Ditto.
1313 (vashl<mode>3): Ditto.
1314
ab931c71
AI		 13152013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1316 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1317 Sergey Lega <sergey.s.lega@intel.com>
1318 Anna Tikhonova <anna.tikhonova@intel.com>
1319 Ilya Tocar <ilya.tocar@intel.com>
1320 Andrey Turetskiy <andrey.turetskiy@intel.com>
1321 Ilya Verbin <ilya.verbin@intel.com>
1322 Kirill Yukhin <kirill.yukhin@intel.com>
1323 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1324
1325 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1326 UNSPEC_SCATTER.
1327 (VI48F_512): New.
1328 (avx512fmaskmode): Ditto.
1329 (bcstscalarsuff): Ditto.
1330 (avx512f_blendm<mode>): Ditto.
1331 (cmp_imm_predicate): Ditto.
1332 (avx512f_cmp<mode>3): Ditto.
1333 (avx512f_vec_dup<mode>): Ditto.
1334 (avx512f_vec_dup_mem<mode>): Ditto.
1335 (avx512f_vpermi2var<mode>3): Ditto.
1336 (avx512f_vpermt2var<mode>3): Ditto.
1337 (vec_init<mode>): Ditto.
1338 (avx512f_gathersi<mode>): Ditto.
1339 (*avx512f_gathersi<mode>): Ditto.
1340 (*avx512f_gathersi<mode>_2): Ditto.
1341 (avx512f_gatherdi<mode>): Ditto.
1342 (*avx512f_gatherdi<mode>): Ditto.
1343 (*avx512f_gatherdi<mode>_2): Ditto.
1344 (avx512f_scattersi<mode>): Ditto.
1345 (*avx512f_scattersi<mode>): Ditto.
1346 (avx512f_scatterdi<mode>): Ditto.
1347 (*avx512f_scatterdi<mode>): Ditto.
1348 (sseintprefix): Extened with wider modes.
1349 (VEC_GATHER_IDXSI): Ditto.
1350 (VEC_GATHER_IDXDI): Ditto.
1351 (VEC_GATHER_SRCDI): Ditto.
1352
c9acb877
MGD		 13532013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1354 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1355
1356 * config/arm/t-aprofile: New file.
1357 * config.gcc: Handle --with-multilib-list option.
1358
ecbda484
BS		 13592013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1360
6f1abb55
UB		 1361 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1362 the input for r2.
ecbda484 1363
24f161fd
RB		 13642013-10-15 Richard Biener <rguenther@suse.de>
1365
1366 * tree-loop-distribution.c (build_empty_rdg): Inline into
1367 single user.
1368 (rdg_flag_vertex): Inline into single user.
1369 (rdg_flag_vertex_and_dependent): Likewise.
1370 (build_rdg_partition_for_vertex): Remove processed bitmap.
1371 (rdg_build_partitions): Simplify.
1372
5de989ed
RB		 13732013-10-15 Richard Biener <rguenther@suse.de>
1374
1375 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1376 Restructure forwarding through conversions and copies to
1377 avoid performing copy-propagation the wrong way. Adjust
1378 recursion invocations.
1379 (forward_propagate_addr_expr): Add argument stating if we
1380 are recursing from a single-use.
1381 (ssa_forward_propagate_and_combine): Adjust.
1382
47e0da37
DM		 13832013-10-14 David Malcolm <dmalcolm@redhat.com>
1384
1385 * dumpfile.h (gcc::dump_manager): New class, to hold state
1386 relating to dumpfile management.
1387 (get_dump_file_name): Remove in favor of method of dump_manager.
1388 (dump_initialized_p): Likewise.
1389 (dump_start): Likewise.
1390 (dump_finish): Likewise.
1391 (dump_switch_p): Likewise.
1392 (dump_register): Likewise.
1393 (get_dump_file_info): Likewise.
1394 * context.c (gcc::context::context): Construct the dump_manager
1395 instance.
1396 * context.h (gcc::context::get_dumps): New.
1397 (gcc::context::m_dumps): New.
1398 * coverage.c (coverage_init): Port to dump_manager API.
6f1abb55 1399 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
47e0da37
DM		 1400 (extra_dump_files_in_use): Likewise.
1401 (extra_dump_files_alloced): Likewise.
1402 (gcc::dump_manager::dump_manager): New.
1403 (dump_register): Convert to...
1404 (gcc::dump_manager::dump_register): ...method, replacing
1405 function-static next_dump with m_next_dump field.
1406 (get_dump_file_info): Convert to...
1407 (gcc::dump_manager::get_dump_file_info): ...method.
1408 (get_dump_file_name): Convert to...
1409 (gcc::dump_manager::get_dump_file_name): ...method.
1410 (dump_start): Convert to...
1411 (gcc::dump_manager::dump_start): ...method.
1412 (dump_finish): Convert to...
1413 (gcc::dump_manager::dump_finish): ...method.
1414 (dump_begin): Replace body with...
1415 (gcc::dump_manager::dump_begin): ...new method.
1416 (dump_phase_enabled_p): Convert to...
1417 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1418 (dump_phase_enabled_p): Convert to...
1419 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1420 (dump_initialized_p): Convert to...
1421 (gcc::dump_manager::dump_initialized_p): ...method.
1422 (dump_flag_name): Replace body with...
1423 (gcc::dump_manager::dump_flag_name): ...new method.
1424 (dump_enable_all): Convert to...
1425 (gcc::dump_manager::dump_enable_all): ...new method.
1426 (opt_info_enable_passes): Convert to...
1427 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1428 (dump_switch_p_1): Convert to...
1429 (gcc::dump_manager::dump_switch_p_1): ...new method.
1430 (dump_switch_p): Convert to...
1431 (gcc::dump_manager::dump_switch_p): ...new method.
1432 (opt_info_switch_p): Port to dump_manager API.
1433 (enable_rtl_dump_file): Likewise.
1434 * opts-global.c (handle_common_deferred_options): Port to new
1435 dump_manager API.
1436 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1437 (pass_manager::register_one_dump_file): Likewise.
1438 (pass_manager::register_pass): Likewise.
1439 (pass_init_dump_file): Likewise.
1440 (pass_fini_dump_file): Likewise.
1441 * statistics.c (statistics_early_init): Likewise.
1442
83fd5d11
RB		 14432013-10-14 Richard Biener <rguenther@suse.de>
1444
1445 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1446 iterative_hash_canonical_type, gimple_canonical_type_hash,
1447 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1448 gimple_register_canonical_type, print_gimple_types_stats,
1449 free_gimple_type_tables): Move to lto/lto.c
1450 (gt-gimple.h): Do not include.
1451 * gimple.h (gimple_register_canonical_type,
1452 print_gimple_types_stats, free_gimple_type_tables): Remove.
1453 * Makefile.in (GTFILES): Remove gimple.c.
1454
a6dbd0f7
TS		 14552013-10-14 Travis Snoozy <quandary@remstate.com>
1456
1457 PR target/58716
1458 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1459 scanning for msp430x targets.
1460
b5ccb9ed
EB		 14612013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1462
1463 PR bootstrap/58509
1464 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1465 (registers_ok_for_ldd_peep): Move around.
1466 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1467 * config/sparc/sparc.md (widening peepholes): Use it.
1468
fd7e5084
RB		 14692013-10-14 Richard Biener <rguenther@suse.de>
1470
1471 PR middle-end/58712
1472 PR middle-end/55358
1473 * gimple.c (iterative_hash_canonical_type): Make sure to
1474 record the hash into the correct hashtable slot.
1475
3246afaa
EB		 14762013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1477
1478 PR rtl-optimization/58662
1479 * combine.c (try_combine): Take into account death nodes on I2 when
1480 splitting a PARALLEL of two independent SETs. Fix dump message.
1481
5d30dc5b
OE		 14822013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1483
1484 PR target/51244
1485 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1486 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1487 * config/sh/t-sh (sh_treg_combine.o): New entry.
1488 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1489 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1490 (register_sh_passes): New function. Register sh_treg_combine pass.
1491 (sh_option_override): Invoke it.
1492 (sh_canonicalize_comparison): Handle op0_preserve_value.
1493 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1494 opportunities. Canonicalize branch condition.
1495 (nott): Allow only if pseudos can be created for non-SH2A.
1496
37d1c5db
L		 14972013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1498
1499 PR target/58690
1500 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1501 (ix86_expand_movmem): Replace copy_addr_to_reg with
1502 ix86_copy_addr_to_reg.
1503 (ix86_expand_setmem): Likewise.
1504
b86a70ff
AM		 15052013-10-12 Alexander Monakov <amonakov@ispras.ru>
1506
1507 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1508 provided by ix86_fp_compare_mode instead of CCFPUmode.
1509
6e1d5961
JG		 15102013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1511
1512 * config/aarch64/arm_neon.h
1513 (vtbx<1,3>_<psu>8): Fix register constriants.
1514
ef3cfba2
JL		 15152013-10-11 Jeff Law <law@redhat.com>
1516
1517 PR tree-optimization/58640
adcfd489
UB		 1518 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1519 threading paths that cross over two loop entry points.
ef3cfba2 1520
6e8b7d9c
BS		 15212013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1522
1523 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1524 handle vector float as well.
1525 (*vsx_le_perm_load_v4si): Likewise.
1526 (*vsx_le_perm_store_v2di): Likewise.
1527 (*vsx_le_perm_store_v4si): Likewise.
1528
f200869a
BS		 15292013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1530
1531 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1532 directly to circumvent subtract from splat{31} workaround.
1533 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1534 prototype.
1535 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1536 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1537 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1538 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1539 separate big and little endian logic.
1540 (*altivec_vperm_<mode>_internal): New define_insn.
1541 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1542 separate big and little endian logic.
1543 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1544 (vec_permv16qi): Add little endian logic.
1545
8d28e3fc
MG		 15462013-10-11 Marc Glisse <marc.glisse@inria.fr>
1547
1548 * doc/extend.texi (returns_nonnull): Remove arguments.
1549
3c87b77b
AI		 15502013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1551 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1552 Sergey Lega <sergey.s.lega@intel.com>
1553 Anna Tikhonova <anna.tikhonova@intel.com>
1554 Ilya Tocar <ilya.tocar@intel.com>
1555 Andrey Turetskiy <andrey.turetskiy@intel.com>
1556 Ilya Verbin <ilya.verbin@intel.com>
1557 Kirill Yukhin <kirill.yukhin@intel.com>
1558 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1559
1560 * config/i386/sse.md (VI48F_256_512): New.
1561 (avx2_permvar<mode>): Change to ...
1562 (<avx2_avx512f>_permvar<mode>): This.
1563
50e60d7d
AI		 15642013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1565 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1566 Sergey Lega <sergey.s.lega@intel.com>
1567 Anna Tikhonova <anna.tikhonova@intel.com>
1568 Ilya Tocar <ilya.tocar@intel.com>
1569 Andrey Turetskiy <andrey.turetskiy@intel.com>
1570 Ilya Verbin <ilya.verbin@intel.com>
1571 Kirill Yukhin <kirill.yukhin@intel.com>
1572 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1573
6f1abb55 1574 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
50e60d7d
AI		 1575 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1576 * config/i386/sse.md (VI4_AVX): New.
1577 (sf2simodelower): Ditto.
1578 (sse2_cvtps2dq): Change to ...
1579 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1580
f62ce24f
AI		 15812013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1582 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1583 Sergey Lega <sergey.s.lega@intel.com>
1584 Anna Tikhonova <anna.tikhonova@intel.com>
1585 Ilya Tocar <ilya.tocar@intel.com>
1586 Andrey Turetskiy <andrey.turetskiy@intel.com>
1587 Ilya Verbin <ilya.verbin@intel.com>
1588 Kirill Yukhin <kirill.yukhin@intel.com>
1589 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1590
1591 * config/i386/sse.md (V_512): New.
1592 (VI_512): Ditto.
1593 (vcond<V_512:mode><VF_512:mode>): Ditto.
1594 (vcond<V_512:mode><VI_512:mode>): Ditto.
1595 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1596
558d9f79
AI		 15972013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1598 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1599 Sergey Lega <sergey.s.lega@intel.com>
1600 Anna Tikhonova <anna.tikhonova@intel.com>
1601 Ilya Tocar <ilya.tocar@intel.com>
1602 Andrey Turetskiy <andrey.turetskiy@intel.com>
1603 Ilya Verbin <ilya.verbin@intel.com>
1604 Kirill Yukhin <kirill.yukhin@intel.com>
1605 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1606
1607 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1608 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1609 (FMAMODE): Ditto.
1610 (fma<mode>4): Removed condition.
1611 (fms<mode>4): Ditto.
1612 (fnma<mode>4): Ditto.
1613 (fnms<mode>4): Ditto.
1614 (fma4i_fmadd_<mode>): Ditto.
1615 (*fma_fmadd_<mode>): Ditto.
1616 (*fma_fmsub_<mode>): Ditto.
1617 (*fma_fnmadd_<mode>): Ditto.
1618 (*fma_fnmsub_<mode>): Ditto.
1619 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1620 (*fma_fmaddsub_<mode>): Ditto.
1621 (*fma_fmsubadd_<mode>): Ditto.
1622 (*fmai_fmadd_<mode>): Ditto.
1623 (*fmai_fmsub_<mode>): Ditto.
1624 (*fmai_fnmadd_<mode>): Ditto.
1625 (*fmai_fnmsub_<mode>): Ditto.
1626
e8d08206
AI		 16272013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1628 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1629 Sergey Lega <sergey.s.lega@intel.com>
1630 Anna Tikhonova <anna.tikhonova@intel.com>
1631 Ilya Tocar <ilya.tocar@intel.com>
1632 Andrey Turetskiy <andrey.turetskiy@intel.com>
1633 Ilya Verbin <ilya.verbin@intel.com>
1634 Kirill Yukhin <kirill.yukhin@intel.com>
1635 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1636
1637 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1638 (VI124_256): Changed to ...
1639 (VI124_256_48_512): This.
1640 (ssepackmode): Extended with wider modes.
1641 (<code><mode>3): Changed iterator.
1642 (*avx2_<code><mode>3): Ditto.
1643 (vec_pack_trunc_<mode>): Ditto.
1644
16821545
AI		 16452013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1646 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1647 Sergey Lega <sergey.s.lega@intel.com>
1648 Anna Tikhonova <anna.tikhonova@intel.com>
1649 Ilya Tocar <ilya.tocar@intel.com>
1650 Andrey Turetskiy <andrey.turetskiy@intel.com>
1651 Ilya Verbin <ilya.verbin@intel.com>
1652 Kirill Yukhin <kirill.yukhin@intel.com>
1653 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1654
1655 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1656 (VI8F_256_512): Ditto.
1657 (abs<mode>2): Changed iterator.
1658 (avx2_perm<mode>): Changed to ...
1659 (<avx2_avx512f>_perm<mode>): This.
1660 (avx2_perm<mode>_1): Changed to ...
1661 (<avx2_avx512f>_perm<mode>_1): This.
1662
5348cff8
AI		 16632013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1664 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1665 Sergey Lega <sergey.s.lega@intel.com>
1666 Anna Tikhonova <anna.tikhonova@intel.com>
1667 Ilya Tocar <ilya.tocar@intel.com>
1668 Andrey Turetskiy <andrey.turetskiy@intel.com>
1669 Ilya Verbin <ilya.verbin@intel.com>
1670 Kirill Yukhin <kirill.yukhin@intel.com>
1671 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1672
1673 * config/i386/sse.md (VI48_AVX512F): New.
1674 (VI48_AVX2): Changed to ...
1675 (VI48_AVX2_48_AVX512F): This.
1676 (avx2_ashrv<mode>): Changed to ...
1677 (<avx2_avx512f>_ashrv<mode>): This.
1678 (avx2_<shift_insn>v<mode>): Changed to ...
1679 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1680
f5f41d88
AI		 16812013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1682 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1683 Sergey Lega <sergey.s.lega@intel.com>
1684 Anna Tikhonova <anna.tikhonova@intel.com>
1685 Ilya Tocar <ilya.tocar@intel.com>
1686 Andrey Turetskiy <andrey.turetskiy@intel.com>
1687 Ilya Verbin <ilya.verbin@intel.com>
1688 Kirill Yukhin <kirill.yukhin@intel.com>
1689 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1690
1691 * config/i386/sse.md (VI4_AVX512F): New.
1692 (VI8_AVX2_AVX512F): Ditto.
1693 (mul<mode>3): Extended with wider modes.
1694 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1695 (mul<mode>3): Ditto.
1696 (vec_widen_<s>mult_odd_<mode>): Ditto.
1697
3bdf6340
AI		 16982013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1699 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1700 Sergey Lega <sergey.s.lega@intel.com>
1701 Anna Tikhonova <anna.tikhonova@intel.com>
1702 Ilya Tocar <ilya.tocar@intel.com>
1703 Andrey Turetskiy <andrey.turetskiy@intel.com>
1704 Ilya Verbin <ilya.verbin@intel.com>
1705 Kirill Yukhin <kirill.yukhin@intel.com>
1706 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1707
1708 * config/i386/sse.md (VI2_AVX512F): New.
1709 (VI124_AVX512F): Ditto.
1710 (sseunpackmode): Extended with wider modes.
1711 (sseunpackfltmode): Ditto.
1712 (vec_unpacks_float_hi_<mode>): Ditto.
1713 (vec_unpacks_float_lo_<mode>): Ditto.
1714 (vec_unpacku_float_hi_<mode>): Ditto.
1715 (vec_unpacku_float_lo_<mode>): Ditto.
1716 (vec_unpacks_lo_<mode>): Ditto.
1717 (vec_unpacks_hi_<mode>): Ditto.
1718 (vec_unpacku_lo_<mode>): Ditto.
1719 (vec_unpacku_hi_<mode>): Ditto.
1720
67f783cb
AI		 17212013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1722 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1723 Sergey Lega <sergey.s.lega@intel.com>
1724 Anna Tikhonova <anna.tikhonova@intel.com>
1725 Ilya Tocar <ilya.tocar@intel.com>
1726 Andrey Turetskiy <andrey.turetskiy@intel.com>
1727 Ilya Verbin <ilya.verbin@intel.com>
1728 Kirill Yukhin <kirill.yukhin@intel.com>
1729 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1730
1731 * config/i386/i386.md (multdiv): New.
1732 (multdiv_mnemonic): Ditto.
1733 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1734 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1735 (<sse>_vmdiv<mode>3): Removed.
1736
ec5e777c
AI		 17372013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1738 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1739 Sergey Lega <sergey.s.lega@intel.com>
1740 Anna Tikhonova <anna.tikhonova@intel.com>
1741 Ilya Tocar <ilya.tocar@intel.com>
1742 Andrey Turetskiy <andrey.turetskiy@intel.com>
1743 Ilya Verbin <ilya.verbin@intel.com>
1744 Kirill Yukhin <kirill.yukhin@intel.com>
1745 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1746
1747 * config/i386/sse.md (V): Extended with wider modes.
1748 (VF2): Ditto.
1749 (ssehalfvecmode): Ditto.
1750 (i128): Ditto.
1751 (ssepackfltmode): Ditto.
1752 (avx_vec_concat<mode>): Ditto.
1753 (V_256_512): New iterator.
1754 (VF2_512_256): Ditto.
1755 (si2dfmode): New attribute.
1756 (si2dfmodelower): Ditto.
1757 (sf2dfmode): Ditto.
1758 (concat_tg_mode): Ditto.
1759 (floatv4siv4df2): Changed to ...
1760 (float<si2dfmodelower><mode>2): This.
1761 (avx_cvtps2pd256): Changed to ...
1762 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1763 (vec_pack_trunc_v4df): Changed to ...
1764 (vec_pack_trunc_<mode>): This.
1765 (avx_vpermil<mode>): Changed to ...
1766 (<sse2_avx_avx512f>_vpermil<mode>): This.
1767 (<fixsuffix>fix_truncv8dfv8si2): New.
1768 (vec_pack_sfix_trunc_v8df): Ditto.
1769 (avx512f_rndscale<mode>): Ditto.
1770 (avx512f_roundpd512): Ditto.
1771 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1772
a9ccbba2
AI		 17732013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1774 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1775 Sergey Lega <sergey.s.lega@intel.com>
1776 Anna Tikhonova <anna.tikhonova@intel.com>
1777 Ilya Tocar <ilya.tocar@intel.com>
1778 Andrey Turetskiy <andrey.turetskiy@intel.com>
1779 Ilya Verbin <ilya.verbin@intel.com>
1780 Kirill Yukhin <kirill.yukhin@intel.com>
1781 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1782
1783 * config/i386/i386.md (any_fix): New iterator.
1784 (fixsuffix): New attribute.
1785 * config/i386/sse.md (VF1): Extened with wider modes.
1786 (VI): Ditto.
1787 (VI_AVX2): Ditto.
1788 (VI8): Ditto.
1789 (sseintvecmodelower): Ditto.
1790 (ssescalarmode): Ditto.
1791 (ssescalarnum): Ditto.
1792 (VF1_128_256): New.
1793 (ssexmmmode): Ditto.
1794 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1795 (<sse>_rcp<mode>2): Change iterator.
1796 (rsqrt<mode>2): Ditto.
1797 (<sse>_rsqrt<mode>2): Ditto.
1798 (avx2_vec_dup<mode>): Ditto.
1799 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1800 (round<mode>2_sfix): Ditto.
1801 (avx2_pbroadcast<mode>): Ditto.
1802 (*andnot<mode>3): Handle XI mode.
1803 (*<code><mode>3): Ditto.
1804 (AVXTOSSEMODE): Removed.
1805 (avx_vpermil<mode>): Changed to ...
1806 (<sse2_avx_avx512f>_vpermil<mode>): This.
1807
2b1ebb0c
AI		 18082013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1809 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1810 Sergey Lega <sergey.s.lega@intel.com>
1811 Anna Tikhonova <anna.tikhonova@intel.com>
1812 Ilya Tocar <ilya.tocar@intel.com>
1813 Andrey Turetskiy <andrey.turetskiy@intel.com>
1814 Ilya Verbin <ilya.verbin@intel.com>
1815 Kirill Yukhin <kirill.yukhin@intel.com>
1816 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1817
1818 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1819 (<sse>_comi): Ditto.
1820 (<sse>_ucomi): Ditto.
1821 (sse_cvtss2siq_2): Ditto.
1822 (sse2_cvtsd2si): Ditto.
1823 (sse2_cvtsd2siq): Ditto.
1824 (sse2_cvttsd2si): Ditto.
1825 (sse2_cvttsd2siq): Ditto.
1826 (<shift_insn><mode>3): Ditto.
1827 (sse2_cvtsi2sdq): Update constraint and prefix.
1828 (sse_cvtsi2ss): Update prefix.
1829 (sse_cvtsi2ssq): Ditto.
1830
d89124ed
JJ		 18312013-10-11 Jakub Jelinek <jakub@redhat.com>
1832
1833 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1834 ignore internal calls.
1835
e6ad28c3
RB		 18362013-10-11 Richard Biener <rguenther@suse.de>
1837
6f1abb55
UB		 1838 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1839 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1840 TREE_OVERFLOW is set.
e6ad28c3 1841
d1227447
TS		 18422013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1843
1ac4b8e4
TS		 1844 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1845
7d05cebb
TS		 1846 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1847 * gimple.h: Likewise.
1848
adcfd489 1849 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 1850
1a79fb8a
TS		 1851 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1852 macros.
1853
3afd2873
TS		 1854 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1855
adcfd489 1856 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS		 1857 * configure: Regenerate.
1858
acf0174b
JJ		 18592013-10-11 Jakub Jelinek <jakub@redhat.com>
1860
1861 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1862 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1863 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1864 similarly pp_right_paren.
1865 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1866 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1867 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1868 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1869 clauses.
1870 (omp_declare_simd_clauses_equal,
1871 omp_remove_redundant_declare_simd_attrs): New functions.
1872 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1873 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1874 * tree.h (OMP_LOOP_CHECK): Define.
1875 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1876 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1877 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1878 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1879 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1880 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1881 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1882 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1883 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1884 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1885 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1886 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1887 (omp_remove_redundant_declare_simd_attrs): New prototype.
1888 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1889 GIMPLE_OMP_TEAMS): New codes.
1890 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1891 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1892 fields.
1893 (target_nesting_level): New variable.
1894 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1895 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1896 for collapse > 1 static schedule unless ordered.
1897 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1898 (determine_parallel_type): Adjust get_ws_args_for caller.
1899 (install_var_field): Handle mask & 4 for double indirection.
1900 (scan_sharing_clauses): Ignore shared clause on teams construct.
1901 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1902 (create_omp_child_function): If inside target or declare target
adcfd489 1903 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ		 1904 (find_combined_for): New function.
1905 (scan_omp_parallel): Handle combined loops.
1906 (scan_omp_target, scan_omp_teams): New functions.
1907 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1908 restrictions and set ctx->cancellable for cancellable constructs.
1909 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1910 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1911 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 1912 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b 1913 (omp_clause_aligned_alignment): New function.
6f1abb55 1914 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
acf0174b
JJ		 1915 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1916 on teams constructs. Handle user defined reductions and new
1917 OpenMP 4.0 clauses.
1918 (lower_reduction_clauses): Don't set placeholder to address of ref
1919 if it has already the right type.
1920 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1921 (expand_parallel_call): Use the new non-_start suffixed builtins,
1922 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1923 and GOMP_parallel_end after the call.
1924 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1925 (expand_omp_for_init_counts): Handle combined loops.
1926 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1927 loops.
1928 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1929 end of cancellable loops.
1930 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1931 Likewise. Handle collapse > 1 loops.
1932 (expand_omp_simd): Handle combined loops.
1933 (expand_omp_for): Add inner_stmt argument, adjust callers of
1934 expand_omp_for* functions, use expand_omp_for_static*chunk even
1935 for collapse > 1 unless ordered.
1936 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1937 of cancellable sections.
1938 (expand_omp_single): Remove need_barrier variable, just rely on
1939 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1940 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1941 (expand_omp_atomic_load, expand_omp_atomic_store,
1942 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1943 (expand_omp_target): New function.
1944 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1945 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1946 (build_omp_regions_1): Immediately close region for
1947 GF_OMP_TARGET_KIND_UPDATE.
1948 (maybe_add_implicit_barrier_cancel): New function.
1949 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1950 cancellation.
1951 (lower_omp_single): Likewise. Add clobber after the barrier.
1952 (lower_omp_taskgroup): New function.
1953 (lower_omp_for): Handle combined loops. Adjust
1954 lower_rec_input_clauses caller. Handle cancellation.
1955 (lower_depend_clauses): New function.
1956 (lower_omp_taskreg): Lower depend clauses. Adjust
1957 lower_rec_input_clauses caller. Add clobber after the call. Handle
1958 cancellation.
1959 (lower_omp_target, lower_omp_teams): New functions.
1960 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1961 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1962 and GOMP_cancellation_point calls.
1963 (lower_omp): Fold stmts inside of target region.
1964 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1965 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1966 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1967 (BT_FN_VOID_OMPFN_PTR_UINT,
1968 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1969 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1970 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1971 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1972 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1973 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1974 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1975 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1976 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1977 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1978 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1979 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1980 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1981 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1982 BUILT_IN_GOMP_PARALLEL_END.
1983 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1984 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1985 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1986 GF_OMP_FOR_KIND_DISTRIBUTE.
1987 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1988 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1989 (dump_gimple_omp_return): Print lhs if it has any.
1990 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1991 gimple_omp_atomic_seq_cst_p.
1992 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1993 and GIMPLE_OMP_TEAMS.
1994 * langhooks.c (lhd_omp_mappable_type): New function.
1995 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1996 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1997 hook.
1998 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1999 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6f1abb55 2000 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
acf0174b
JJ		 2001 (struct gimplify_omp_ctx): Add combined_loop field.
2002 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2003 on stmts inside of target region.
2004 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2005 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2006 ORT_TARGET and ORT_TARGET_DATA.
2007 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2008 Handle GOVD_MAP.
2009 (omp_notice_threadprivate_variable): Complain about threadprivate
2010 variables in target region.
2011 (omp_notice_variable): Complain about vars with non-mappable type
2012 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2013 (omp_check_private): Ignore ORT_TARGET* regions.
2014 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2015 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2016 (find_combined_omp_for): New function.
2017 (gimplify_omp_for): Handle gimplification of combined loops.
2018 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2019 OMP_TEAMS.
2020 (gimplify_omp_target_update): New function.
2021 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2022 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2023 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2024 (gimplify_body): If fndecl has "omp declare target" attribute, add
2025 implicit ORT_TARGET context around it.
2026 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2027 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2028 * tree-nested.c (convert_nonlocal_reference_stmt,
2029 convert_local_reference_stmt, convert_gimple_call): Handle
2030 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2031 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2032 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2033 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2034 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2035 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2036 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2037 BUILT_IN_GOMP_LOOP_END_CANCEL,
2038 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2039 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2040 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2041 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2042 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2043 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2044 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2045 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2046 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2047 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2048 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2049 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2050 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2051 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2052 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2053 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2054 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2055 gimple_build_omp_teams): New functions.
2056 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2057 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2058 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2059 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2060 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2061 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2062 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2063 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2064 GF_OMP_ATOMIC_SEQ_CST): New.
2065 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2066 gimple_build_omp_teams): New prototypes.
2067 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2068 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2069 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2070 GIMPLE_OMP_SINGLE as end of range.
2071 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2072 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2073 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2074 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2075 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2076 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2077 gimple_omp_target_kind, gimple_omp_target_set_kind,
2078 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2079 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2080 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2081 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2082 gimple_omp_teams_set_clauses): New inlines.
2083 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2084 and GIMPLE_OMP_TASKGROUP.
adcfd489 2085 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ		 2086 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2087 enum omp_clause_proc_bind_kind): New.
2088 (union omp_clause_subcode): Add depend_kind, map_kind and
2089 proc_bind_kind fields.
2090 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2091 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2092 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2093 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2094 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2095
67fa7880
TJ		 20962013-10-10 Teresa Johnson <tejohnson@google.com>
2097
2098 * predict.c (tree_estimate_probability): Add new parameter
2099 for estimate_bb_frequencies.
2100 (estimate_bb_frequencies): Add new parameter to force estimation.
2101 (rebuild_frequencies): When max frequency in function is small,
2102 recompute counts from frequencies.
2103 * predict.h (estimate_bb_frequencies): New parameter.
2104
b591a8b7
DM		 21052013-10-10 David Malcolm <dmalcolm@redhat.com>
2106
2107 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2108 optimizations are disabled.
2109
782f0db2
DM		 21102013-10-10 David Malcolm <dmalcolm@redhat.com>
2111
2112 * coverage.c (coverage_finish): Fix leak of da_file_name.
2113
13ef00fa
JH		 21142013-10-10 Jan Hubicka <jh@suse.cz>
2115
2116 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2117 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2118 Bobcat and generic.
2119
a1d50386
JJ		 21202013-10-10 Jakub Jelinek <jakub@redhat.com>
2121
2122 PR middle-end/58670
2123 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2124 if any labels are in FALLTHRU_BB, use a special label emitted
2125 immediately after the asm goto insn rather than label_rtx
2126 of the LABEL_DECL.
2127 (expand_asm_stmt): Adjust caller.
2128 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2129 edge if the last insn in predecessor is a jump with single successor,
2130 but it isn't simplejump_p.
2131
e2c2fde2
RB		 21322013-10-10 Richard Biener <rguenther@suse.de>
2133
2134 PR tree-optimization/58656
6f1abb55 2135 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
e2c2fde2 2136
de6a5608
AM		 21372013-10-10 Andrew MacLeod <amacleod@redhat.com>
2138
adcfd489
UB		 2139 * gimplify.c: Include expr.h and tm_p.h for targets with special
2140 va-arg padding requirements.
de6a5608 2141
b184c8f1
AM		 21422013-10-10 Andrew MacLeod <amacleod@redhat.com>
2143
2144 * tree-flow.h: Move some prototypes to gimple.h.
2145 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2146 * gimple.h: Relocate some prototypes from tree-flow.h
2147 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2148 Move to gimplify.c.
2149 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2150 (build_va_arg_indirect_ref): Relocate and make static.
2151 (std_gimplify_va_arg_expr): Relocate here.
2152 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2153 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2154
13b72c22
AK		 21552013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2156
2157 * doc/md.texi: Document the mnemonic attribute.
2158
5db40447
AK		 21592013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2160
2161 PR target/57377
2162 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2163 (set_attr_alternative x ...) when searching for user defined
2164 mnemonic attribute.
2165
ed018d4b
AM		 21662013-10-10 Andrew MacLeod <amacleod@redhat.com>
2167
2168 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2169
4d5b5e9f
ER		 21702013-10-09 Easwaran Raman <eraman@google.com>
2171
2172 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2173 * cfgexpand.c (defer_stack_allocation): ...use here
2174 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2175
c9ef86a1
ZC		 21762013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2177
2178 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2179 (operand_equal_for_value_replacement): New function, extracted from
2180 value_replacement and enhanced to catch more cases.
2181 (value_replacement): Use operand_equal_for_value_replacement.
2182
2431114f
AM		 21832013-10-09 Andrew MacLeod <amacleod@redhat.com>
2184
adcfd489 2185 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM		 2186 get_max_loop_iterations.
2187
1dd2a9a0
KT		 21882013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2189
2190 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2191 Place comment above function.
2192
c1bf2a39
AM		 21932013-10-09 Andrew MacLeod <amacleod@redhat.com>
2194
2195 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2196 are not related to tree-cfg.c.
2197 * tree-ssa-address.h: New file. Relocate prototypes.
2198 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2199 (addr_for_mem_ref): New. Combine call to get_address_description and
2200 return addr_for_mem_ref.
2201 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2202 * tree-ssa-live.h: Adjust prototypes.
2203 * passes.c: Include tree-ssa-live.h.
2204 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2205 * graphite.c (graphite_transform_loops): Make static.
2206 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 2207 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM		 2208 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2209 * ipa-pure-const.c (warn_function_noreturn): Make static.
2210 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2211 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2212 Relocate from tree-cfg.c
2213 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2214 static.
2215 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2216 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2217 Move to ipa-pure-const.c.
adcfd489
UB		 2218 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2219 Relocate from tree-optimize.c.
c1bf2a39
AM		 2220 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2221 make_pass_fixup_cfg): Move to tree-cfg.c.
2222 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2223 Relocate some prototypes.
2224 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2225 * tree-dump.c (dump_function_to_file): Remove prototype.
2226 Add tree-flow.h to the include file.
2227 * tree-dump.h: Remove prototype.
2228 * tree-parloops.h: New File. Add prototypes.
2229 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2230 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2231 from tree-ssa-loop.c.
2232 * tree-predcom.c (run_tree_predictive_commoning,
2233 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2234 Relocate here from tree-ssa-loop.c.
adcfd489 2235 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM		 2236 ssa_name_values.release ().
2237 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2238 (ssa_name_values): Relocate from tree-flow.h.
2239 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2240 * tree-ssa-loop.c (run_tree_predictive_commoning,
2241 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2242 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2243 make_pass_graphite, pass_data_graphite_transforms,
2244 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2245 tree_parallelize_loops, pass_data_parallelize_loops,
2246 make_pass_parallelize_loops): Move to other files.
2247 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2248 moved here.
2249 * tree.h: Remove prototypes from tree-address.c.
2250
cc524fc7
AM		 22512013-10-09 Andrew MacLeod <amacleod@redhat.com>
2252
2253 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2254 (struct int_tree_map): Move to tree-hasher.h
2255 (SCALE, LABEL, PERCENT): Move to gimple.h
2256 * tree-flow-inline.h: Delete. Move functions to other files.
2257 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2258 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2259 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2260 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2261 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2262 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2263 * tree-hasher.h: Don't include tree-flow.h.
2264 (struct int_tree_map): Relocate from tree-flow.h.
2265 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2266 tree-flow-inline.h and make static.
adcfd489
UB		 2267 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2268 tree-flow-inline.h.
cc524fc7
AM		 2269 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2270 tree-flow-inline.h and make static.
2271 * tree.h (is_global_var, may_be_aliased): Relocate from
2272 tree-flow-inline.h.
2273 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2274 * value-prof.c: No longer include tree-flow-inline.h.
2275 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2276
71343877
AM		 22772013-10-09 Andrew MacLeod <amacleod@redhat.com>
2278
2279 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2280 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2281 (enum move_pos): Move to tree-ssa-loop-im.h
2282 * cfgloop.h: Move some prototypes.
2283 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2284 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2285 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2286 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2287 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2288 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2289 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2290 (enum move_pos): Relocate here.
2291 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2292 tree-ssa-loop.c.
2293 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB		 2294 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2295 make_pass_lim): Relocate here from tree-ssa-loop.c.
2296 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM		 2297 tree-ssa-loop.c.
2298 (loop_edge_to_cancel, unloop_loops): Make static.
2299 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2300 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2301 (tree_complete_unroll, gate_tree_complete_unroll,
2302 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2303 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2304 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2305 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2306 (stmt_invariant_in_loop_p): Remove unused function.
2307 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2308 * tree-ssa-loop-manip.h: New file. Add prototypes.
2309 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2310 (gcov_type_to_double_int): Move to cfgloop.h.
2311 (double_int_cmp, bound_index,
2312 estimate_numbers_of_iterations_loop): Make static.
2313 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2314 (max_loop_iterations): Factor out get_max_loop_iterations.
2315 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2316 cfgloop.c.
2317 * tree-ssa-loop-niter.h: New file. Add prototypes.
2318 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2319 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2320 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2321 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2322 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2323 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2324 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2325 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2326 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2327 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2328 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2329 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2330 pass_data_complete_unroll, make_pass_complete_unroll,
2331 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2332 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2333 tree-ssa-loop-ivcanon.c.
2334 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2335 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2336 tree-ssa-loop-prefetch.c.
2337 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2338 tree-ssa-loop-im.c.
2339 (get_lsm_tmp_name): Relocate and add suffix parameter.
2340 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2341 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2342 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2343 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 2344 (get_estimated_loop_iterations): Factor out accessor from
71343877 2345 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB		 2346 (get_max_loop_iterations): Factor out accessor from
2347 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM		 2348 * loop-unroll.c (decide_unroll_constant_iterations,
2349 decide_unroll_runtime_iterations, decide_peel_simple,
2350 decide_unroll_stupid): Use new get_* accessors.
2351
826cacfe
MG		 23522013-10-09 Marc Glisse <marc.glisse@inria.fr>
2353
2354 PR tree-optimization/20318
2355 * doc/extend.texi (returns_nonnull): New function attribute.
2356 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2357 attribute.
2358 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2359 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2360
378f8976
EB		 23612013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2362
2363 PR middle-end/58570
2364 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2365 false if both components are bitfields.
2366
0fe04f5c
AV		 23672013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2368
2369 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2370 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2371 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2372 * config/aarch64/aarch64.h
2373 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2374 * config/aarch64/aarch64-simd-builtins.def
2375 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2376
bed9bae4
AV		 23772013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2378
2379 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2380
7f3d8b19
AV		 23812013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2382
2383 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2384
7df625a6
AV		 23852013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2386
2387 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2388 (vneg_f64): New intrinsic.
2389 (vneg_s8): Asm replaced with C.
2390 (vneg_s16): Likewise.
2391 (vneg_s32): Likewise.
2392 (vneg_s64): New intrinsic.
2393 (vnegq_f32): Asm replaced with C.
2394 (vnegq_f64): Likewise.
2395 (vnegq_s8): Likewise.
2396 (vnegq_s16): Likewise.
2397 (vnegq_s32): Likewise.
2398 (vnegq_s64): Likewise.
2399
2b86fca7
RL		 24002013-10-09 Renlin Li <Renlin.Li@arm.com>
2401
2402 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2403
e179df83
AK		 24042013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2405
2406 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2407 packed stack special handling.
2408 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2409 back to fixed stack slots for FPRs saved due to stdarg.
2410
74129172
AK		 24112013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2412
2413 * config/s390/s390.c (s390_frame_info): Restructure function.
2414
6455a49e
AK		 24152013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2416
2417 * config/s390/s390.c (struct s390_frame_layout): New field
2418 gpr_save_slots.
2419 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2420 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2421 regs_ever_clobbered to char*.
adcfd489
UB		 2422 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2423 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK		 2424 (s390_register_info): Enable FPR save slots. Move/Copy some
2425 functionality into ...
2426 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2427 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2428 function.
2429 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2430 not imply a stack frame.
2431 (s390_init_frame_layout): Remove variable clobbered_regs.
2432 (s390_update_register_info): Remove function.
2433 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2434 cannot be used for register renaming.
2435 (s390_hard_regno_scratch_ok): New function.
2436 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2437 (s390_initial_elimination_offset): Change offset calculation of
2438 the return address pointer.
adcfd489 2439 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 2440 (restore_gprs): Set frame related flag.
adcfd489 2441 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK		 2442 (s390_emit_prologue): Call s390_register_info instead of
2443 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2444 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2445 (s390_optimize_prologue): Call s390_optimize_register_info.
2446 Try to remove also FPR slot save/restore INSNs. Remove frame
2447 related flags from restore INSNs.
2448
59d96342
DD		 24492013-10-08 DJ Delorie <dj@redhat.com>
2450
0dc7adad
DD		 2451 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2452 (movhi): Likewise.
2453
59d96342
DD		 2454 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2455 avoid conflict with the MI use of %c.
2456 * config/rl78/rl78-real.md: change %c to %C throughout.
2457 * config/rl78/rl78-virt.md: Likewise.
2458
a81169d8
JH		 24592013-10-08 Jan Hubicka <jh@suse.cz>
2460
2461 * config/i386/i386.c (ix86_option_override_internal): Switch
2462 to SSE math for -ffast-math when target ISA supports SSE2.
2463
481d1b81
AM		 24642013-10-08 Andrew MacLeod <amacleod@redhat.com>
2465
2466 * tree-flow.h: Remove some prototypes.
2467 * tree.h: Remove some protypes, add a couple.
2468 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2469 using_eh_for_cleanups_p): Add interface routines for front ends.
2470 * tree-eh.h: New file. Add protoptyes.
2471 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2472 (add_stmt_to_eh_lp_fn): Make static.
2473 (lower_try_finally): Use new using_eh_for_cleanups_p.
2474 * emit-rtl.c: Include tree-eh.h.
2475 * gimple.h: Include tree-eh.h.
2476
0e6a0e48
MG		 24772013-10-08 Marc Glisse <marc.glisse@inria.fr>
2478
2479 PR tree-optimization/58480
2480 * tree-vrp.c (infer_nonnull_range): New function.
2481 (infer_value_range): Call infer_nonnull_range.
2482
ef6179d1
DC		 24832013-10-08 Dehao Chen <dehao@google.com>
2484
2485 PR tree-optimization/58619
2486 * tree-inline.c (copy_phis_for_bb): Combine location data
2487 only if non-null.
2488
07d964d5
ZC		 24892013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2490
2491 PR target/58423
2492 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2493 RTX_FRAME_RELATED_P on INSN.
2494
19e9b2a3
BS		 24952013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2496
2497 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2498 (altivec_expand_vec_perm_const): Call it.
2499
0cf68694
BS		 25002013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2501
2502 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2503 sequences for LE VSX loads and stores at expand time.
2504 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2505 prototype.
2506 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2507 (rs6000_gen_le_vsx_permute): New.
2508 (rs6000_gen_le_vsx_load): New.
2509 (rs6000_gen_le_vsx_store): New.
2510 (rs6000_gen_le_vsx_move): New.
2511 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2512 (*vsx_le_perm_load_v4si): New.
2513 (*vsx_le_perm_load_v8hi): New.
2514 (*vsx_le_perm_load_v16qi): New.
2515 (*vsx_le_perm_store_v2di): New.
2516 (*vsx_le_perm_store_v4si): New.
2517 (*vsx_le_perm_store_v8hi): New.
2518 (*vsx_le_perm_store_v16qi): New.
2519 (*vsx_xxpermdi2_le_<mode>): New.
2520 (*vsx_xxpermdi4_le_<mode>): New.
2521 (*vsx_xxpermdi8_le_V8HI): New.
2522 (*vsx_xxpermdi16_le_V16QI): New.
2523 (*vsx_lxvd2x2_le_<mode>): New.
2524 (*vsx_lxvd2x4_le_<mode>): New.
2525 (*vsx_lxvd2x8_le_V8HI): New.
2526 (*vsx_lxvd2x16_le_V16QI): New.
2527 (*vsx_stxvd2x2_le_<mode>): New.
2528 (*vsx_stxvd2x4_le_<mode>): New.
2529 (*vsx_stxvd2x8_le_V8HI): New.
2530 (*vsx_stxvd2x16_le_V16QI): New.
2531
9520e1eb
RL		 25322013-10-07 Renlin Li <Renlin.Li@arm.com>
2533
2534 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2535
36c0bd4f
AK		 25362013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2537
2538 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2539 loop to work also for 31bit ABI.
2540 Save the stack pointer for frame_size > 0.
2541
ee163e72
AK		 25422013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2543
2544 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2545 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2546 constraint letters from expanders.
2547 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2548 retry count to general_operand.
2549 ("tabort"): Give operand 0 a mode.
2550 ("tabort_1"): Add mode and constraint letter for operand 0.
2551 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2552
45d99234
JL		 25532013-10-04 Jeff Law <law@redhat.com>
2554
2555 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2556
adcfd489
UB		 2557 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2558 out of ...
45d99234
JL		 2559 (thread_across_edge): Here. Call it.
2560
7c327f7b
CC		 25612013-10-04 Cary Coutant <ccoutant@google.com>
2562
2563 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2564 discarding a location list expression (or a piece of one).
7c327f7b 2565
6867e128
JH		 25662013-10-03 Jan Hubicka <jh@suse.cz>
2567
65efa7e7
UB		 2568 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2569 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2570 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2571
c7f36d55
JH		 25722013-10-03 Jan Hubicka <jh@suse.cz>
2573
65efa7e7 2574 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH		 2575 accumulate-outgoing-args when producing unwind info.
2576
4f0bee4c
WM		 25772013-10-03 Wei Mi <wmi@google.com>
2578
2579 * lra-constraints.c (insert_move_for_subreg): New function
2580 extracted from simplify_operand_subreg.
2581 (simplify_operand_subreg): Add reload for paradoxical subreg.
2582
111c3f39
RX		 25832013-10-03 Rong Xu <xur@google.com>
2584
65efa7e7
UB		 2585 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2586 the candidate of builtin_expect such that we should fix the
2587 size/time estimation.
2588 (estimate_function_body_sizes): Do the acutally size/time fix-up
2589 for builtin_expect.
111c3f39 2590
942df739
RX		 25912013-10-03 Rong Xu <xur@google.com>
2592
65efa7e7
UB		 2593 * predict.c (tree_predict_by_opcode): Get the probability
2594 for builtin_expect from param builtin_expect_probability.
2595 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2596 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2597 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2598
2284b034
MG		 25992013-10-03 Marc Glisse <marc.glisse@inria.fr>
2600
2601 PR c++/19476
2602 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2603 * calls.c (alloca_call_p): Use get_callee_fndecl.
2604 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2605 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2606 Likewise.
2607 (vrp_visit_stmt): Remove duplicated code.
2608
0609bdf2
MM		 26092013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2610
2611 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2612 ceildf2, btruncdf2, instead of vsx_* name.
2613
2614 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2615 iterators to only do V2DF and V4SF here. Move the DF code to
2616 rs6000.md where it is combined with SF mode. Replace <VSv> with
2617 just 'v' since only vector operations are handled with these insns
2618 after moving the DF support to rs6000.md.
2619 (vsx_sub<mode>3): Likewise.
2620 (vsx_mul<mode>3): Likewise.
2621 (vsx_div<mode>3): Likewise.
2622 (vsx_fre<mode>2): Likewise.
2623 (vsx_neg<mode>2): Likewise.
2624 (vsx_abs<mode>2): Likewise.
2625 (vsx_nabs<mode>2): Likewise.
2626 (vsx_smax<mode>3): Likewise.
2627 (vsx_smin<mode>3): Likewise.
2628 (vsx_sqrt<mode>2): Likewise.
2629 (vsx_rsqrte<mode>2): Likewise.
2630 (vsx_fms<mode>4): Likewise.
2631 (vsx_nfma<mode>4): Likewise.
2632 (vsx_copysign<mode>3): Likewise.
2633 (vsx_btrunc<mode>2): Likewise.
2634 (vsx_floor<mode>2): Likewise.
2635 (vsx_ceil<mode>2): Likewise.
2636 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2637 (vsx_sminsf3): Likewise.
2638 (vsx_fmadf4): Likewise.
2639 (vsx_fmsdf4): Likewise.
2640 (vsx_nfmadf4): Likewise.
2641 (vsx_nfmsdf4): Likewise.
2642 (vsx_cmpdf_internal1): Likewise.
2643
2644 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2645 simpler to select whether a target has SPE or traditional floating
2646 point support in iterators.
2647 (TARGET_DF_SPE): Likewise.
2648 (TARGET_SF_FPR): Likewise.
2649 (TARGET_DF_FPR): Likewise.
2650 (TARGET_SF_INSN): Macros to say whether floating point support
2651 exists for a given operation for expanders.
2652 (TARGET_DF_INSN): Likewise.
2653
2654 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2655 combining of SF/DF mode operations, using both traditional and VSX
2656 registers.
2657 (Fvsx): Likewise.
2658 (Ff): Likewise.
2659 (Fv): Likewise.
2660 (Fs): Likewise.
2661 (Ffre): Likewise.
2662 (FFRE): Likewise.
2663 (abs<mode>2): Combine SF/DF modes using traditional floating point
2664 instructions. Add support for using the upper DF registers with
2665 VSX support, and SF registers with power8-vector support. Update
2666 expanders for operations supported by both the SPE and traditional
2667 floating point units.
2668 (abs<mode>2_fpr): Likewise.
2669 (nabs<mode>2): Likewise.
2670 (nabs<mode>2_fpr): Likewise.
2671 (neg<mode>2): Likewise.
2672 (neg<mode>2_fpr): Likewise.
2673 (add<mode>3): Likewise.
2674 (add<mode>3_fpr): Likewise.
2675 (sub<mode>3): Likewise.
2676 (sub<mode>3_fpr): Likewise.
2677 (mul<mode>3): Likewise.
2678 (mul<mode>3_fpr): Likewise.
2679 (div<mode>3): Likewise.
2680 (div<mode>3_fpr): Likewise.
2681 (sqrt<mode>3): Likewise.
2682 (sqrt<mode>3_fpr): Likewise.
2683 (fre<Fs>): Likewise.
2684 (rsqrt<mode>2): Likewise.
2685 (cmp<mode>_fpr): Likewise.
2686 (smax<mode>3): Likewise.
2687 (smin<mode>3): Likewise.
2688 (smax<mode>3_vsx): Likewise.
2689 (smin<mode>3_vsx): Likewise.
2690 (negsf2): Delete SF operations that are merged with DF.
2691 (abssf2): Likewise.
2692 (addsf3): Likewise.
2693 (subsf3): Likewise.
2694 (mulsf3): Likewise.
2695 (divsf3): Likewise.
2696 (fres): Likewise.
2697 (fmasf4_fpr): Likewise.
2698 (fmssf4_fpr): Likewise.
2699 (nfmasf4_fpr): Likewise.
2700 (nfmssf4_fpr): Likewise.
2701 (sqrtsf2): Likewise.
2702 (rsqrtsf_internal1): Likewise.
2703 (smaxsf3): Likewise.
2704 (sminsf3): Likewise.
2705 (cmpsf_internal1): Likewise.
2706 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2707 (negdf2): Delete DF operations that are merged with SF.
2708 (absdf2): Likewise.
2709 (nabsdf2): Likewise.
2710 (adddf3): Likewise.
2711 (subdf3): Likewise.
2712 (muldf3): Likewise.
2713 (divdf3): Likewise.
2714 (fred): Likewise.
2715 (rsqrtdf_internal1): Likewise.
2716 (fmadf4_fpr): Likewise.
2717 (fmsdf4_fpr): Likewise.
2718 (nfmadf4_fpr): Likewise.
2719 (nfmsdf4_fpr): Likewise.
2720 (sqrtdf2): Likewise.
2721 (smaxdf3): Likewise.
2722 (smindf3): Likewise.
2723 (cmpdf_internal1): Likewise.
2724 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2725 (btrunc<mode>2): Delete separate expander, and combine with the
2726 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2727 (btrunc<mode>2_fpr): Likewise.
2728 (ceil<mode>2): Likewise.
2729 (ceil<mode>2_fpr): Likewise.
2730 (floor<mode>2): Likewise.
2731 (floor<mode>2_fpr): Likewise.
2732 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2733 Add support for using the upper registers with VSX and
2734 power8-vector. Move insns to be closer to the define_expands. On
2735 VSX systems, prefer the traditional form of FMA over the VSX
2736 version, since the traditional form allows the target not to
2737 overlap with the inputs.
2738 (fms<mode>4_fpr): Likewise.
2739 (nfma<mode>4_fpr): Likewise.
2740 (nfms<mode>4_fpr): Likewise.
2741
5bea0c6c 27422013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 2743 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT		 2744
2745 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2746 (struct mult_cost_table): Likewise.
2747 (struct mem_cost_table): Likewise.
2748 (struct fp_cost_table): Likewise.
2749 (struct vector_cost_table): Likewise.
2750 (cpu_cost_table): Likewise.
2751 * config/arm/arm.opt (mold-rts-costs): New option.
2752 (mnew-generic-costs): Likewise.
2753 * config/arm/arm.c (generic_extra_costs): New table.
2754 (cortexa15_extra_costs): Likewise.
2755 (arm_slowmul_tune): Use NULL as new costs.
2756 (arm_fastmul_tune): Likewise.
2757 (arm_strongarm_tune): Likewise.
2758 (arm_xscale_tune): Likewise.
2759 (arm_9e_tune): Likewise.
2760 (arm_v6t2_tune): Likewise.
2761 (arm_cortex_a5_tune): Likewise.
2762 (arm_cortex_a9_tune): Likewise.
2763 (arm_v6m_tune): Likewise.
2764 (arm_fa726te_tune): Likewise.
2765 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2766 (arm_cortex_tune): Use generict_extra_costs.
2767 (shifter_op_p): New function.
2768 (arm_unspec_cost): Likewise.
2769 (LIBCALL_COST): Define.
2770 (arm_new_rtx_costs): New function.
2771 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2772 table is available. Use old costs otherwise unless mnew-generic-costs
2773 is specified.
2774 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2775 (cpu_cost_table): Declare.
2776
24c56925
MS		 27772013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2778
2779 PR target/58460
2780 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2781 (*subs_mul_imm_<mode>)
2782 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2783 (*sub_<shift>_<mode>)
2784 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2785 Remove k constraint.
2786
2ab8f063
IB		 27872013-10-03 Ian Bolton <ian.bolton@arm.com>
2788
2789 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2790 code.
2791 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2792
79221839
TJ		 27932013-10-02 Teresa Johnson <tejohnson@google.com>
2794
2795 * predict.c (probably_never_executed): New function.
2796 (probably_never_executed_bb_p): Invoke probably_never_executed.
2797 (probably_never_executed_edge_p): Ditto.
2798 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2799 Treat profile insanities conservatively.
2800
d441afe0
JDA		 28012013-10-02 John David Anglin <danglin@gcc.gnu.org>
2802
2803 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2804
80f466c4
VM		 28052013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2806
2807 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2808 use it. Use smaller increase for scratch. Don't increase reject
2809 for early clobber scratch.
2810 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2811 setting eliminated regs except setting fp from hfp.
2812 (lra_eliminate): Check lra_insn_recog_data on NULL.
2813
6e228b4b
MM		 28142013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2815
2816 PR target/58587
2817 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 2818 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM		 2819 (RS6000_CPU, power7 defaults): Likewise.
2820
11b54a5a
UB		 28212013-10-02 Uros Bizjak <ubizjak@gmail.com>
2822
2823 * config/x-linux (host-linux.o): Remove header dependencies.
2824 Use $(COMPILE) and $(POSTCOMPILE).
2825 * config/t-linux-android (linux-android.o): Ditto.
2826
bbc02b69
UB		 28272013-10-02 Uros Bizjak <ubizjak@gmail.com>
2828
2829 * Makefile.in (expmed.o-warn): Remove.
2830
440917de
AM		 28312013-10-02 Andrew MacLeod <amacleod@redhat.com>
2832
2833 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2834 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2835
99206ca9
TJ		 28362013-10-02 Teresa Johnson <tejohnson@google.com>
2837
2838 * dojump.c (do_jump_1): Divide probability between
2839 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2840
56e82b14
TT		 28412013-10-02 Tom Tromey <tromey@redhat.com>
2842
2843 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2844
3d9c733e
AM		 28452013-10-02 Andrew MacLeod <amacleod@redhat.com>
2846
2847 * tree-flow.h: Remove some prototypes.
2848 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2849 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2850 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2851 mark_virtual_phi_result_for_renaming): Relocate here.
2852 * tree-into-ssa.h: Add prototypes.
bbc02b69 2853 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM		 2854 single_pred_before_succ_order.
2855 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2856 (nonfreeing_call_p) Move to gimple.c.
2857 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2858 tree-ssa-phiopt.c.
2859 * basic-block.h (single_pred_before_succ_order): Add prototype.
2860 * gimple.c (nonfreeing_call_p): Relocate here.
2861 * gimple.h: Add prototype.
2862 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2863 * tree-ssa-dom.h: New file. Relocate prototypes here.
2864 * tree-ssa.h: Include tree-ssa-dom.h.
2865
78cedfb1
UB		 28662013-10-02 Uros Bizjak <ubizjak@gmail.com>
2867
2868 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2869 Use $(COMPILE) and $(POSTCOMPILE).
2870
2871 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2872
744730a4
AM		 28732013-10-02 Andrew MacLeod <amacleod@redhat.com>
2874
2875 * tree-flow.h: Remove some prototypes.
2876 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2877 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2878 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2879 propagate_tree_value*): Move from here to...
2880 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2881 propagate_tree_value*): Relocate here.
2882 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2883 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2884 * gimple-fold.c: Remove gimple-fold.h from include list.
2885 * tree-vrp.c: Remove gimple-fold.h from include list.
2886 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2887 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2888 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2889 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2890 * sese.c: Add tree-ssa-propagate.h to include list.
2891
826a536d
RB		 28922013-10-02 Richard Biener <rguenther@suse.de>
2893
2894 * tree-loop-distribution.c: Include tree-vectorizer.h for
2895 find_loop_location.
2896 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 2897 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB		 2898 (partition_alloc): Adjust.
2899 (partition_builtin_p): Likewise.
2900 (partition_has_writes): Remove.
2901 (partition_reduction_p): New function.
2902 (partition_merge_into): Likewise.
2903 (generate_code_for_partition): Commonize builtin partition
2904 handling tail.
2905 (rdg_cannot_recompute_vertex_p): Remove.
2906 (already_processed_vertex_p): Likewise.
2907 (rdg_flag_vertex): Do not set has_writes.
2908 (classify_partition): Adjust.
2909 (rdg_build_partitions): Do not set has_writes, treat all
2910 partitions as useful.
78cedfb1 2911 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB		 2912 (tree_loop_distribution): Report number of loops and library
2913 calls generated as opt-info.
2914
4b403ece
AM		 29152013-10-02 Andrew MacLeod <amacleod@redhat.com>
2916
2917 * tree-flow.h: Include new .h files. Move prototypes.
2918 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2919 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2920 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2921 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2922 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2923 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2924 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2925 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 2926 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM		 2927 tree-dfa.h.
2928 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2929 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2930 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2931 * tree-scalar-evolution.c: Include tree.h.
2932 * sese.c: Include tree.h.
2933 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2934 * dwarf2out.c: Include tree-dfa.h.
2935 * tree-chrec.c: Include tree.h.
2936 * tree-data-ref.c: Include tree.h.
2937
1d2151c6
YZ		 29382013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2939
2940 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2941 Fix whitespace.
2942
441ad147
RO		 29432013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2944
2945 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2946 Use $(COMPILE) and $(POSTCOMPILE).
2947 (sol2-cxx.o): Likewise.
2948 (sol2-stubs.o): Likewise.
2949 (sol2.o): Likewise.
2950 * config/x-solaris (host-solaris.o): Likewise.
2951
2952 * config/sparc/t-sparc (sparc.o): Remove.
2953 (sparc-c.o): Remove header dependencies.
2954 Use $(COMPILE) and $(POSTCOMPILE).
2955 * config/sparc/x-sparc: Likewise.
2956
2deaf8b0
JR		 29572013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2958
2959 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2960 * config/arc/arc700.md: Likewise.
2961 * config/arc/arc-modes.def: Likewise.
2962 * config/arc/arc-simd.h: Likewise.
2963 * config/arc/t-arc-uClibc: Likewise.
2964 * config/arc/t-arc-newlib: Likewise.
2965
f6fe771a
RL		 29662013-10-02 Renlin Li <renlin.li@arm.com>
2967
78cedfb1
UB		 2968 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2969 plus_constant.
f6fe771a
RL		 2970 (aarch64_expand_epilogue): Likewise.
2971
0916f876
YZ		 29722013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2973 Yufeng Zhang <yufeng.zhang@arm.com>
2974
2975 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2976 declaration.
2977 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2978 'base_in' represent a conversion and legal_cast_p_1 holds; set
2979 'base_in' with the returned value from get_unwidened.
2980
ebfcd719
KT		 29812013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2982
2983 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2984 plus_constant is not used.
2985
157ca3e9
WM		 29862013-10-01 Wei Mi <wmi@google.com>
2987
78cedfb1 2988 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM		 2989 * config/i386/i386.md: Add define_peephole2 to
2990 break partial reg stall for cvtss2sd/cvtsd2ss.
2991
cd4dd8f0
JR		 29922013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2993
2994 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 2995 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 2996
1465e5cf
JL		 29972013-10-01 Jeff Law <law@redhat.com>
2998
2999 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3000 outgoing_edge and intermediate_edge fields. Instead store the path.
3001 (redirection_data::hash): Hash on the last edge's destination index.
3002 (redirection_data::equal): Check the entire thread path.
3003 (lookup_redirectio_data): Corresponding changes.
3004 (create_edge_and_update_destination_phis): Likewise.
3005 (thread_single_edge): Likewise.
3006
25c606cb 30072013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 3008 Diego Novillo <dnovillo@google.com>
25c606cb
JR		 3009
3010 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3011 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3012 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3013 versions of these patterns.
3014
0ccbc132
JR		 3015 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3016 (__builtin_arc_aligned): Likewise.
3017
f55d4a20
JR		 3018 * config/arc/arc.md: Expand adc_0 comment stating the intended
3019 purpose and why it isn't ready.
3020 Replace commented out call_value_via_label_mixed with a
3021 plain comment about bl_s.
3022
5719867d 3023 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR		 3024 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3025 Remove if (1) condition.
3026 (arc_encode_section_info): Fix comment.
3027
6462fab0
JR		 30282013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3029
3030 * config/arc/arc.c (arc_conditional_register_usage):
3031 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3032 Also set reg_alloc_order for DMA config regs.
3033
5d5f6720
JR		 30342013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3035 Jeremy Bennett <jeremy.bennett@embecosm.com>
3036
3037 * doc/install.texi (--with-cpu): Mention ARC.
3038 (arc-*-elf32): New paragraph.
3039 (arc-linux-uclibc): Likewise.
3040 * doc/md.texi (Machine Constraints): Add ARC part.
3041 * doc/invoke.texi: (menu): Add ARC Options.
3042 (Machine Dependent Options) <ARC Options>: Add synopsis.
3043 (node ARC Options): Add.
3044 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3045 (ARC Built-in Functions): New section defining
3046 generic ARC built-in functions.
3047 (ARC SIMD Built-in Functions): New section defining SIMD specific
3048 built-in functions.
3049 (Declaring Attributes of Functions): Extended
3050 description of short_call and long_call attributes for ARC and
3051 added index entries.
3052
526b7aee
SV		 30532013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3054 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3055 Joern Rennecke <joern.rennecke@embecosm.com>
3056 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3057 Brendan Kehoe <brendan@zen.org>
3058 Michael Eager <eager@eagercon.com>
3059 Simon Cook <simon.cook@embecosm.com>
3060 Jeremy Bennett <jeremy.bennett@embecosm.com>
3061
3062 * config/arc, common/config/arc: New directories.
3063
53426f6c
JR		 30642013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3065 Brendan Kehoe <brendan@zen.org>
3066 Simon Cook <simon.cook@embecosm.com>
3067
3068 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3069
1df9f5a9
AM		 30702013-10-01 Andrew MacLeod <amacleod@redhat.com>
3071
3072 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3073 * tree-ssa-coalesce.h: New. Move prototype to here.
3074 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3075 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3076 (gimple_can_coalesce_p): Move to...
3077 * gimple.c (gimple_can_coalesce_p): Here.
3078
80560f95
AM		 30792013-10-01 Andrew MacLeod <amacleod@redhat.com>
3080
3081 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3082 (dump_decl_set): Move to gimple.c.
3083 * gimple.h: Don't include tree-ssa-operands.h.
3084 (dump_decl_set): Add prototype.
3085 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3086 Move to gimple-ssa.h.
3087 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3088 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3089 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3090 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3091 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3092 than PHI_ARG_DEF.
3093 (dump_decl_set): Relocate here.
3094 * gimple-ssa.h: New file.
3095 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3096 Relocate from gimple.h.
3097 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3098 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3099 swap_tree_operands and remove non-ssa path.
3100 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3101 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3102 swap_ssa_operands.
3103 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3104 vect_is_simple_reduction_1): Use swap_ssa_operands.
3105 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3106 (enum need_phi_state): Move to tree-into-ssa.c.
3107 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3108 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3109 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3110 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3111 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3112 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3113 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3114 num_imm_uses): Move to ssa-iterators.h.
3115 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7ab4168e 3116 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
80560f95
AM		 3117 tree-phinodes.h.
3118 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3119 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3120 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3121 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3122 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3123 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3124 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3125 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3126 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3127 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3128 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3129 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3130 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3131 (set_phi_nodes): Move to tree-phinodes.h.
3132 * tree-ssa-operands.h (enum ssa_op_iter_type,
3133 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3134 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3135 (dump_decl_set): Remove prototype.
3136 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3137 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3138 (set_phi_nodes): Relocate from tree-flow-inline.h.
3139 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3140 tree-flow-inline.h
3141 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3142 include list. Temporarily add gimple.h to include list.
3143 * ssa-iterators.h: New file.
3144 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3145 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3146 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3147 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3148 Relocate from tree-ssa-operands.h.
3149 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3150 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3151 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3152 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3153 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3154 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3155 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3156 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3157 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3158 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3159 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3160 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3161 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3162 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3163 Relocate from tree-flow-inline.h.
3164 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3165
8da00d65
VP		 31662013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3167
3168 * aarch64-simd.md
3169 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3170 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3171 Insert '\t' to output template.
3172 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3173 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3174 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3175 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3176
e2ebe1c2
UB		 31772013-10-01 Uros Bizjak <ubizjak@gmail.com>
3178
3179 * doc/install.texi (Host/target specific installation notes for GCC):
3180 Put @anchor before @heading.
3181 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3182 Use @email for email addresses.
3183
aee2d611
JL		 31842013-10-01 Jeff Law <law@redhat.com>
3185
3186 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3187 a vec. Only delete the path if we create one without successfully
3188 registering a jump thread.
3189 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3190 as a pointer.
3191 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3192 (paths): New vector of jump threading paths.
3193 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3194 (THREAD_PATH): New accessor macro for the entire thread path.
3195 (lookup_redirection_data): Get intermediate and final outgoing edge
3196 from the thread path.
3197 (create_edge_and_update_destination_phis): Copy the threading path.
3198 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3199 jump threading path.
3200 (ssa_redirect_edges): Get edges and block types from the jump threading
3201 path. Free the path vector.
3202 (thread_block): Get edges from the jump threading path. Look at the
3203 entire path to see if we thread to a loop exit. If we cancel a jump
3204 thread request, then free the path vector.
3205 (thread_single_edge): Get edges and block types from the jump threading
3206 path. Free the path vector.
3207 (thread_through_loop_header): Get edges and block types from the jump
3208 threading path. Free the path vector.
3209 (mark_threaded_blocks): Iterate over the vector of paths and store
3210 the path on the appropriate edge. Get edges and block types from the
3211 jump threading path.
3212 (mark_threaded_blocks): Get edges and block types from the jump
3213 threading path. Free the path vector.
3214 (thread_through_all_blocks): Use the vector of paths rather than
3215 a vector of 3-edge sets.
3216 (register_jump_thread): Accept pointer to a path vector rather
3217 than the path vector itself. Store the path vector for later use.
3218 Simplify.
3219
966f97ac 32202013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 3221 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ		 3222
3223 PR target/58574
3224 * config/s390/s390.c (s390_split_branches): Modify check for table
3225 jump insns.
3226 (s390_chunkify_start): Rearrange table jump insn check in order to
3227 deal with compare and branch insns correctly.
3228
3a323a38
KV		 32292013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3230
3231 PR target/58578
3232 Revert
3233 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3234 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3235 define_insn_and_split.
3236 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3237 (shiftsi3_compare): New pattern.
3238 (rrx): New pattern.
3239 * config/arm/unspecs.md (UNSPEC_RRX): New.
3240
90444831
AM		 32412013-10-01 Alan Modra <amodra@gmail.com>
3242
3243 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3244 casing inout operands.
3245
818625cf
RB		 32462013-10-01 Richard Biener <rguenther@suse.de>
3247
3248 PR tree-optimization/58553
3249 * tree-loop-distribution.c (struct partition_s): Add niter member.
3250 (classify_partition): Populate niter member for the partition
3251 and properly identify whether the relevant store happens before
3252 or after the loop exit.
3253 (generate_memset_builtin): Use niter member from the partition.
3254 (generate_memcpy_builtin): Likewise.
3255
30f641cd
RS		 32562013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3257
3258 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3259 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3260
65d3284b
RS		 32612013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3262
3263 * basic-block.h (edge_list): Prefix member names with "m_".
3264 * context.h (context): Likewise.
3265 * domwalk.h (dom_walker): Likewise.
3266 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3267 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3268 * hash-table.h (hash_table): Likewise.
3269 * machmode.h (bit_field_mode_iterator): Likewise.
3270 * pass_manager.h (pass_list): Likewise.
3271 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3272 * tree-pass.h (pass_data): Likewise.
3273 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3274 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3275 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3276 * asan.c (pass_data_asan): Update accordingly.
3277 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3278 (control_dependences::control_dependences): Likewise.
3279 (control_dependences::~control_dependences): Likewise.
3280 (control_dependences::~control_dependences): Likewise.
3281 (control_dependences::get_edges_dependent_on): Likewise.
3282 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3283 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3284 * context.c (gcc::context::context): Likewise.
3285 * cprop.c (pass_rtl_cprop::clone): Likewise.
3286 * domwalk.c (dom_walker::walk): Likewise.
3287 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3288 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3289 * mode-switching.c (pass_mode_switching::clone): Likewise.
3290 * passes.c (opt_pass::opt_pass): Likewise.
3291 (pass_manager::pass_manager): Likewise.
3292 * predict.c (pass_strip_predict_hints::clone): Likewise.
3293 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3294 (pass_split_all_insns::clone): Likewise.
3295 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3296 Likewise.
3297 (bit_field_mode_iterator::next_mode): Likewise.
3298 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3299 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3300 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3301 * tree-complex.c (pass_lower_complex::clone): Likewise.
3302 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3303 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3304 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3305 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3306 (pass_fold_builtins::clone): Likewise.
3307 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3308 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3309 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3310 (pass_cd_dce::clone): Likewise.
3311 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3312 (pass_phi_only_cprop::clone): Likewise.
3313 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3314 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3315 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3316 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3317 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3318 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3319 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3320 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3321 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3322 * tree-vrp.c (pass_vrp::clone): Likewise.
3323 * tsan.c (pass_tsan::clone): Likewise.
3324
f66d0891
JJ		 33252013-09-30 Jakub Jelinek <jakub@redhat.com>
3326
d2a365a8
JJ		 3327 PR middle-end/58564
3328 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3329 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3330
f66d0891
JJ		 3331 PR middle-end/58564
3332 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3333 optimization, punt if sign_bit_p looked through any zero extension.
3334
05357ac3
TJ		 33352013-09-30 Teresa Johnson <tejohnson@google.com>
3336
3337 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3338 Update redirected out edge count in joiner case.
3339 (ssa_redirect_edges): Common the joiner and non-joiner cases
3340 so that joiner case gets profile updates.
3341
ca406576
RB		 33422013-09-30 Richard Biener <rguenther@suse.de>
3343
3344 PR tree-optimization/58554
e2ebe1c2
UB		 3345 * tree-loop-distribution.c (classify_partition): Require
3346 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB		 3347 (tree_loop_distribution): Calculate dominance info.
3348
92d649c4
VK		 33492013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3350
3351 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3352 (NO_PROFILE_COUNTERS): Likewise.
3353 (PROFILE_HOOK): Likewise.
3354 (FUNCTION_PROFILER): Likewise.
3355 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3356
bd9534e2
IS		 33572013-09-30 Iain Sandoe <iain@codesourcery.com>
3358
3359 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3360 calls and defines in TARGET_MACHO conditional.
3361 (load_macho_picbase_di): Likewise.
3362 (reload_macho_picbase): Likewise.
3363 (reload_macho_picbase_si): Likewise.
3364 (reload_macho_picbase_di): Likewise.
3365 (nonlocal_goto_receiver): Likewise.
3366
4494fbc9
NC		 33672013-09-30 Nick Clifton <nickc@redhat.com>
3368
3369 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3370 that use the MSP430X ISA.
3371 (msp430_option_override): Scan -mmcu command line option for any
3372 MCU name that supports the MSP430X ISA.
3373 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3374 -mmcu options which enable the MSP430X ISA.
3375
791d9044
RB		 33762013-09-30 Richard Biener <rguenther@suse.de>
3377
3378 PR middle-end/58532
3379 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3380 before looking for setjmp-like calls.
3381
e1c5c877
IS		 33822013-09-29 Iain Sandoe <iain@codesourcery.com>
3383
3384 PR target/10901
3385 * config/darwin-protos.h (machopic_get_function_picbase): New.
3386 * config/darwin.c (machopic_get_function_picbase): New.
3387 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3388 label for a new func. (load_macho_picbase_di): Likewise.
3389 (reload_macho_picbase): New expand.
3390 (reload_macho_picbase_si): New insn.
3391 (reload_macho_picbase_di): New insn.
3392 (nonlocal_goto_receiver): New define and split.
3393 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3394 (unspecv enum): Add UNSPECV_NLGR.
3395
749278c5
IS		 33962013-09-29 Iain Sandoe <iain@codesourcery.com>
3397
3398 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3399 that altivec registers are correctly sized on Darwin.
3400
2c43a51c
IS		 34012013-09-29 Iain Sandoe <iain@codesourcery.com>
3402
3403 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3404 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3405 * config/x-darwin (host-darwin.o): Likewise.
3406 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3407 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3408 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3409
7a5911d3
UB		 34102013-09-29 Uros Bizjak <ubizjak@gmail.com>
3411
3412 * doc/invoke.texi: Fix usage of @tie{} command.
3413
3f67a73d
EB		 34142013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3415
3416 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3417
9ac2f538
JH		 34182013-09-28 Jan Hubicka <jh@suse.cz>
3419
3420 * config/i386/x86-tune.def: Add documentation for each of the options;
3421 add whitespace.
3422
b0ff06da
JH		 34232013-09-28 Jan Hubicka <jh@suse.cz>
3424
7a5911d3
UB		 3425 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3426 generic.
b0ff06da
JH		 3427 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3428 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3429 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3430 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3431 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3432
c3284718
RS		 34332013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3434
3435 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3436 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3437 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3438 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3439 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3440 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3441 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3442 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3443 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3444 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3445 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3446 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3447 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3448 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3449 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3450 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3451 gimple.h, godump.c, graphite-clast-to-gimple.c,
3452 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3453 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3454 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3455 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3456 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3457 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3458 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3459 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3460 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3461 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3462 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3463 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3464 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3465 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3466 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3467 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3468 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3469 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3470 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3471 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3472 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3473 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3474 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3475 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3476 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3477 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3478 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3479 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3480 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3481 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3482 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3483 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3484 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3485 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3486 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3487 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3488 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3489 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3490 whitespace before "(".
3491
c6285bd7
SL		 34922013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3493
3494 * expr.h (extract_bit_field): Remove packedp parameter.
3495 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3496 from forward declaration.
3497 (store_split_bit_field): Remove packedp arg from calls to
3498 extract_fixed_bit_field.
3499 (extract_bit_field_1): Remove packedp parameter and packedp
3500 argument from recursive calls and calls to extract_fixed_bit_field.
3501 (extract_bit_field): Remove packedp parameter and corresponding
3502 arg to extract_bit_field_1.
3503 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3504 to issue warnings.
3505 (extract_split_bit_field): Remove packedp arg from call to
3506 extract_fixed_bit_field.
3507 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3508 (copy_blkmode_from_reg): Likewise.
3509 (copy_blkmode_to_reg): Likewise.
3510 (read_complex_part): Likewise.
3511 (store_field): Likewise.
3512 (expand_expr_real_1): Likewise.
3513 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3514 to extract_bit_field.
3515 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3516 call to extract_bit_field.
3517 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3518 call to extract_bit_field.
3519 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3520 and special packedp behavior from -fstrict-volatile-bitfields
3521 documentation.
3522
522d4efc
JBG		 35232013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3524
3525 * lra-eliminations.c (init_elim_table): Guard value_p.
3526
19cdb489
MM		 35272013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3528
3529 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3530 DFmode, DImode, and SFmode in the upper VSX registers based on the
3531 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3532 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3533 of the VSX register handling.
3534
3535 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3536 (f32_sv): Likewise.
3537 (zero_extendsidi2_lfiwzx): Add support for loading into the
3538 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3539 only do VSX memory options on Altivec registers.
3540 (extendsidi2_lfiwax): Likewise.
3541 (extendsfdf2_fpr): Likewise.
3542 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3543 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3544 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3545 (movdi_internal64): Likewise.
3546
d6d11272
XDL		 35472013-09-27 Xinliang David Li <davidxl@google.com>
3548
3549 * opts.c (finish_options): Adjust parameters
3550 according to vect cost model.
3551 (common_handle_option): Set dynamic vect cost
3552 model for FDO.
3553 targhooks.c (default_add_stmt_cost): Compute stmt cost
3554 unconditionally.
3555 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3556 Use helper function.
3557 * tree-vectorizer.h (unlimited_cost_model): New function.
3558 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3559 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3560 function.
3561 (vect_enhance_data_refs_alignment): Ditto.
3562 * flag-types.h: New enum.
3563 * common/config/i386/i386-common.c (ix86_option_init_struct):
3564 No need to initialize vect_cost_model flag.
3565 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3566 unconditionally.
3567
ac1857a3
DN		 35682013-09-27 Diego Novillo <dnovillo@google.com>
3569
3570 * gimple.h (enum ssa_mode): Remove.
3571
0d0e2af6
PM		 35722013-09-27 Paulo Matos <pmatos@broadcom.com>
3573
3574 * cfgloop.h (number_of_loops): Fix typo in check for null.
3575
09dc585e
JJ		 35762013-09-27 Jakub Jelinek <jakub@redhat.com>
3577
3578 PR middle-end/58551
3579 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3580 are children of outermost saved_cfun's loop, and set it up to
3581 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3582 if loop != loop0 and SESE region contains bbs that belong to loop0.
3583
ec5a3504
RS		 35842013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3585
3586 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3587 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3588 (set_address_segment, set_address_base, set_address_index)
3589 (set_address_disp): Accept the argument unconditionally.
3590 (baseness): Remove must_be_base_p and must_be_index_p checks.
3591 (decompose_normal_address): Classify as much as possible in the
3592 main loop.
3593
f91aec98
RS		 35942013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3595
3596 * cse.c (count_reg_usage): Handle INT_LIST.
3597 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3598 * reginfo.c (reg_scan_mark_refs): Likewise.
3599 * reload1.c (eliminate_regs_1): Likewise.
3600
a9195970
IS		 36012013-09-27 Iain Sandoe <iain@codesourcery.com>
3602
3603 PR middle-end/58547
3604 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3605 signed.
3606
e6f1c509
RB		 36072013-09-27 Richard Biener <rguenther@suse.de>
3608
3609 PR tree-optimization/58459
3610 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3611 restriction not propagating into loops.
3612
84f48495
FW		 36132013-09-26 Florian Weimer <fw@deneb.enyo.de>
3614
3615 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3616 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3617 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3618
74fc8b8a
RB		 36192013-09-26 Richard Biener <rguenther@suse.de>
3620
7a5911d3 3621 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3622
b4ada065
RB		 36232013-09-26 Richard Biener <rguenther@suse.de>
3624
3625 * alias.h (component_uses_parent_alias_set): Rename to ...
3626 (component_uses_parent_alias_set_from): ... this.
3627 * alias.c (component_uses_parent_alias_set): Rename to ...
3628 (component_uses_parent_alias_set_from): ... this and return
3629 the desired parent.
3630 (reference_alias_ptr_type_1): Use the result from
3631 component_uses_parent_alias_set_from instead of stripping
3632 components one at a time.
3633 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3634
78bca40d
AM		 36352013-09-26 Andrew MacLeod <amacleod@redhat.com>
3636
7a5911d3
UB		 3637 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3638 Move prototypes to...
78bca40d
AM		 3639 * tree-ssa-ter.h: New File. Move prototypes here.
3640 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3641 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3642 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3643 from is_replaceable_p.
78bca40d
AM		 3644 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3645 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3646 (process_replaceable): Use ter_is_replaceable_p.
3647 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3648 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3649 newly refactored ssa_is_replaceable_p.
3650 * cfgexpand.c: Include tree-outof-ssa.h.
3651 * ssaexpand.h: Delete.
3652
ff2a63a7
AM		 36532013-09-26 Andrew MacLeod <amacleod@redhat.com>
3654
3655 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3656 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3657 tree-ssa.c
3658 (create_gimple_tmp): Delete.
3659 (get_expr_type, build_assign, build_type_cast): Move to...
3660 * gimple-builder.c: New File.
3661 (get_expr_type): Relocate from gimple.c.
3662 (build_assign, build_type_cast): Change to only create ssanames.
3663 * gimple.h: Move prototypes to...
3664 * gimple-builder.h: New File. Here.
3665 * tree-ssa.h: And here.
3666 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3667 count_uses_and_derefs): Relocate from gimple.c.
3668 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3669 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3670 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3671 gimple_replace_ssa_lhs.
3672 * asan.c: Include gimple-builder.h.
3673 * Makefile.in: Add gimple-builder.o.
3674
a2544177
RB		 36752013-09-26 Richard Biener <rguenther@suse.de>
3676
3677 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3678 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3679 (loe_visit_block): Use gcc_checking_assert.
3680 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3681 gimple_assign_ssa_name_copy_p.
3682 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3683 change.
3684
6d1ca9a0
DE		 36852013-09-26 David Edelsohn <dje.gcc@gmail.com>
3686
3687 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3688 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3689
46e950db
RB		 36902013-09-26 Richard Biener <rguenther@suse.de>
3691
3692 PR tree-optimization/58539
3693 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3694 the fact that debug statements are not taking part in loop-closed
3695 SSA construction.
3696
40ada30a
NC		 36972013-09-26 Nick Clifton <nickc@redhat.com>
3698
3699 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3700 time warning message.
3701 (msp430_print_operand_raw): Delete unused letter parameter.
3702 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3703 (msp430_print_operand_address): New function.
3704 (msp430_print_operand): Move address printing code from here to
3705 new function.
3706 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3707 assembler.
3708 (zero_extendpsisi2): Likewise.
3709 (extendpsisi2): New pattern.
3710 (andneghi3): New pattern.
3711
38e8f663
YR		 37122013-09-26 Yvan Roux <yvan.roux@linaro.org>
3713
3714 * config/aarch64/aarch64.opt (mlra): New option.
3715 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3716 (TARGET_LRA_P): Define.
3717
4167e8d4
EB		 37182013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3719
3720 * expr.c (expand_assignment): Remove obsolete comment.
3721
5254eac4
JL		 37222013-09-25 Jeff Law <law@redhat.com>
3723
3724 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3725 tree-ssa-threadupdate.h.
3726 (register_jump_thread): Similarly.
3727 * tree-ssa-threadupdate.h: New header file.
3728 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3729 * tree-vrp.c: Likewise.
3730 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3731 (thread_around_empty_blocks): Change type of path vector argument to
3732 an edge,type pair from just an edge. Initialize both elements when
3733 appending to a jump threading path. Tweak references to elements
3734 appropriately.
3735 (thread_across_edge): Similarly. Release memory for the elements
3736 as needed.
3737 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3738 (dump_jump_thread_path): New function broken out from
3739 register_jump_thread.
3740 (register_jump_thread): Use dump_jump_thread_path. Change type of
3741 path vector entries. Search the path for NULL edges and dump
3742 the path if one is found. Tweak the conversion of path to 3-edge
3743 form to use the block copy type information embedded in the path.
3744
20f114a3
YR		 37452013-09-25 Yvan Roux <yvan.roux@linaro.org>
3746
3747 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3748
3936bafc
YR		 37492013-09-25 Yvan Roux <yvan.roux@linaro.org>
3750 Vladimir Makarov <vmakarov@redhat.com>
3751
3752 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3753 from the least significant bit.
3754 (strip_address_mutations): Add bitfield operations handling.
3755 (must_be_index_p): Add shifting and rotate operations handling.
3756 (set_address_base): Use must_be_base_p predicate.
4167e8d4 3757 (set_address_index): Use must_be_index_p predicate.
3936bafc 3758
b86f6e9e
AI		 37592013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3760 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3761 Sergey Lega <sergey.s.lega@intel.com>
3762 Anna Tikhonova <anna.tikhonova@intel.com>
3763 Ilya Tocar <ilya.tocar@intel.com>
3764 Andrey Turetskiy <andrey.turetskiy@intel.com>
3765 Ilya Verbin <ilya.verbin@intel.com>
3766 Kirill Yukhin <kirill.yukhin@intel.com>
3767 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3768
3769 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3770 Use new names.
3771 (ix86_expand_vector_move_misalign): Support new unaligned load and
3772 stores and use new names.
3773 (CODE_FOR_sse2_storedqu): Rename to ...
3774 (CODE_FOR_sse2_storedquv16qi): ... this.
3775 (CODE_FOR_sse2_loaddqu): Rename to ...
3776 (CODE_FOR_sse2_loaddquv16qi): ... this.
3777 (CODE_FOR_avx_loaddqu256): Rename to ...
3778 (CODE_FOR_avx_loaddquv32qi): ... this.
3779 (CODE_FOR_avx_storedqu256): Rename to ...
3780 (CODE_FOR_avx_storedquv32qi): ... this.
3781 * config/i386/i386.md (fpint_logic): New.
3782 * config/i386/sse.md (VMOVE): Extend for AVX512.
3783 (VF): Ditto.
3784 (VF_128_256): New.
3785 (VF_512): Ditto.
3786 (VI_UNALIGNED_LOADSTORE): Ditto.
3787 (sse2_avx_avx512f): Ditto.
3788 (sse2_avx2): Extend for AVX512.
3789 (sse4_1_avx2): Ditto.
3790 (avx2_avx512f): New.
3791 (sse): Extend for AVX512.
3792 (sse2): Ditto.
3793 (sse4_1): Ditto.
3794 (avxsizesuffix): Ditto.
3795 (sseintvecmode): Ditto.
3796 (ssePSmode): Ditto.
3797 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3798 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3799 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3800 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3801 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3802 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3803 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3804 (STORENT_MODE): Extend for AVX512.
3805 (*absneg<mode>2): Replace constraint "x" with "v".
3806 (*mul<mode>3): Ditto.
3807 (*ieee_smin<mode>3): Ditto.
3808 (*ieee_smax<mode>3): Ditto.
3809 (avx_cmp<mode>3): Replace VF with VF_128_256.
3810 (*<sse>_maskcmp<mode>3_comm): Ditto.
3811 (<sse>_maskcmp<mode>3): Ditto.
3812 (<sse>_andnot<mode>3): Extend for AVX512.
3813 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3814 (<code><mode>3, fpint_logic): New.
3815 (*<code><mode>3): Extend for AVX512.
3816 (avx512flogicsuff): New.
3817 (avx512f_<logic><mode>): Ditto.
3818 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3819 VF_128_256.
3820 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3821 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3822 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3823 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3824 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3825 (xop_vpermil2<mode>3): Ditto.
3826 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3827 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3828 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3829 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3830
4d44d03c
TT		 38312013-09-25 Tom Tromey <tromey@redhat.com>
3832
3833 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3834 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3835 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3836 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3837 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3838 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3839 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3840 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3841 (GRAPHITE_HTAB_H): Remove.
3842
86aaf68d
TT		 38432013-09-25 Tom Tromey <tromey@redhat.com>
3844
3845 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3846
c987deb8
TT		 38472013-09-25 Tom Tromey <tromey@redhat.com>
3848
3849 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3850
9439747e
TT		 38512013-09-25 Tom Tromey <tromey@redhat.com>
3852
3853 * config/i386/t-i386 (i386.o): Remove.
3854 (i386-c.o): Use COMPILE and POSTCOMPILE.
3855
aefc31a1
TT		 38562013-09-25 Tom Tromey <tromey@redhat.com>
3857
3858 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3859
b6541edc
TT		 38602013-09-25 Tom Tromey <tromey@redhat.com>
3861
3862 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3863 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3864 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3865 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3866 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3867 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3868 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3869 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3870 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3871 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3872 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3873 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3874 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3875 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3876 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3877 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3878 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3879 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3880 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3881 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3882 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3883 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3884 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3885 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3886 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3887 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3888 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3889 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3890 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3891 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3892 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3893 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3894 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3895 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3896 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3897 (graphite-clast-to-gimple.o, graphite-dependences.o)
3898 (graphite-interchange.o, graphite-poly.o)
3899 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3900 (graphite-optimize-isl.o, tree-vect-loop.o)
3901 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3902 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3903 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3904 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3905 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3906 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3907 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3908 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3909 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3910 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3911 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3912 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3913 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3914 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3915 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3916 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3917 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3918 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3919 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3920 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3921 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3922 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3923 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3924 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3925 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3926 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3927 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3928 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3929 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3930 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3931 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3932 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3933 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3934 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3935 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3936 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3937 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3938 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3939 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3940 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3941 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3942 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3943 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3944 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3945 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3946 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3947 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3948 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3949 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3950 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3951 (gcov-dump.o): Remove.
3952 (default-c.o): Use COMPILE and POSTCOMPILE.
3953 (CFLAGS-gcc.o): New variable.
3954 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3955
8c796d19
TT		 39562013-09-25 Tom Tromey <tromey@redhat.com>
3957
3958 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3959 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3960 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3961 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3962 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3963 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3964 (c-family/array-notation-common.o, c-family/stub-objc.o)
3965 (c-family/c-ubsan.o): Remove.
3966
0a6c2227
TT		 39672013-09-25 Tom Tromey <tromey@redhat.com>
3968
3969 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3970
f3bc55f0
TT		 39712013-09-25 Tom Tromey <tromey@redhat.com>
3972
3973 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3974 to add -DENABLE_SHARED_LIBGCC.
3975 (gcc.o): Don't use subshell.
3976
c11c795e
TT		 39772013-09-25 Tom Tromey <tromey@redhat.com>
3978
3979 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3980 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3981 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3982 * configure, config.in: Rebuild.
3983
043378c3
TT		 39842013-09-25 Tom Tromey <tromey@redhat.com>
3985
3986 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3987 (COMPILE, POSTCOMPILE): New variables.
3988 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3989 (DEPFILES): New variable.
3990 Include ".Po" files.
3991 * configure.ac: Add checks for dependency checking.
3992 * configure, aclocal.m4: Regenerate.
3993
eab34643
TT		 39942013-09-25 Tom Tromey <tromey@redhat.com>
3995
3996 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 3997 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 3998
d2db5e91
TT		 39992013-09-25 Tom Tromey <tromey@redhat.com>
4000
4001 * Makefile.in (generated_files): Add options.h,
4002 target-hooks-def.h, insn-opinit.h,
4003 common/common-target-hooks-def.h, pass-instances.def,
4004 c-family/c-target-hooks-def.h.
4005
af33629e
JL		 40062013-09-25 Jeff Law <law@redhat.com>
4007
4008 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4009 than foo[foo.length () - 1] to access last member in a vec.
4010 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4011
ed0d3051
RB		 40122013-09-25 Richard Biener <rguenther@suse.de>
4013
4014 PR middle-end/58521
4015 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4016
123485ca
JH		 40172013-09-25 Jan Hubicka <jh@suse.cz>
4018
4019 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4020 test.
4021
d70b0f1f
MP		 40222013-09-25 Marek Polacek <polacek@redhat.com>
4023
4024 PR sanitizer/58420
4025 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4026 when determining the type name.
4027
e3f02798
OE		 40282013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4029
4030 * config/sh/sh.md: Fix formatting.
4031
7a5911d3 40322013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 4033
7a5911d3
UB		 4034 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4035 max peel iterations parameter.
4f17aa0b
XDL		 4036 * param.def: New parameter.
4037 * doc/invoke.texi: Document New parameter.
4038
0498471b
CL		 40392013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4040
4041 * gimple-pretty-print.c: Various whitespace tweaks.
4042 * tree-core.h: Likewise.
4043 * tree-pretty-print.c: Likewise.
4044 * tree-ssa-alias.c: Likewise.
4045 * tree-ssa-copy.c: Likewise.
4046 * tree-ssanames.c: Likewise.
4047 * tree-ssanames.h: Likewise.
4048 * tree-vrp.c: Likewise.
4049
5a8c2b57
AM		 40502013-09-24 Alan Modra <amodra@gmail.com>
4051
4052 PR middle-end/57134
4053 PR middle-end/57586
7a5911d3
UB		 4054 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4055 for output operands that disallow regs. Don't use EXPAND_WRITE on
4056 inout operands.
5a8c2b57 4057
f40333af
RB		 40582013-09-24 Richard Biener <rguenther@suse.de>
4059
4060 PR middle-end/58513
4061 * tree.c (reference_alias_ptr_type): Move ...
4062 * alias.c (reference_alias_ptr_type): ... here and implement
4063 in terms of the new reference_alias_ptr_type_1.
4064 (ref_all_alias_ptr_type_p): New helper.
4065 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4066 use ref_all_alias_ptr_type_p.
4067 (get_deref_alias_set): Add flag_strict_aliasing check here.
4068 (reference_alias_ptr_type_1): New function, split out from ...
4069 (get_alias_set): ... here.
4070 (alias_ptr_types_compatible_p): New function.
4071 * alias.h (reference_alias_ptr_type): Declare.
4072 (alias_ptr_types_compatible_p): Likewise.
4073 * tree.h (reference_alias_ptr_type): Remove.
4074 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4075 to compare MEM_REF alias types.
4076
583e8bf5
RB		 40772013-09-24 Richard Biener <rguenther@suse.de>
4078
4079 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4080
5e8586d7
MM		 40812013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4082
19c3e797
MM		 4083 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4084 reload helper function arrays into a single array reg_addr.
4085 (reload_fpr_gpr): Likewise.
4086 (reload_gpr_vsx): Likewise.
4087 (reload_vsx_gpr): Likewise.
4088 (struct rs6000_reg_addr): Likewise.
4089 (reg_addr): Likewise.
4090 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4091 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4092 (rs6000_init_hard_regno_mode_ok): Likewise.
4093 (rs6000_secondary_reload_direct_move): Likewise.
4094 (rs6000_secondary_reload): Likewise.
4095
5e8586d7
MM		 4096 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4097 constraints: wu, ww, and wy. Repurpose wv constraint added during
4098 power8 changes. Put wg constraint in alphabetical order.
4099
4100 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4101 for future work to add ISA 2.07 VSX single precision support.
4102 (-mvsx-scalar-double): Change default from -1 to 1, update
4103 documentation comment.
4104 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4105 (-mupper-regs-df): New debug switch to control whether DF values
4106 can go in the traditional Altivec registers.
4107 (-mupper-regs-sf): New debug switch to control whether SF values
4108 can go in the traditional Altivec registers.
4109
4110 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4111 and wy constraints.
4112 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4113 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4114 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4115 (rs6000_debug_legitimate_address_p): Print if we are running
4116 before, during, or after reload.
4117 (rs6000_secondary_reload): Add a comment.
4118 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4119
4120 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4121 constraints. Update documentation string.
4122 (wd constraint): Likewise.
4123 (wf constraint): Likewise.
4124 (wg constraint): Likewise.
4125 (wn constraint): Likewise.
4126 (ws constraint): Likewise.
4127 (wt constraint): Likewise.
4128 (wx constraint): Likewise.
4129 (wz constraint): Likewise.
4130 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4131 instructions.
4132 (ww constraint): Likewise.
4133 (wy constraint): Likewise.
358e1993
MM		 4134 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4135 in the previous submissions.
5e8586d7
MM		 4136 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4137
e5af9ddd
RS		 41382013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4139
4140 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4141 (REG_BR_PROB): Say that the probability is stored in an int_list.
4142 * reg-notes.def: Update commentary to mention INT_LIST.
4143 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4144 (INT_LIST): New rtx.
4145 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4146 * rtlanal.c (int_reg_note_p): New function.
4147 (alloc_reg_note): Assert that the note does not have an int argument.
4148 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4149 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4150 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4151 rather than an INSN_LIST. Handle INT_LIST.
4152 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4153 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4154 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4155 Manipulate them as ints rather than rtxes.
4156 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4157 * regmove.c (copy_src_to_dest): Likewise.
4158 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4159
4160 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4161 into the cases that need it.
4162 * config/arm/arm.c (arm_unwind_emit): Likewise.
4163
4164 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4165 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4166 * loop-doloop.c (add_test, doloop_modify): Likewise.
4167 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4168 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4169 * predict.c (combine_predictions_for_insn): Likewise.
4170 * print-rtl.c (print_rtx): Handle INT_LIST.
4171 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4172 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4173 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4174 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4175 (ix86_split_fp_branch, predict_jump): Likewise.
4176 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4177 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4178 * config/spu/spu.c (ea_load_store_inline): Likewise.
4179
4180 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4181 value of a REG_BR_PROB note.
4182 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4183 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4184 * emit-rtl.c (try_split): Likewise.
4185 * predict.c (br_prob_note_reliable_p): Likewise.
4186 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4187 * reorg.c (mostly_true_jump): Likewise.
4188 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4189 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4190 * config/i386/i386.c (ix86_print_operand): Likewise.
4191 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4192 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4193 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4194 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4195 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4196 * config/sparc/sparc.c (output_cbranch): Likewise.
4197 * config/spu/spu.c (get_branch_target): Likewise.
4198 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4199 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4200
450ad0cd
JH		 42012013-09-23 Jan Hubicka <jh@suse.cz>
4202
4203 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4204 for ipa-devirt.
4205 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4206 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4207 of external calls
4208 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4209 (gimple_fold_call): Dump inconsistent devirtualizations; add
4210 sanity check for type based devirtualizations.
4211 * ipa-prop.c: Include ipa-utils.h
4212 (ipa_intraprocedural_devirtualization): Add sanity check.
4213 (try_make_edge_direct_virtual_call): Likewise.
4214
cefce769
EB		 42152013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4216
4217 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4218 assignment statements.
4219
a895a2b8
KV		 42202013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4221
4222 * gimple-pretty-print.c (dump_ssaname_info): New function.
4223 (dump_gimple_phi): Call it.
4224 (pp_gimple_stmt_1): Likewise.
4225 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4226 (range_info_def): Declare.
4227 * tree-pretty-print.c (pp_double_int): New function.
4228 (dump_generic_node): Call it.
4229 * tree-pretty-print.h (pp_double_int): Declare.
4230 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4231 * tree-ssanames.h (range_info_def): New structure.
4232 (value_range_type): Move definition here.
4233 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4234 Declare.
4235 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4236 initialization.
4237 (set_range_info): New function.
4238 (get_range_info): Likewise.
4239 (duplicate_ssa_name_range_info): Likewise.
4240 (duplicate_ssa_name_fn): Check pointer type and call
4241 duplicate_ssa_name_range_info.
4242 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4243 * tree-vrp.c (value_range_type): Remove definition, now in
4244 tree-ssanames.h.
7a5911d3 4245 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV		 4246 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4247 (SSA_NAME_RANGE_INFO): New macro.
4248
984af6ac
RB		 42492013-09-23 Richard Biener <rguenther@suse.de>
4250
4251 PR tree-optimization/58464
4252 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 4253 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB		 4254 (phi_translate_1): Remove recursion detection here.
4255 (phi_translate): Pre-seed the cache with NULL to catch
4256 recursion here in a more generic way.
4257 (bitmap_find_leader): Adjust comment.
4258 (get_representative_for): Dump value-numbers.
4259 (create_expression_by_pieces): Likewise.
4260 (insert_into_preds_of_block): Likewise.
4261
0e26cf79
CB		 42622013-09-23 Christian Bruel <christian.bruel@st.com>
4263
4264 PR target/58475
4265 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4266 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4267
4d0e5b54
JG		 42682013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4269
4270 Revert r202780:
4271 2013-09-20 Renlin Li <renlin.li@arm.com>
4272
7a5911d3
UB		 4273 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4274 plus_constant.
4d0e5b54
JG		 4275 (aarch64_expand_epilogue): Likewise.
4276 (aarch64_legitimize_reload_address): Likewise.
4277
ca081cc8
EB		 42782013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4279
4280 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4281 NULL_TREE before pushing them onto the vector. Likewise for labels.
4282
8fc53a5f
EB		 42832013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4284
4285 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4286 comparison operators when -fno-trapping-math is in effect.
4287 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4288 comparison operators in TFmode and assert that unsupported operators
4289 cannot reach here.
4290 (ia64_print_operand): Likewise.
4291
42922013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH		 4293
4294 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4295 (sse_typeless_stores): Enable for core
4296 (sse_load0_by_pxor): Likewise.
4297 (four_jump_limit): Disable for core.
4298 (pad_returns): Likewise.
4299 (avoid_vector_decode): Likewise.
4300 (fuse_cmp_and_branch): Enable for cores.
4301 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4302
578e51e6
JDA		 43032013-09-20 John David Anglin <danglin@gcc.gnu.org>
4304
96ce28a3
JDA		 4305 PR middle-end/56791
4306 * config/pa/pa.c (pa_option_override): Disable auto increment and
4307 decrement instructions until reload is completed.
4308
578e51e6
JDA		 4309 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4310 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4311 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4312
6fcd3a13 43132013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 4314 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD		 4315
4316 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4317 (need_to_save): Save bank 0 on interrupts.
4318 (characterize_address): Strip far address wrappers.
4319 (rl78_as_legitimate_address): Likewise.
4320 (transcode_memory_rtx): Likewise.
4321 (rl78_peep_movhi_p): Disable this peephole after devirt.
4322 (rl78_propogate_register_origins): Forget all origins when a
4323 CLOBBER is seen.
4324 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4325 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4326 peephole2's.
4327 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4328 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4329 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4330 relocating.
4331 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4332
4a08db33
JDA		 43332013-09-20 John David Anglin <danglin@gcc.gnu.org>
4334
4335 * config/pa/pa.md: In "scc" insn patterns, change output template to
4336 handle const0_rtx in reg_or_0_operand operands.
4337
7691132c
JBG		 43382013-09-20 Martin Husemann <martin@NetBSD.org>
4339
4340 PR target/56875
4341 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4342 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4343
a3cc13cc
RB		 43442013-09-20 Richard Biener <rguenther@suse.de>
4345
4346 PR middle-end/58484
4347 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4348 remembering SSA name version and block index.
4349 (new_scev_info_str): Adjust.
4350 (hash_scev_info): Likewise. Also hash the block index.
4351 (eq_scev_info): Adjust.
4352 (find_var_scev_info): Likewise.
4353 (struct instantiate_cache_entry): Remove.
4354 (struct instantiate_cache_type): Use a htab to map name, block
4355 to chrec.
4356 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4357 (get_instantiated_value_entry): Likewise.
4358 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4359 (instantiate_scev_name): Adjust.
4360
925f3871
JL		 43612013-09-20 Jeff Law <law@redhat.com>
4362
4363 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4364
2888c331
YZ		 43652013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4366
4367 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4368 Call aarch64_simd_expand_args to update op[argc].
4369
0a811e96
BS		 43702013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4371
7a5911d3
UB		 4372 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4373 plugin argument.
0a811e96 4374
0078f462
BS		 43752013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4376
7a5911d3
UB		 4377 * gengtype.c (file_rules): Added rule for *.cc files.
4378 (get_output_file_with_visibility): Give fatal message when no
4379 rules found.
0078f462 4380
d20e48b0
RL		 43812013-09-20 Renlin Li <renlin.li@arm.com>
4382
4383 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4384 (aarch64_expand_epilogue): Likewise.
4385 (aarch64_legitimize_reload_address): Likewise.
4386
59bb1544
BE		 43872013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4388
4389 PR middle-end/57748
4390 * expr.c (expand_assignment): Remove misalignp code path.
4391
a5430019
MP		 43922013-09-20 Marek Polacek <polacek@redhat.com>
4393
4394 PR sanitizer/58413
4395 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4396 TYPE_PRECISION. Add asserts.
4397
9fed7f3a
RB		 43982013-09-20 Richard Biener <rguenther@suse.de>
4399
4400 PR tree-optimization/58453
4401 * tree-loop-distribution.c (distribute_loop): Apply the cost
4402 model for -ftree-loop-distribute-patterns, too.
4403
a820c834
RB		 44042013-09-20 Richard Biener <rguenther@suse.de>
4405
4406 PR middle-end/58473
4407 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4408 make type comparison less strict.
4409
3cbe17f7
AM		 44102013-09-20 Alan Modra <amodra@gmail.com>
4411
4412 * configure: Regenerate.
4413 * aclocal.m4: Regenerate.
4414
87fccdbb
MP		 44152013-09-20 Marek Polacek <polacek@redhat.com>
4416
4417 PR other/58467
4418 * doc/extend.texi: Document that attribute used is meant to be used
4419 on variables with static storage duration.
4420
8928eff3
JJ		 44212013-09-19 Jakub Jelinek <jakub@redhat.com>
4422
4423 PR tree-optimization/58472
4424 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4425 simd_lane_access set inv_p = false.
4426 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4427 the simduid magic VAR_DECL.
4428
b7b28c53
JH		 44292013-09-19 Jan Hubicka <jh@suse.cz>
4430
4431 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4432
83ae86f5
JL		 44332013-09-17 Jeff Law <law@redhat.com>
4434
4435 * tree-ssa-dom.c (record_temporary_equivalences): New function
4436 split out of dom_opt_dom_walker::after_dom_children.
4437 (dom_opt_dom_walker::thread_across_edge): Move common code
4438 in here from dom_opt_dom_walker::after_dom_children.
4439 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4440
9d532162
JH		 44412013-09-19 Jan Hubicka <jh@suse.cz>
4442
4443 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4444 (TARGET_GENERIC): Use PROCESOR_GENERIC
4445 (enum processor_type): Unify generic32 and 64.
4446 * i386.md (cpu): Likewise.
4447 * x86-tune.def (use_leave): Enable for generic32.
4448 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4449 * athlon.md: Change generic64 to generic in all occurences.
4450 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4451 (ix86_target_macros_internal): Likewise.
4452 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB		 4453 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4454 to ..
9d532162
JH		 4455 (generic_memcpy, generic_memset, generic_cost): This one.
4456 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4457 (m_GENERIC32, m_GENERIC64): Remove.
4458 (m_GENERIC): Turn into one flag.
4459 (processor_target): Unify generic tunnings.
4460 (ix86_option_override_internal): Replace generic32/64 by generic.
4461 (ix86_issue_rate): Likewise.
4462 (ix86_adjust_cost): Likewise.
4463
188c7d00
JH		 44642013-09-19 Jan Hubicka <jh@suse.cz>
4465
4466 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4467 of speculative flag.
4468
4befd127
JJ		 44692013-09-19 Jakub Jelinek <jakub@redhat.com>
4470
4471 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4472 GOMP_sections_start, even if !exit_reachable.
4473
8a26ad39
VM		 44742013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4475
4476 * lra-constraints.c (need_for_all_save_p): Use macro
4477 HARD_REGNO_CALL_PART_CLOBBERED.
4478 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4479 set up pseudo conflict hard regs.
4480
09487185
MM		 44812013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4482
4483 PR target/58452
4484 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4485 operands.
4486
080cbf9e
VM		 44872013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4488
4489 PR rtl-optimization/58438
4490 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4491 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4492 from previous subpasses.
4493
a0f70fcb
RE		 44942013-09-18 Richard Earnshaw <rearnsha@arm.com>
4495
4496 * arm.c (arm_get_frame_offsets): Validate architecture supports
4497 LDRD/STRD before accepting the tuning preference.
4498 (arm_expand_prologue): Likewise.
4499 (arm_expand_epilogue): Likewise.
4500
0547c9b6
RB		 45012013-09-18 Richard Biener <rguenther@suse.de>
4502
4503 PR tree-optimization/58417
4504 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4505 have chrecs with symbols defined in the loop as operands.
4506 (chrec_fold_multiply): Likewise.
4507 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4508 parameters before folding binary operations.
4509 (struct instantiate_cache_entry_hasher): Remove.
4510 (struct instantiate_cache_type): Use a pointer-map.
4511 (instantiate_cache_type::instantiate_cache_type): New function.
4512 (instantiate_cache_type::get): Likewise.
4513 (instantiate_cache_type::set): Likewise.
4514 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4515 (get_instantiated_value_entry): Likewise.
4516 (global_cache): New global.
4517 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4518 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4519 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4520 (instantiate_scev_name): Adjust.
4521 (instantiate_scev): Construct global instead of local cache.
4522 (resolve_mixers): Likewise.
4523
309dc1aa
DM		 45242013-09-18 Daniel Morris <danielm@ecoscentric.com>
4525 Paolo Carlini <paolo.carlini@oracle.com>
4526
4527 PR c++/58458
4528 * doc/implement-cxx.texi: Fix references to the C++ standards.
4529
995a1b4a
JJ		 45302013-09-18 Jakub Jelinek <jakub@redhat.com>
4531
4532 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4533 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4534 simd_lane_access drs, update also DR_ALIGNED_TO.
4535
ce6923c5
MP		 45362013-09-18 Marek Polacek <polacek@redhat.com>
4537
4538 PR sanitizer/58411
4539 * doc/extend.texi: Document no_sanitize_undefined attribute.
4540 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4541 no_sanitize_undefined attribute.
4542
d30d00a2
NC		 45432013-09-18 Nick Clifton <nickc@redhat.com>
4544
4545 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4546 (ASM_DECLARE_FUNCTION_NAME): Define.
4547
4d9192b5
TS		 45482013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4549
4550 * compare-elim.c (find_comparison_dom_walker): New class
4551 (find_comparisons_in_bb): Rename to
4552 find_comparison_dom_walker::before_dom_children
4553 (find_comparisons): Adjust
4554 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4555 adjust.
4556 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4557 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4558 (init_walk_dominator_tree): Remove declaration.
4559 (fini_walk_dominator_tree): Remove declaration.
4560 * fwprop.c (single_def_use_dom_walker): New class
4561 (single_def_use_enter_block): Convert to
4562 single_def_use_dom_walker::before_dom_children.
4563 (single_def_use_leave_block): Convert to
4564 single_def_use_dom_walker::after_dom_children.
4565 (build_single_def_use_links): Adjust.
4566 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4567 class.
4568 (find_candidates_in_block): Convert to
4569 find_candidates_dom_walker::before_dom_children.
4570 (execute_strength_reduction): Adjust.
4571 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4572 (sese_dom_walker): New class.
4573 (sese_dom_walker::sese_dom_walker): New constructor.
4574 (sese_dom_walker::~sese_dom_walker): New destructor.
4575 (build_sese_conditions_before): Convert to
4576 sese_dom_walker::before_dom_children.
4577 (build_sese_conditions_after): Convert to
4578 sese_dom_walker::after_dom_children.
4579 (build_poly_scop): Adjust
4580 * tree-into-ssa.c (rewrite_dom_walker): New class
4581 (rewrite_enter_block): Convert to
4582 rewrite_dom_walker::before_dom_children.
4583 (rewrite_leave_block): Convert to
4584 rewrite_dom_walker::after_dom_children.
4585 (rewrite_update_dom_walker): New class.
4586 (rewrite_update_enter_block): Convert to
4587 rewrite_update_dom_walker::before_dom_children.
4588 (rewrite_update_leave_block): Convert to
4589 rewrite_update_dom_walker::after_dom_children.
4590 (rewrite_blocks, rewrite_into_ssa): Adjust.
4591 (mark_def_dom_walker): New class.
4592 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4593 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4594 (mark_def_sites_blocks): Convert to
4595 mark_def_dom_walker::before_dom_children.
4596 (mark_def_site_blocks): Remove.
4597 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4598 (tree_ssa_dominator_optimize): Adjust.
4599 (dom_thread_across_edge): Convert to method
4600 dom_opt_dom_walker::thread_across_edge.
4601 (dom_opt_enter_block): Convert to member function
4602 dom_opt_dom_walker::before_dom_children.
4603 (dom_opt_leave_block): Convert to member function
4604 dom_opt_dom_walker::after_dom_children.
4605 * tree-ssa-dse.c (dse_dom_walker): New class.
4606 (dse_enter_block): Convert to member function
4607 dse_dom_walker::before_dom_children.
4608 (tree_ssa_dse): Adjust.
4609 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4610 (determine_invariantness_stmt): Convert to method
4611 invariantness_dom_walker::before_dom_children.
4612 (determine_invariantness): Remove
4613 (move_computations_dom_walker): New class.
4614 (move_computations_stmt): Convert to method
4615 move_computations_dom_walker::before_dom_children.
4616 (move_computations, tree_ssa_lim): Adjust.
4617 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4618 (nt_init_block): Make method
4619 notrappping_dom_walker::before_dom_children.
4620 (nt_fini_block): Make
4621 method nontrapping_dom_walker::after_dom_children.
4622 (get_non_trapping): Adjust.
4623 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4624 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4625 (eliminate_leave_block): Make method.
4626 eliminate_dom_walker::after_dom_children.
4627 (eliminate): Adjust
4628 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4629 (strlen_enter_block): Make method
4630 strlen_dom_walker::before_dom_children.
4631 (strlen_leave_block): Make
4632 method strlen_dom_walker::after_dom_children.
4633 (tree_ssa_strlen): Adjust.
4634 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4635 (tree_ssa_uncprop): Adjust.
4636 (uncprop_leave_block): Make method
4637 uncprop_dom_walker::after_dom_children.
4638 (uncprop_leave_block): Make method
4639 uncprop_dom_walker::before_dom_children.
4640
3b3cc26b
BC		 46412013-09-18 Bin Cheng <bin.cheng@arm.com>
4642
7a5911d3
UB		 4643 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4644 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4645
b5ba3a9b
NC		 46462013-09-17 Nick Clifton <nickc@redhat.com>
4647
4648 * config/rl78/rl78-real.md (bf): New pattern.
4649 (bt): New pattern.
4650 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4651 (rl78_print_operand): Do not put a # before a %B.
4652 * config/rl78/rl78.opt: Tweak doc strings.
4653
09d5094b
DD		 46542013-09-17 DJ Delorie <dj@redhat.com>
4655
4656 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4657 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4658 (rl78_option_override): Likewise, if -mallregs.
4659 (is_virtual_register): Likewise.
4660 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4661 (REGNO_OK_FOR_BASE_P): Likewise.
4662 * config/rl78/rl78.opt (-mallregs): New.
4663
2ba1dca1
NC		 46642013-09-17 Nick Clifton <nickc@redhat.com>
4665
4666 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4667 For interrupt functions: save all call clobbered registers if the
4668 interrupt handler is not a leaf function.
4669 (rl78_expand_prologue): Always recompute the frame information.
4670 For interrupt functions: only select bank 0 if one of the bank 0
4671 registers is going to be psuhed.
4672
2e7c3f21
DD		 46732013-09-17 DJ Delorie <dj@redhat.com>
4674
4675 * config/rl78/constraints.md: For each W* constraint, rename to C*
4676 and create a W* constraint that checks for an optional ES: prefix
4677 pattern also.
4678 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4679 (es_addr): New. Used to wrap far addresses.
4680 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4681 (rl78_es_base): New.
4682 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4683 wrapped far addresses.
4684 (rl78_print_operand_1): Unwrap far addresses before processing.
4685 (rl78_lo16): Wrap far addresses in unspecs.
4686 (rl78_es_addr): New.
4687 (rl78_es_base): New.
4688 (insn_ok_now): Check for not-yet-wrapped far addresses.
4689 (transcode_memory_rtx): Properly re-wrap far addresses.
4690
5c26a69a
SH		 46912013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4692
4693 * config/sparc/t-rtems: Add leon3 multilibs.
4694
181f5f3e
CH		 46952013-09-17 Cong Hou <congh@google.com>
4696
4697 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4698 when checking the dot production pattern. The type of rhs operand
4699 of multiply is now checked correctly.
4700
8d34e421
JL		 47012013-09-17 Jeff Law <law@redhat.com>
4702
4703 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4704 edge implied equivalences into successor phis.
4705 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4706 here from tree-ssa-threadedge.c.
4707 (mark_threaded_blocks): When threading through a joiner, if both
4708 successors of the joiner's clone reach the same block, verify the
4709 PHI arguments are equal. If not, cancel the jump threading request.
4710 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4711 tree-ssa-threadupdate.c
4712 (thread_across_edge): Don't check PHI argument equality when
4713 threading through joiner block here.
7a5911d3 4714
c152901f
AM		 47152013-09-17 Andrew MacLeod <amacleod@redhat.com>
4716
4717 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4718 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB		 4719 (warn_uninit, warn_uninitialized_vars,
4720 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4721 Move to tree-ssa-uninit.c.
78cedfb1 4722 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 4723 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM		 4724 ssa_undefined_value_p.
4725 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB		 4726 (warn_uninit, warn_uninitialized_vars,
4727 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4728 Move from tree-ssa.c.
4729 * tree-ssa.h: Adjust prototypes.
c152901f 4730
b631d45a
JH		 47312013-09-17 Jan Hubicka <jh@suse.cz>
4732
4733 PR middle-end/58332
4734 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4735 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4736 FUNCTION_NOT_OPTIMIZED.
4737 * ipa-inline-analysis.c (compute_inline_parameters): Function
4738 not optimized is not inlinable unless it is alwaysinline.
4739 (inline_analyze_function): Force calls in not optimized
4740 function not inlinable.
4741
5b79657a
JH		 47422013-09-17 Jan Hubicka <jh@suse.cz>
4743
4744 PR middle-end/58329
4745 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4746 to return NULL.
4747 * ipa.c (function_and_variable_visibility): Likewise.
4748 * ipa-profile.c (ipa_profile): Likewise.
4749
27442c24
BE		 47502013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4751
4752 PR ipa/58398
4753 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4754 attribute, and don't inline the resolver in this case.
4755
753b9bc3
TJ		 47562013-09-17 Teresa Johnson <tejohnson@google.com>
4757
4758 * coverage.c (get_coverage_counts): Add missing newline.
4759
e3d3cfb4
KT		 47602013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4761
4762 PR tree-optimization/58088
4763 * fold-const.c (mask_with_trailing_zeros): New function.
4764 (fold_binary_loc): Make sure we don't recurse infinitely
4765 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4766 Use mask_with_trailing_zeros where appropriate.
4767
9a0ac98f
YR		 47682013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4769
4770 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4771 of 'prev' var to get better distance estimation.
4772
78bbd765
EB		 47732013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4774
4775 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4776 * tree-inline.c (is_parameter_of): New predicate.
4777 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4778 a parameter has been remapped.
4779 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4780 (optimize_inline_calls): Initialize transform_parameter.
4781 (copy_gimple_seq_and_replace_locals): Likewise.
4782 (tree_function_versioning): Likewise.
4783 (maybe_inline_call_in_expr): Likewise.
4784
cad055a4
NC		 47852013-09-17 Nick Clifton <nickc@redhat.com>
4786
4787 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4788 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4789 interrupt handlers.
4790 (is_attr_func): New function.
4791 (msp430_is_interrupt_func): New function.
4792 (is_naked_func): New function.
4793 (is_reentrant_func): New function.
4794 (is_critical_func): New function.
4795 (msp430_start_function): Add annotations for function attributes.
4796 (msp430_attr): New function.
4797 (msp430_attribute_table): New.
4798 (msp430_function_section): New function.
4799 (TARGET_ASM_FUNCTION_SECTION): Define.
4800 (msp430_builtin): New enum.
4801 (msp430_init_builtins): New function.
4802 (msp430_builtin_devl): New function.
4803 (msp430_expand_builtin): New function.
4804 (TARGET_INIT_BUILTINS): Define.
4805 (TARGET_EXPAND_BUILTINS): Define.
4806 (TARGET_BUILTIN_DECL): Define.
4807 (msp430_expand_prologue): Add support for naked, interrupt,
4808 critical and reentrant functions.
4809 (msp430_expand_epilogue): Likewise.
4810 (msp430_print_operand): Handle 'O' character.
4811 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4812 NO_TRAMPOLINES.
4813 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4814 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4815 (pushm): Use a 'n' rather than an 'i' constraint.
4816 (msp_return): Add generation of the interrupt return instruction.
4817 (disable_interrupts): New pattern.
4818 (enable_interrupts): New pattern.
4819 (push_intr_state): New pattern.
4820 (pop_intr_state): New pattern.
4821 (bic_SR): New pattern.
4822 (bis_SR): New pattern.
4823 * doc/extend.texi: Document MSP430 function attributes and builtin
4824 functions.
4825
deb6c11a
RB		 48262013-09-17 Richard Biener <rguenther@suse.de>
4827
4828 PR tree-optimization/58432
4829 * tree-loop-distribution.c (tree_loop_distribution): Also
4830 scan PHIs for outside loop uses and seed a partition from them.
4831
c068654b
BC		 48322013-09-17 Bin Cheng <bin.cheng@arm.com>
4833
4834 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4835 (restructure_reference): Call backtrace_base_for_ref.
4836
567969e4
AM		 48372013-09-17 Alan Modra <amodra@gmail.com>
4838
4839 PR target/57589
7a5911d3
UB		 4840 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4841 patch.
567969e4 4842
5c0029de
DD		 48432013-09-16 DJ Delorie <dj@redhat.com>
4844
4845 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4846 vregs location for RL78/G10.
4847 (rl78_expand_prologue): Avoid SEL on G10.
4848 (rl78_expand_epilogue): Likewise.
4849 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4850 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4851 __RL78_G10__ when appropriate.
4852 (ASM_SPEC): Pass -mg10 along to the assembler.
4853 * config/rl78/rl78.md (sel_rb): Disable for G10.
4854 * config/rl78/rl78.opt: Add -mg10 option.
4855 * config/rl78/t-rl78: Add -mg10 multilib.
4856
ea0f3e87
XDL		 48572013-09-16 Xinliang David Li <davidxl@google.com>
4858
4859 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4860 * omp-low.c (omp_max_vf): Ditto.
4861 (expand_omp_simd): Ditto.
4862 * tree-vectorizer.c (vectorize_loops): Ditto.
4863 (gate_vect_slp): Ditto.
4864 (gate_increase_alignment): Ditto.
4865 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4866 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4867 (gate_tree_vectorize): Name change.
4868 (tree_vectorize): Ditto.
4869 (pass_vectorize::gate): Call new function.
4870 (pass_vectorize::execute): Ditto.
7a5911d3 4871 * opts.c: O3 default setting change.
ea0f3e87
XDL		 4872 (finish_options): Check new flag.
4873 * doc/invoke.texi: Document new flags.
4874 * common.opt: New flags.
4875
4f3da2e5
AS		 48762013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4877
4878 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4879 (Condition Code): Don't reference it.
4880 * doc/tm.texi: Regenerate.
4881
aa940f7c
VM		 48822013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4883
4884 PR middle-end/58418
4885 * lra-constraints.c (undo_optional_reloads): Consider all optional
4886 reload even if it did not get a hard reg.
4887
e645e942
TJ		 48882013-09-16 Teresa Johnson <tejohnson@google.com>
4889
4890 * dumpfile.c (dump_loc): Remove newline emission.
4891 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4892 emission to dump_printf_loc calls where missing.
4893 (vect_mark_for_runtime_alias_test): Ditto.
4894 (vect_analyze_data_ref_dependence): Ditto.
4895 (vect_analyze_data_ref_dependences): Ditto.
4896 (vect_slp_analyze_data_ref_dependence): Ditto.
4897 (vect_slp_analyze_data_ref_dependences): Ditto.
4898 (vect_compute_data_ref_alignment): Ditto.
4899 (vect_update_misalignment_for_peel): Ditto.
4900 (vect_verify_datarefs_alignment): Ditto.
4901 (vector_alignment_reachable_p): Ditto.
4902 (vect_get_data_access_cost): Ditto.
4903 (vect_enhance_data_refs_alignment): Ditto.
4904 (vect_find_same_alignment_drs): Ditto.
4905 (vect_analyze_data_refs_alignment): Ditto.
4906 (vect_analyze_group_access): Ditto.
4907 (vect_analyze_data_ref_access): Ditto.
4908 (vect_analyze_data_ref_accesses): Ditto.
4909 (vect_prune_runtime_alias_test_list): Ditto.
4910 (vect_analyze_data_refs): Ditto.
4911 (vect_create_addr_base_for_vector_ref): Ditto.
4912 (vect_create_data_ref_ptr): Ditto.
4913 (vect_grouped_store_supported): Ditto.
4914 (vect_grouped_load_supported): Ditto.
4915 * value-prof.c (check_counter): Ditto.
4916 (check_ic_target): Ditto.
4917 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4918 (vect_recog_widen_mult_pattern): Ditto.
4919 (vect_recog_widen_sum_pattern): Ditto.
4920 (vect_recog_over_widening_pattern): Ditto.
4921 (vect_recog_widen_shift_pattern): Ditto.
4922 (vect_recog_rotate_pattern): Ditto.
4923 (vect_recog_vector_vector_shift_pattern): Ditto.
4924 (vect_recog_divmod_pattern): Ditto.
4925 (vect_recog_mixed_size_cond_pattern): Ditto.
4926 (vect_recog_bool_pattern): Ditto.
4927 (vect_pattern_recog_1): Ditto.
4928 (vect_pattern_recog): Ditto.
4929 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4930 (vect_is_simple_iv_evolution): Ditto.
4931 (vect_analyze_scalar_cycles_1): Ditto.
4932 (vect_get_loop_niters): Ditto.
4933 (vect_analyze_loop_1): Ditto.
4934 (vect_analyze_loop_form): Ditto.
4935 (vect_analyze_loop_operations): Ditto.
4936 (vect_analyze_loop_2): Ditto.
4937 (vect_analyze_loop): Ditto.
4938 (report_vect_op): Ditto.
4939 (vect_is_slp_reduction): Ditto.
4940 (vect_is_simple_reduction_1): Ditto.
4941 (vect_get_known_peeling_cost): Ditto.
4942 (vect_estimate_min_profitable_iters): Ditto.
4943 (vect_model_reduction_cost): Ditto.
4944 (vect_model_induction_cost): Ditto.
4945 (get_initial_def_for_induction): Ditto.
4946 (vect_create_epilog_for_reduction): Ditto.
4947 (vectorizable_reduction): Ditto.
4948 (vectorizable_induction): Ditto.
4949 (vectorizable_live_operation): Ditto.
4950 (vect_loop_kill_debug_uses): Ditto.
4951 (vect_transform_loop): Ditto.
4952 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4953 (vect_stmt_relevant_p): Ditto.
4954 (process_use): Ditto.
4955 (vect_mark_stmts_to_be_vectorized): Ditto.
4956 (vect_model_simple_cost): Ditto.
4957 (vect_model_promotion_demotion_cost): Ditto.
4958 (vect_model_store_cost): Ditto.
4959 (vect_get_store_cost): Ditto.
4960 (vect_model_load_cost): Ditto.
4961 (vect_get_load_cost): Ditto.
4962 (vect_init_vector_1): Ditto.
4963 (vect_get_vec_def_for_operand): Ditto.
4964 (vect_finish_stmt_generation): Ditto.
4965 (vectorizable_call): Ditto.
4966 (vectorizable_conversion): Ditto.
4967 (vectorizable_assignment): Ditto.
4968 (vectorizable_shift): Ditto.
4969 (vectorizable_operation): Ditto.
4970 (vectorizable_store): Ditto.
4971 (vectorizable_load): Ditto.
4972 (vectorizable_condition): Ditto.
4973 (vect_analyze_stmt): Ditto.
4974 (vect_transform_stmt): Ditto.
4975 (vect_is_simple_use): Ditto.
4976 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4977 (vect_can_advance_ivs_p): Ditto.
4978 (vect_update_ivs_after_vectorizer): Ditto.
4979 (vect_do_peeling_for_loop_bound): Ditto.
4980 (vect_gen_niters_for_prolog_loop): Ditto.
4981 (vect_update_inits_of_drs): Ditto.
4982 (vect_create_cond_for_alias_checks): Ditto.
4983 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4984 (vect_build_slp_tree_1): Ditto.
4985 (vect_supported_load_permutation_p): Ditto.
4986 (vect_analyze_slp_instance): Ditto.
4987 (vect_analyze_slp): Ditto.
4988 (vect_make_slp_decision): Ditto.
4989 (vect_detect_hybrid_slp): Ditto.
4990 (vect_bb_vectorization_profitable_p): Ditto.
4991 (vect_slp_analyze_bb_1): Ditto.
4992 (vect_update_slp_costs_according_to_vf): Ditto.
4993 (vect_get_mask_element): Ditto.
4994 (vect_transform_slp_perm_load): Ditto.
4995 (vect_schedule_slp_instance): Ditto.
4996 (vect_schedule_slp): Ditto.
4997 (vect_slp_transform_bb): Ditto.
4998 * profile.c (read_profile_edge_counts): Ditto.
4999 (compute_branch_probabilities): Ditto.
5000 * coverage.c (get_coverage_counts): Ditto.
5001
46f851f3
DN		 50022013-09-16 Diego Novillo <dnovillo@google.com>
5003
7a5911d3 5004 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 5005
1632a686
JH		 50062013-09-16 Jan Hubicka <jh@suse.cz>
5007
5008 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 5009 abstract functions; for static functions check the presence of body.
1632a686 5010
828e70c1
JG		 50112013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5012
5013 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5014 * config/aarch64/aarch64-simd.md
5015 (aarch64_mla_elt<mode>): New.
5016 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5017 (aarch64_mls_elt<mode>): Likewise.
5018 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5019 (aarch64_fma4_elt<mode>): Likewise.
5020 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5021 (aarch64_fma4_elt_to_128v2df): Likewise.
5022 (aarch64_fma4_elt_to_64df): Likewise.
5023 (fnma<mode>4): Likewise.
5024 (aarch64_fnma4_elt<mode>): Likewise.
5025 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5026 (aarch64_fnma4_elt_to_128v2df): Likewise.
5027 (aarch64_fnma4_elt_to_64df): Likewise.
5028 * config/aarch64/iterators.md (VDQSF): New.
5029 * config/aarch64/arm_neon.h
5030 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5031 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5032
779aea46
JG		 50332013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5034
5035 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5036 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5037 (aarch64_mul3_elt_to_128df): Likewise.
5038 (aarch64_mul3_elt_to_64v2df): Likewise.
5039 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5040 (VMUL): New.
5041 (VMUL_CHANGE_NLANES) Likewise.
5042 (h_con): Likewise.
5043 (f): Likewise.
5044 * config/aarch64/arm_neon.h
5045 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5046
a407a750
JG		 50472013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5048
5049 * config/aarch64/arm_neon.h
5050 (vcvtx_high_f32_f64): Fix parameters.
5051
f77e2d2d 50522013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 5053 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 5054
a5bd71fa 5055 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 5056
36875e8f
RB		 50572013-09-16 Richard Biener <rguenther@suse.de>
5058
5059 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5060 (dot_rdg_1): Handle control_dd.
5061 (create_edge_for_control_dependence): New function.
5062 (create_rdg_edges): Add control dependences if asked for.
5063 (build_rdg): Likewise.
5064 (generate_loops_for_partition): If there are not necessary
5065 control stmts remove all their dependencies.
5066 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5067 (distribute_loop): Pass on control dependences.
5068 (tree_loop_distribution): Compute control dependences and remove
5069 restriction on number of loop nodes.
5070
96e24d49
JJ		 50712013-09-16 Jakub Jelinek <jakub@redhat.com>
5072
5073 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5074 for internal calls.
5075
40dbb05c
RS		 50762013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5077
5078 * cse.c (try_const_anchors): Punt on CC modes.
5079
97191ad0
JBG		 50802013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5081
5082 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5083
403a3fb7
JDA		 50842013-09-14 John David Anglin <danglin@gcc.gnu.org>
5085
5086 PR target/58382
5087 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5088 calls to word_mode.
5089
fb5610fb
IS		 50902013-09-14 Iain Sandoe <iain@codesourcery.com>
5091
5092 PR target/48094
5093 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG		 5094 seen.
5095 (darwin_objc1_section): Likewise.
fb5610fb
IS		 5096 (darwin_file_end): Emit Image Info section when required.
5097
afdec9bd
JH		 50982013-09-14 Jan Hubicka <jh@suse.cz>
5099
5100 * tree-into-ssa.c (gate_into_ssa): New.
5101 (pass_data_build_ssa): Use it.
5102 * cgraph.h (expand_thunk): Update prototype.
5103 * cgraphunit.c (analyze_function): Expand thunks early.
5104 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5105 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5106 set lowered flag; do not add new function.
5107 (assemble_thunks_and_aliases): Update.
5108 * tree-ssa.c (gate_init_datastructures): New gate.
5109 (pass_data_init_datastructures): Use it.
5110
492d1e0a
IS		 51112013-09-14 Iain Sandoe <iain@codesourcery.com>
5112
5113 PR target/58269
f77e2d2d 5114 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS		 5115 xmm register set described in the psABI.
5116
d13c6b70
EG		 51172013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5118
f77e2d2d
JBG		 5119 * dwarf2out.c (should_emit_struct_debug): Add check
5120 for type_decl variable is not NULL.
d13c6b70 5121
88b0e79e
JC		 51222013-09-13 Jacek Caban <jacek@codeweavers.com>
5123
5124 * config.gcc: Use new winnt-c.c target hooks
5125 * config/t-winnt: New file
5126 * config/winnt-c.c: New file
5127 * doc/tm.texi.in: Document new hook
5128 * doc/tm.texi: Regenerated
5129
19ba6aab
JH		 51302013-09-13 Jan Hubicka <jh@suse.cz>
5131
5132 PR middle-end/58094
5133 * ipa-inline.c (check_callers): New function.
5134 (check_caller_edge): Remove.
5135 (want_inline_function_to_all_callers_p): Also permit alises that are
5136 called dirrectly.
5137 (inline_to_all_callers): Terminate the walk when devirtualization
5138 introduce new calls.
5139
a93c18c8
JH		 51402013-09-13 Jan Hubicka <jh@suse.cz>
5141
5142 * ipa-inline-analysis.c (struct growth_data): Add node.
5143 (do_estimate_growth_1): Fix detection of recursion.
5144
f3b331d1
JJ		 51452013-09-13 Jakub Jelinek <jakub@redhat.com>
5146
5147 PR tree-optimization/58392
5148 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5149 to avoid shadowing of outer loop variable. If
5150 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5151 replace_by_duplicate_decl simduid of loops that have it set and
5152 set dest_cfun->has_simduid_loops and/or
5153 dest_cfun->has_force_vect_loops.
5154 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5155 instead of maybe_lookup_decl.
5156 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5157 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5158 call. Copy over force_vect and copy and remap simduid. Set
5159 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5160 (copy_cfg_body): Remove blocks_to_copy argument. Use
5161 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5162 caller. Don't set cfun->has_simduid_loops and/or
5163 cfun->has_force_vect_loops here.
5164 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5165 caller.
5166 (expand_call_inline, tree_function_versioning): Adjust copy_body
5167 callers.
5168
ae6d0907
MJ		 51692013-09-13 Martin Jambor <mjambor@suse.cz>
5170
5171 PR bootstrap/58388
5172 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5173 the assert if the edge was a speculative one.
5174
2fd5894f
RB		 51752013-09-13 Richard Biener <rguenther@suse.de>
5176
5177 * tree-data-ref.h (known_dependences_p): Move here ...
5178 * tree-loop-distribution.c (known_dependences_p): ... from here.
5179 (dump_rdg_component, debug_rdg_component): Remove.
5180 (dump_rdg): Adjust.
5181 (generate_loops_for_partition): Use gimple_uid instead of
5182 relying on matching stmt visit order.
5183 (rdg_build_partitions): Take starting stmt vector.
5184 (ldist_gen): Merge into ...
5185 (distribute_loop): ... this function. Do not compute starting
5186 vertices vector.
5187 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5188
86eb4bd7
KT		 51892013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5190
5191 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5192 Set type attribute correctly. Set predicable_short_it attribute.
5193 (cmpsi_shiftsi): Remove %? from output template.
5194
83a95546
RB		 51952013-09-13 Richard Biener <rguenther@suse.de>
5196
5197 * tree-loop-distribution.c (struct rdg_component,
5198 rdg_defs_used_in_other_loops_p, free_rdg_components,
5199 rdg_build_components): Remove.
5200 (stmts_from_loop): Do not record virtual PHIs.
5201 (generate_loops_for_partition): Skip virtual PHIs.
5202 (build_rdg_partition_for_component): Rename to ...
5203 (build_rdg_partition_for_vertex): ... this and adjust.
5204 (rdg_build_partitions): Take a vector of starting vertices
5205 instead of components. Remove unnecessary leftover handling.
5206 (ldist_gen): Do not build components or record other stores.
5207 (distribute_loop): Do not distribute loops containing stmts
5208 with side-effects.
5209
c0ecf7f1
CB		 52102013-09-13 Christian Bruel <christian.bruel@st.com>
5211
5212 PR target/58314
5213 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5214
34f56efc
KT		 52152013-09-13 Kai Tietz <ktietz@redhat.com>
5216
740f2bf3
KT		 5217 * config.gcc: Separate cases for mingw and cygwin targets,
5218 and add 64-bit cygwin target case.
5219
34f56efc
KT		 5220 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5221 dll-export inline-functions.
5222 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5223
5224
52252013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL		 5226
5227 PR middle-end/58387
5228 Revert:
5229 2013-09-06 Jeff Law <law@redhat.com>
5230
5231 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5232 edge implied equivalences into successor phis.
5233
aadb52ad
DD		 52342013-09-12 DJ Delorie <dj@redhat.com>
5235
5236 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5237 separators.
5238
90ae7019
BM		 52392013-09-12 Brooks Moses <bmoses@google.com>
5240
5241 PR driver/42955
5242 * Makefile.in: Do not install driver binaries in $(target)/bin.
5243
17b2557c
DD		 52442013-09-12 DJ Delorie <dj@redhat.com>
5245
c5215a9b
DD		 5246 * config/rl78/rl78.opt (mrelax): New.
5247 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5248 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5249
17b2557c
DD		 5250 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5251 between SP and FP.
5252 (rl78_expand_epilogue): Likewise.
5253
ecafcf05
VM		 52542013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5255
5256 PR middle-end/58335
84f6d9fa 5257 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM		 5258 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5259 pointer elimination with using remove_reg_equal_offset_note.
5260
8369f38a
DD		 52612013-09-12 DJ Delorie <dj@redhat.com>
5262
f6a83b4a
DD		 5263 * config/msp430/: New port.
5264 * config.gcc (msp430): Added.
5265 * doc/invoke.texi: Document MSP430 options.
5266 * doc/install.texi: Document msp430-elf
5267 * doc/md.texi: Document msp430-elf
5268 * doc/contrib.texi: Document msp430-elf
5269
8369f38a
DD		 5270 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5271 PSImode.
5272
a854f856
MJ		 52732013-09-12 Martin Jambor <mjambor@suse.cz>
5274
5275 PR ipa/58389
5276 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5277 reference descriptor is NULL.
5278 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5279 edge to NULL.
5280
b12ebd96
AM		 52812013-09-12 Andrew MacLeod <amacleod@redhat.com>
5282
5283 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5284 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5285 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5286 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5287 make_temp_ssa_name): move to tree-ssanames.h
5288 * tree-ssa-alias.h: Move prototype.
5289 * tree-ssa.h: Include tree-ssanames.h.
5290 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5291 * tree-ssanames.h: New. Move items from tree-flow*.h
5292 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5293
e1fb4ad3
RB		 52942013-09-12 Richard Biener <rguenther@suse.de>
5295
5296 PR tree-optimization/58404
5297 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5298 propagate non-invariant addresses into dereferences wrapped
5299 in component references.
5300
154c9188
RB		 53012013-09-12 Richard Biener <rguenther@suse.de>
5302
5303 PR tree-optimization/58402
5304 * passes.def: Move pass_late_warn_uninitialized later.
5305
7a300452
AM		 53062013-09-12 Andrew MacLeod <amacleod@redhat.com>
5307
5308 * tree-ssa.h: New. Move content from tree-flow.h and
5309 tree-flow-inline.h.
5310 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5311 Move prototypes belonging to tree-ssa.c.
f77e2d2d 5312 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM		 5313 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5314 tree-ssa.h.
5315 * gimple.h: Adjust prototypes.
f77e2d2d
JBG		 5316 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5317 to...
5318 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM		 5319 * tree.h: Move prototype to tree-ssa.h.
5320 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5321 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5322 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5323 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5324 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5325 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5326 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5327 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG		 5328 graphite-blocking.c, graphite-clast-to-gimple.c,
5329 graphite-dependences.c, graphite-interchange.c,
5330 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5331 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5332 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5333 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM		 5334 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5335 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5336 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5337 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5338 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5339 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5340 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5341 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5342 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5343 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5344 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5345 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5346 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5347 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5348 value-prof.c, var-tracking.c,
5349 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5350
97463b2b
RB		 53512013-09-12 Richard Biener <rguenther@suse.de>
5352
5353 PR tree-optimization/58396
5354 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5355 (build_rdg): Take a loop-nest parameter, fix memleaks.
5356 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5357
80ecd89d
YR		 53582013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5359
5360 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5361 for SLM.
5362
893e85fa
CM		 53632013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5364
5365 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5366 The FMA instruction names should have a 'v' prefix.
5367
174ec470
RB		 53682013-09-12 Richard Biener <rguenther@suse.de>
5369
5370 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5371 (dot_rdg): Use popen instead of system in optional code.
5372 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5373 (already_processed_vertex_p): Adjust.
5374 (has_anti_or_output_dependence, predecessor_has_mem_write,
5375 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5376 rdg_flag_uses): Remove.
5377 (rdg_flag_vertex): Simplify.
5378 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5379 remove recursion.
5380 (build_rdg_partition_for_component): Process the first vertex
5381 of a component only.
5382 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5383
38c89763
AM		 53842013-09-12 Alan Modra <amodra@gmail.com>
5385
5386 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5387
3e3d9d17
DD		 53882013-09-11 DJ Delorie <dj@redhat.com>
5389 Nick Clifton <nickc@redhat.com>
5390
5391 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5392 (rl78_stack_based_mem): New.
5393 * config/rl78/constraints.md (Iv08): New.
5394 (Iv16): New.
5395 (Iv24): New.
5396 (Is09): New.
5397 (Is17): New.
5398 (Is25): New.
5399 (ISsi): New.
5400 (IShi): New.
5401 (ISqi): New.
5402 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5403 (movhi): Likewise.
5404 (movsi): Change from expand to insn-and-split.
5405 (ashrsi3): Clobber AX.
5406 (lshrsi3): New.
5407 (ashlsi3): New.
5408 (cbranchsi4): New.
5409 * config/rl78/rl78.md (CC_REG): Fix.
5410 (addsi3): Allow memory and immediate operands.
5411 (addsi3_internal): Split into...
5412 (addsi3_internal_virt): ...new, and ...
5413 (addsi3_internal_real): ...new.
5414 (subsi): New.
5415 (subsi3_internal_virt): New.
5416 (subsi3_internal_real): New.
5417 (mulsi3): Add memory operand.
5418 (mulsi3_rl78): Likewise.
5419 (mulsi3_g13): Likewise.
5420 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5421 (cbranchqi4_real): Add more constraint options.
5422 (cbranchhi4_real): Expand pattern.
5423 (cbranchhi4_real_signed): New.
5424 (cbranchhi4_real_inverted): New.
5425 (cbranchsi4_real_lt): New.
5426 (cbranchsi4_real_ge): New.
5427 (cbranchsi4_real_signed): New.
5428 (cbranchsi4_real): New.
5429 (peephole2): New.
893e85fa
CM		 5430 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5431 constant shifts.
3e3d9d17
DD		 5432 (lshrsi3_virt): Likewise.
5433 (ashlsi3_virt): Likewise.
5434 (cbranchqi4_virt_signed): New.
5435 (cbranchhi4_virt_signed): New.
5436 (cbranchsi4_virt): New.
5437 * config/rl78/rl78.c: Whitespace fixes throughout.
5438 (move_elim_pass): New.
5439 (pass_data_rl78_move_elim): New.
5440 (pass_rl78_move_elim): New.
5441 (make_pass_rl78_move_elim): New.
5442 (rl78_devirt_info): Run devirt earlier.
5443 (rl78_move_elim_info): New.
5444 (rl78_asm_file_start): Register it.
5445 (rl78_split_movsi): New.
5446 (rl78_as_legitimate_address): Allow virtual base registers when
5447 appropriate.
5448 (rl78_addr_space_convert): Remove spurious debug stuff.
5449 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5450 (rl78_print_operand): More cases for not printing '#'.
5451 (rl78_expand_compare): Remove most of the logic.
5452 (content_memory): New.
5453 (clear_content_memory): New.
5454 (get_content_index): New.
5455 (get_content_name): New.
5456 (display_content_memory): New.
5457 (update_content): New.
5458 (record_content): New.
5459 (already_contains): New.
5460 (insn_ok_now): Re-recog insns with virtual registers.
5461 (add_postponed_content_update): New.
5462 (process_postponed_content_update): New.
5463 (gen_and_emit_move): New.
893e85fa
CM		 5464 (transcode_memory_rtx): Record new location content.
5465 Use gen_and_emit_move.
3e3d9d17
DD		 5466 (force_into_acc): New.
5467 (move_to_acc): Use gen_and_emit_move.
5468 (move_from_acc): Likewise.
5469 (move_acc_to_reg): Likewise.
5470 (move_to_x): Likewise.
5471 (move_to_hl): Likewise.
5472 (move_to_de): Likewise.
5473 (rl78_alloc_physical_registers_op1): Record location content.
5474 (has_constraint): New.
5475 (rl78_alloc_physical_registers_op2): Record location content.
5476 Optimize use of HL.
5477 (rl78_alloc_physical_registers_ro1): Likewise.
5478 (rl78_alloc_physical_registers_cmp): Likewise.
5479 (rl78_alloc_physical_registers_umul): Likewise.
5480 (rl78_alloc_address_registers_macax): New.
5481 (rl78_alloc_physical_registers): Initialize and set location
5482 content memory as needed.
5483 (rl78_reorg): Make sure split2 is called.
5484 (rl78_rtx_costs): New.
5485
088845a5
RS		 54862013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5487
5488 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5489 for (not (neg ...)) and (neg (not ...)) cases.
5490
5c2961cf
RB		 54912013-09-11 Richard Biener <rguenther@suse.de>
5492
5493 PR middle-end/58377
5494 * passes.def: Split critical edges before late uninit warning passes.
5495 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5496
427e6a14
JJ		 54972013-09-11 Jakub Jelinek <jakub@redhat.com>
5498
5499 PR tree-optimization/58385
5500 * fold-const.c (build_range_check): If both low and high are NULL,
5501 use omit_one_operand_loc to preserve exp side-effects.
5502
df45c0cc
KT		 55032013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5504
5505 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5506
80ab0b19
RB		 55072013-09-11 Richard Biener <rguenther@suse.de>
5508
5509 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5510 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5511 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5512 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5513 stmts_from_loop, known_dependences_p, build_empty_rdg,
5514 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5515 * tree-loop-distribution.c: ... here.
5516 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5517 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5518 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5519 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5520 * tree-loop-distribution.c: ... here.
5521 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5522 (struct partition_s): Add loops member.
5523 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5524 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5525 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5526
85a77221
AI		 55272013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5528 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5529 Sergey Lega <sergey.s.lega@intel.com>
5530 Anna Tikhonova <anna.tikhonova@intel.com>
5531 Ilya Tocar <ilya.tocar@intel.com>
5532 Andrey Turetskiy <andrey.turetskiy@intel.com>
5533 Ilya Verbin <ilya.verbin@intel.com>
5534 Kirill Yukhin <kirill.yukhin@intel.com>
5535 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5536
5537 * config/i386/constraints.md (k): New.
5538 (Yk): Ditto.
5539 * config/i386/i386.c (const regclass_map): Add new mask registers.
5540 (dbx_register_map): Ditto.
5541 (dbx64_register_map): Ditto.
5542 (svr4_dbx_register_map): Ditto.
5543 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5544 disabled.
5545 (ix86_preferred_reload_class): Disable constants for mask registers.
5546 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5547 (ix86_hard_regno_mode_ok): Support new mask registers.
5548 (x86_order_regs_for_local_alloc): Ditto.
5549 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5550 (FIXED_REGISTERS): Add new mask registers.
5551 (CALL_USED_REGISTERS): Ditto.
5552 (REG_ALLOC_ORDER): Ditto.
5553 (VALID_MASK_REG_MODE): New.
5554 (FIRST_MASK_REG): Ditto.
5555 (LAST_MASK_REG): Ditto.
5556 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5557 (MAYBE_MASK_CLASS_P): New.
5558 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5559 (REG_CLASS_CONTENTS): Ditto.
5560 (MASK_REGNO_P): New.
5561 (ANY_MASK_REG_P): Ditto.
5562 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM		 5563 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5564 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5565 mask registers.
85a77221
AI		 5566 (attribute "type"): Add mskmov, msklog.
5567 (attribute "length_immediate"): Support them.
5568 (attribute "memory"): Ditto.
5569 (attribute "prefix_0f"): Ditto.
5570 (*movhi_internal): Support new mask registers.
5571 (*movqi_internal): Ditto.
5572 (define_split): Split out clobber pattern is a logic
5573 insn on mask registers.
5574 (*k<logic><mode>): New.
5575 (*andhi_1): Extend to support mask regs.
5576 (*andqi_1): Extend to support mask regs.
5577 (kandn<mode>): New.
5578 (define_split): Split and-not to and and not if operands
5579 are not mask regs.
5580 (*<code><mode>_1): Separate HI mode to new pattern...
5581 (*<code>hi_1): This.
5582 (*<code>qi_1): Extend to support mask regs.
5583 (kxnor<mode>): New.
5584 (kortestzhi): Ditto.
5585 (kortestchi): Ditto.
5586 (kunpckhi): Ditto.
5587 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5588 (*one_cmplhi2_1): ...Here, now with mask registers support.
5589 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM		 5590 (HI/QImode arithmetics splitter): Don't split if mask registers
5591 are used.
85a77221
AI		 5592 (HI/QImode not splitter): Ditto.
5593 * config/i386/predicated.md (mask_reg_operand): New.
5594 (general_reg_operand): Ditto.
5595
7e570821
AI		 55962013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5597
5598 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5599 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5600
9e1376e9
JL		 56012013-09-10 Jeff Law <law@redhat.com>
5602
5603 PR tree-optimization/58380
5604 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5605 of threading through a buried loop header.
5606
5607 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5608 return value for single successor case.
5609
5bbcb888
JH		 56102013-09-10 Jan Hubicka <jh@suse.cz>
5611
5612 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5613
d742ff4b
RE		 56142013-09-10 Richard Earnshaw <rearnsha@arm.com>
5615
5616 PR target/58361
5617 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5618 support conditional execution.
5619 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5620
b0681c9e
VM		 56212013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5622
5623 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5624 constraint pass.
f77e2d2d
JBG		 5625 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5626 Check destination too to check move insn.
b0681c9e 5627 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG		 5628 changed its allocation and the optional reload was inherited on last
5629 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM		 5630 (lra_undo_inheritance): Add check that inherited pseudo still in
5631 memory.
5632
c9b1c957
JG		 56332013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5634
5635 * config/aarch64/aarch64.md (generic_sched): New.
5636 * config/aarch64/aarch64-generic.md (load): Make conditional
5637 on generic_sched attribute.
5638 (nonload): Likewise.
5639
82d618d3
JH		 56402013-09-10 Jan Hubicka <jh@suse.cz>
5641
5642 * lto-cgraph.c: Include ipa-utils.h.
5643 (compute_ltrans_boundary): Also add possible targets into the boundary.
5644
668b6894
JH		 56452013-09-10 Jan Hubicka <jh@suse.cz>
5646
5647 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5648 VAR_DECL of vtable rather than full expression.
5649
4f90d3e0
JH		 56502013-09-10 Jan Hubicka <jh@suse.cz>
5651 Paolo Carlini <paolo.carlini@oracle.com>
5652
5653 * cgraphunit.c (analyze_functions): Save input_location, set it
5654 to UNKNOWN_LOCATION and restore it at the end.
5655
39e87baf
MJ		 56562013-09-10 Martin Jambor <mjambor@suse.cz>
5657
5658 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5659 represented by a thunk.
5660
581aedec
JL		 56612013-09-10 Jeff Law <law@redhat.com>
5662
5663 PR tree-optimization/58343
5664 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5665 jump threading requests through loop headers buried in the middle
5666 of a jump threading path.
5667
5668 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5669 in return value/type.
5670
84cf4ab6
JJ		 56712013-09-10 Jakub Jelinek <jakub@redhat.com>
5672
5673 PR rtl-optimization/58365
5674 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5675 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5676 it differs.
5677
447dd906
RB		 56782013-09-10 Richard Biener <rguenther@suse.de>
5679
5680 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5681 * tree-data-ref.c (create_rdg_vertices): Collect all data
5682 references, signal failure to the caller, use data-ref API.
893e85fa
CM		 5683 (build_rdg): Compute data references only once. Maintain lifetime
5684 of data references and data dependences from within RDG.
447dd906
RB		 5685 (free_rdg): Free dependence relations.
5686 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5687 inventing extra dependences.
5688 (distribute_loop): Update for RDG API changes.
5689
f14a7324
KT		 56902013-09-10 Kai Tietz <ktietz@redhat.com>
5691
5692 * doc/invoke.texi (fms-extensions): Document changed
5693 behavior for ms-abi targets.
5694 * config/i386/i386.c (ix86_option_override_internal):
5695 Set default value of option -fms-extension for ms-abi targets.
5696
ff131f26
MZ		 56972013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5698
5699 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5700
33115a2c
AM		 57012013-09-10 Alan Modra <amodra@gmail.com>
5702
5703 PR target/58330
5704 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5705
b6872a13
AM		 57062013-09-10 Alan Modra <amodra@gmail.com>
5707
5708 * config/rs6000/predicates.md (add_cint_operand): New.
5709 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5710 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5711 using add_cint_operand.
5712 (largetoc_high_plus_aix): Likewise.
5713
28fd0ba2
JJ		 57142013-09-09 Jakub Jelinek <jakub@redhat.com>
5715
5716 PR tree-optimization/58364
5717 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5718 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5719 the current range can't be an unconditional true or false.
5720
552d99fe
JG		 57212013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5722
5723 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5724
f43245d1
UB		 57252013-09-09 Uros Bizjak <ubizjak@gmail.com>
5726
5727 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5728
c9b0866a
PC		 57292013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5730
5731 PR c++/43452
5732 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5733
78d8b9f0
IB		 57342013-09-09 Ian Bolton <ian.bolton@arm.com>
5735
5736 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5737 NO_REGS for immediate that can't be moved directly into FP_REGS.
5738
274b2532
KT		 57392013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5740
5741 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5742 comparison with negated operand.
c9b0866a
PC		 5743 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5744 RTL form.
274b2532 5745
467a3558
RB		 57462013-09-09 Richard Biener <rguenther@suse.de>
5747
5748 PR middle-end/58326
5749 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5750 of a subloop record all its block as affecting loop-closed SSA form.
5751
088c5368
RS		 57522013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5753
5754 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5755 of an rtx/bitpos pair.
5756 (store_fixed_bit_field): Update accordingly.
5757
69db2d57
RS		 57582013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5759
5760 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5761 GEN_INT.
5762 * builtins.c (expand_errno_check): Likewise.
5763 * dwarf2cfi.c (init_return_column_size): Likewise.
5764 * except.c (sjlj_mark_call_sites): Likewise.
5765 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5766 * lra-constraints.c (emit_inc): Likewise.
5767 * ree.c (combine_set_extension): Likewise.
5768 * regmove.c (fixup_match_2): Likewise.
5769 * reload1.c (inc_for_reload): Likewise.
5770
69a59f0f
RS		 57712013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5772
5773 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5774 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5775 Use gen_int_mode with the mode of the associated simplify_* call.
5776 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5777 * expmed.c (expand_shift_1): Likewise.
5778 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5779 * loop-iv.c (iv_number_of_iterations): Likewise.
5780 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5781 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5782
2f1cd2eb
RS		 57832013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5784
5785 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5786 of the associated expand_* call.
5787 (asan_emit_stack_protection): Likewise.
5788 * builtins.c (round_trampoline_addr): Likewise.
5789 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5790 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5791 (emit_store_flag): Likewise.
5792 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5793 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5794 Likewise.
5795 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5796 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5797 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5798 * modulo-sched.c (generate_prolog_epilog): Likewise.
5799 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5800 (expand_ctz, expand_ffs, expand_unop): Likewise.
5801
4789c0ce
RS		 58022013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5803
5804 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5805 of the associated gen_rtx_* call.
5806 * caller-save.c (init_caller_save): Likewise.
5807 * combine.c (find_split_point, make_extraction): Likewise.
5808 (make_compound_operation): Likewise.
5809 * dwarf2out.c (mem_loc_descriptor): Likewise.
5810 * explow.c (plus_constant, probe_stack_range): Likewise.
5811 * expmed.c (expand_mult_const): Likewise.
5812 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5813 * reload1.c (init_reload): Likewise.
5814 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5815 * var-tracking.c (adjust_mems): Likewise.
5816 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5817 rather than gen_rtx_fmt_ee.
5818
a811f7cb
JH		 58192013-09-09 Jan Hubicka <jh@suse.cz>
5820
5821 PR middle-end/58294
5822 * value-prof.c (gimple_ic): Copy also abnormal edges.
5823
dcad1dd3
RS		 58242013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5825
5826 * asan.c (asan_shadow_cst): Use gen_int_mode.
5827
daf5c770
JH		 58282013-09-08 Jan Hubicka <jh@suse.cz>
5829
5830 * ipa-profile.c: Add toplevel comment.
5831 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5832 (contains_hot_call_p): New function.
5833 (ipa_propagate_frequency): Set frequencies based on counts when
5834 profile is read.
5835 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5836 profile; do not tamper with profile after inlining if it is read.
5837
a12cd2db
JH		 58382013-09-08 Jan Hubicka <jh@suse.cz>
5839
5840 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5841 speculative edges.
5842
856a1588
JH		 58432013-09-08 Jan Hubicka <jh@suse.cz>
5844
5845 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5846 summary generation.
5847
4f4b0b73
JL		 58482013-09-08 Jeff Law <law@redhat.com>
5849
5850 PR bootstrap/58340
5851 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5852 of 'found'.
5853
cc4029ee
AK		 58542013-09-08 Andi Kleen <ak@linux.intel.com>
5855
5856 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5857
7b395ddd
JH		 58582013-09-08 Jan Hubicka <jh@suse.cz>
5859
5860 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5861
c34c46dd
RS		 58622013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5863
5864 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5865 for non-debug insns.
5866 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5867 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5868 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5869 accordingly.
5870
3462aa02
JH		 58712013-09-08 Jan Hubicka <jh@suse.cz>
5872
5873 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5874 targets and devirtualize to BUILT_IN_UNREACHABLE.
5875 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5876 * ipa.c (walk_polymorphic_call_targets): New function.
5877 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5878 functions; use the new timevar.
5879 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5880 was removed from the program.
5881 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5882 not consider it in the walk when its vtable is dead.
5883 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5884 record_binfo.
5885 (devirt_variable_node_removal_hook): New function.
5886 (possible_polymorphic_call_targets): Also register
5887 devirt_variable_node_removal_hook.
5888 (ipa_devirt): Do not do non-speculative devirtualization.
5889 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5890
26e5b0fd
JH		 58912013-09-08 Jan Hubicka <jh@suse.cz>
5892
5893 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5894 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5895 varpool_remove_variable_insertion_hook): Declare.
5896 * varpool.c (varpool_node_hook_list): New structure.
5897 (first_varpool_node_removal_hook,
5898 first_varpool_variable_insertion_hook): New variables.
5899 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5900 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5901 varpool_remove_variable_insertion_hook,
5902 varpool_call_variable_insertion_hooks): New functions.
5903 (varpool_remove_node): Use it.
5904
97f1f314 59052013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 5906
97f1f314
PC		 5907 PR c++/54941
5908 * diagnostic.c (diagnostic_build_prefix): When s.file is
5909 "<built-in>" don't output line and column numbers.
6df2ca09 5910
bcb650cb
JH		 59112013-09-06 Jan Hubicka <jh@suse.cz>
5912
5913 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5914 * lto-streamer-out.c: Stream thunks, too.
5915 * lto-streamer-in.c (input_function): Pop cfun here
5916 (lto_read_body): Instead of here.
5917
87f3fea8
CT		 59182013-09-06 Caroline Tice <cmtice@google.com>
5919
5920 * doc/install.texi: Add documentation for the --enable-vtable-verify
5921 and the --disable-libvtv configure options.
5922
5e8d9b45
JL		 59232013-09-06 Jeff Law <law@redhat.com>
5924
5925 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5926 edge implied equivalences into successor phis.
5927
3a001aff
JR		 59282013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5929
5930 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5931
59322013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 5933
5934 * resource.c (mark_target_live_regs): Compute resources taking
5935 into account if a call is predicated or not.
5936
2c641f8b
EB		 59372013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5938
5939 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5940 the compiler in the function names.
5941
a81b0a3d
JH		 59422013-09-06 Jan Hubicka <jh@suse.cz>
5943
5944 PR middle-end/58094
5945 * ipa-inline.c (has_caller_p): New function.
5946 (want_inline_function_to_all_callers_p): Use it.
5947 (sum_callers, inline_to_all_callers): Break out from ...
5948 (ipa_inline): ... here.
5949
68e4ca7e
JH		 59502013-09-06 Jan Hubicka <jh@suse.cz>
5951
2c641f8b
EB		 5952 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5953 only when AVX is enabled.
68e4ca7e 5954
3c73e2fe
JG		 59552013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5956
5957 * config/aarch64/aarch64.md
5958 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5959 is fpsimd_<load/store>2.
5960 (load_pair<mode>): Likewise.
5961 (store_pair<mode>): Likewise.
5962
8c48eecd
JG		 59632013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5964
5965 * config/arm/types.md (type): Add "mrs" type.
5966 * config/aarch64/aarch64.md
5967 (aarch64_load_tp_hard): Make type "mrs".
5968 * config/arm/arm.md
5969 (load_tp_hard): Make type "mrs".
5970 * config/arm/cortex-a15.md: Update with new attributes.
5971 * config/arm/cortex-a5.md: Update with new attributes.
5972 * config/arm/cortex-a53.md: Update with new attributes.
5973 * config/arm/cortex-a7.md: Update with new attributes.
5974 * config/arm/cortex-a8.md: Update with new attributes.
5975 * config/arm/cortex-a9.md: Update with new attributes.
5976 * config/arm/cortex-m4.md: Update with new attributes.
5977 * config/arm/cortex-r4.md: Update with new attributes.
5978 * config/arm/fa526.md: Update with new attributes.
5979 * config/arm/fa606te.md: Update with new attributes.
5980 * config/arm/fa626te.md: Update with new attributes.
5981 * config/arm/fa726te.md: Update with new attributes.
5982
9ea187a5
JG		 59832013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5984
5985 * config/aarch64/aarch64.md
5986 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5987 (*movtf_aarch64): Likewise.
5988 * config/arm/arm.md
5989 (thumb1_movdi_insn): Use "multiple" for type where more than one
5990 instruction is used for a move.
5991 (*arm32_movhf): Likewise.
5992 (*thumb_movdf_insn): Likewise.
5993
292b89b3
JG		 59942013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5995
5996 * config/arm/types.md (type): Rename fcpys to fmov.
5997 * config/arm/vfp.md
5998 (*arm_movsi_vfp): Rename type fcpys as fmov.
5999 (*thumb2_movsi_vfp): Likewise
6000 (*movhf_vfp_neon): Likewise
6001 (*movhf_vfp): Likewise
6002 (*movsf_vfp): Likewise
6003 (*thumb2_movsf_vfp): Likewise
6004 (*movsfcc_vfp): Likewise
6005 (*thumb2_movsfcc_vfp): Likewise
6006 * config/aarch64/aarch64-simd.md
6007 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6008 * config/aarch64/aarch64.md
6009 (*movsi_aarch64): Replace type mov_reg with fmovs.
6010 (*movdi_aarch64): Likewise
6011 (*movsf_aarch64): Likewise
6012 (*movdf_aarch64): Likewise
6013 * config/arm/arm.c
6014 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6015 * config/arm/iwmmxt.md
6016 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6017 * config/arm/arm1020e.md: Update with new attributes.
6018 * config/arm/cortex-a15-neon.md: Update with new attributes.
6019 * config/arm/cortex-a5.md: Update with new attributes.
6020 * config/arm/cortex-a53.md: Update with new attributes.
6021 * config/arm/cortex-a7.md: Update with new attributes.
6022 * config/arm/cortex-a8-neon.md: Update with new attributes.
6023 * config/arm/cortex-a9.md: Update with new attributes.
6024 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6025 * config/arm/cortex-r4f.md: Update with new attributes.
6026 * config/arm/marvell-pj4.md: Update with new attributes.
6027 * config/arm/vfp11.md: Update with new attributes.
6028
6a24a83e
JG		 60292013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6030
6031 * config/aarch64/aarch64.md
6032 (*madd<mode>): Fix type attribute.
6033 (*maddsi_uxtw): Likewise.
6034 (*msub<mode>): Likewise.
6035 (*msubsi_uxtw): Likewise.
6036 (<su_optab>maddsidi4): Likewise.
6037 (<su_optab>msubsidi4): Likewise.
6038
b86923f0
JG		 60392013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6040
6041 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6042 * config/arm/arm.md (core_cycles): Remove fdiv.
6043 * config/arm/vfp.md:
6044 (*sqrtsf2_vfp): Update for attribute changes.
6045 (*sqrtdf2_vfp): Likewise.
6046 * config/aarch64/aarch64.md:
6047 (sqrt<mode>2): Update for attribute changes.
6048 * config/arm/arm1020e.md: Update with new attributes.
6049 * config/arm/cortex-a15-neon.md: Update with new attributes.
6050 * config/arm/cortex-a5.md: Update with new attributes.
6051 * config/arm/cortex-a53.md: Update with new attributes.
6052 * config/arm/cortex-a7.md: Update with new attributes.
6053 * config/arm/cortex-a8-neon.md: Update with new attributes.
6054 * config/arm/cortex-a9.md: Update with new attributes.
6055 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6056 * config/arm/cortex-r4f.md: Update with new attributes.
6057 * config/arm/marvell-pj4.md: Update with new attributes.
6058 * config/arm/vfp11.md: Update with new attributes.
6059
7b49c9e1
JG		 60602013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6061
6062 * config/arm/types.md
6063 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6064 * config/aarch64/aarch64.md
6065 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6066 new attributes.
6067 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6068 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6069 (float<GPI:mode><GPF:mode>2): Likewise.
6070 * config/arm/vfp.md
6071 (*truncsisf2_vfp): Update with new attributes.
6072 (*truncsidf2_vfp): Likewise.
6073 (fixuns_truncsfsi2): Likewise.
6074 (fixuns_truncdfsi2): Likewise.
6075 (*floatsisf2_vfp): Likewise.
6076 (*floatsidf2_vfp): Likewise.
6077 (floatunssisf2): Likewise.
6078 (floatunssidf2): Likewise.
6079 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6080 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6081 * config/arm/arm1020e.md: Update with new attributes.
6082 * config/arm/cortex-a15-neon.md: Update with new attributes.
6083 * config/arm/cortex-a5.md: Update with new attributes.
6084 * config/arm/cortex-a53.md: Update with new attributes.
6085 * config/arm/cortex-a7.md: Update with new attributes.
6086 * config/arm/cortex-a8-neon.md: Update with new attributes.
6087 * config/arm/cortex-a9.md: Update with new attributes.
6088 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6089 * config/arm/cortex-r4f.md: Update with new attributes.
6090 * config/arm/marvell-pj4.md: Update with new attributes.
6091 * config/arm/vfp11.md: Update with new attributes.
6092
e7df8af8
JG		 60932013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6094
6095 * config/aarch64/arm_neon.h
6096 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6097 (vqtbx<1,2,3,4><q>_s8): Likewise.
6098
594726e4
JG		 60992013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6100
f43245d1
UB		 6101 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6102 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG		 6103 (add<mode>3): Add type attribute.
6104 (add<mode>3): Likewise.
6105 (usadd<mode>3): Likewise.
6106 (ssadd<mode>3): Likewise.
6107 (sub<mode>3): Likewise.
6108 (sub<mode>3): Likewise.
6109 (ussub<mode>3): Likewise.
6110 (sssub<mode>3): Likewise.
6111 (ssmulsa3): Likewise.
6112 (usmulusa3): Likewise.
6113 (arm_usatsihi): Likewise.
6114 * config/arm/vfp.md
6115 (*movdi_vfp): Add types for all instructions.
6116 (*movdi_vfp_cortexa8): Likewise.
6117 (*movhf_vfp_neon): Likewise.
6118 (*movhf_vfp): Likewise.
6119 (*movdf_vfp): Likewise.
6120 (*thumb2_movdf_vfp): Likewise.
6121 (*thumb2_movdfcc_vfp): Likewise.
6122 * config/arm/arm.md: Add type attribute to all insn patterns.
6123 (*thumb1_adddi3): Add type attribute.
6124 (*arm_adddi3): Likewise.
6125 (*adddi_sesidi_di): Likewise.
6126 (*adddi_zesidi_di): Likewise.
6127 (*thumb1_addsi3): Likewise.
6128 (addsi3_compare0): Likewise.
6129 (*addsi3_compare0_scratch): Likewise.
6130 (*compare_negsi_si): Likewise.
6131 (cmpsi2_addneg): Likewise.
6132 (*addsi3_carryin_<optab>): Likewise.
6133 (*addsi3_carryin_alt2_<optab>): Likewise.
6134 (*addsi3_carryin_clobercc_<optab>): Likewise.
6135 (*subsi3_carryin): Likewise.
6136 (*subsi3_carryin_const): Likewise.
6137 (*subsi3_carryin_compare): Likewise.
6138 (*subsi3_carryin_compare_const): Likewise.
6139 (*arm_subdi3): Likewise.
6140 (*thumb_subdi3): Likewise.
6141 (*subdi_di_zesidi): Likewise.
6142 (*subdi_di_sesidi): Likewise.
6143 (*subdi_zesidi_di): Likewise.
6144 (*subdi_sesidi_di): Likewise.
6145 (*subdi_zesidi_ze): Likewise.
6146 (thumb1_subsi3_insn): Likewise.
6147 (*arm_subsi3_insn): Likewise.
6148 (*anddi3_insn): Likewise.
6149 (*anddi_zesidi_di): Likewise.
6150 (*anddi_sesdi_di): Likewise.
6151 (*ne_zeroextracts): Likewise.
6152 (*ne_zeroextracts): Likewise.
6153 (*ite_ne_zeroextr): Likewise.
6154 (*ite_ne_zeroextr): Likewise.
6155 (*anddi_notdi_di): Likewise.
6156 (*anddi_notzesidi): Likewise.
6157 (*anddi_notsesidi): Likewise.
6158 (andsi_notsi_si): Likewise.
6159 (thumb1_bicsi3): Likewise.
6160 (*iordi3_insn): Likewise.
6161 (*iordi_zesidi_di): Likewise.
6162 (*iordi_sesidi_di): Likewise.
6163 (*thumb1_iorsi3_insn): Likewise.
6164 (*xordi3_insn): Likewise.
6165 (*xordi_zesidi_di): Likewise.
6166 (*xordi_sesidi_di): Likewise.
6167 (*arm_xorsi3): Likewise.
6168 (*andsi_iorsi3_no): Likewise.
6169 (*smax_0): Likewise.
6170 (*smax_m1): Likewise.
6171 (*arm_smax_insn): Likewise.
6172 (*smin_0): Likewise.
6173 (*arm_smin_insn): Likewise.
6174 (*arm_umaxsi3): Likewise.
6175 (*arm_uminsi3): Likewise.
6176 (*minmax_arithsi): Likewise.
6177 (*minmax_arithsi_): Likewise.
6178 (*satsi_<SAT:code>): Likewise.
6179 (arm_ashldi3_1bit): Likewise.
6180 (arm_ashrdi3_1bit): Likewise.
6181 (arm_lshrdi3_1bit): Likewise.
6182 (*arm_negdi2): Likewise.
6183 (*thumb1_negdi2): Likewise.
6184 (*arm_negsi2): Likewise.
6185 (*thumb1_negsi2): Likewise.
6186 (*negdi_extendsid): Likewise.
6187 (*negdi_zero_extend): Likewise.
6188 (*arm_abssi2): Likewise.
6189 (*thumb1_abssi2): Likewise.
6190 (*arm_neg_abssi2): Likewise.
6191 (*thumb1_neg_abss): Likewise.
6192 (one_cmpldi2): Likewise.
6193 (extend<mode>di2): Likewise.
6194 (*compareqi_eq0): Likewise.
6195 (*arm_extendhisi2addsi): Likewise.
6196 (*arm_movdi): Likewise.
6197 (*thumb1_movdi_insn): Likewise.
6198 (*arm_movt): Likewise.
6199 (*thumb1_movsi_insn): Likewise.
6200 (pic_add_dot_plus_four): Likewise.
6201 (pic_add_dot_plus_eight): Likewise.
6202 (tls_load_dot_plus_eight): Likewise.
6203 (*thumb1_movhi_insn): Likewise.
6204 (*thumb1_movsf_insn): Likewise.
6205 (*movdf_soft_insn): Likewise.
6206 (*thumb_movdf_insn): Likewise.
6207 (cbranchsi4_insn): Likewise.
6208 (cbranchsi4_scratch): Likewise.
6209 (*negated_cbranchsi4): Likewise.
6210 (*tbit_cbranch): Likewise.
6211 (*tlobits_cbranch): Likewise.
6212 (*tstsi3_cbranch): Likewise.
6213 (*cbranchne_decr1): Likewise.
6214 (*addsi3_cbranch): Likewise.
6215 (*addsi3_cbranch_scratch): Likewise.
6216 (*arm_cmpdi_insn): Likewise.
6217 (*arm_cmpdi_unsig): Likewise.
6218 (*arm_cmpdi_zero): Likewise.
6219 (*thumb_cmpdi_zero): Likewise.
6220 (*deleted_compare): Likewise.
6221 (*mov_scc): Likewise.
6222 (*mov_negscc): Likewise.
6223 (*mov_notscc): Likewise.
6224 (*cstoresi_eq0_thumb1_insn): Likewise.
6225 (cstoresi_nltu_thumb1): Likewise.
6226 (cstoresi_ltu_thu): Likewise.
6227 (thumb1_addsi3_addgeu): Likewise.
6228 (*arm_jump): Likewise.
6229 (*thumb_jump): Likewise.
6230 (*check_arch2): Likewise.
6231 (arm_casesi_internal): Likewise.
6232 (thumb1_casesi_dispatch): Likewise.
6233 (*arm_indirect_jump): Likewise.
6234 (*thumb1_indirect_jump): Likewise.
6235 (nop): Likewise.
6236 (*and_scc): Likewise.
6237 (*ior_scc): Likewise.
6238 (*compare_scc): Likewise.
6239 (*cond_move): Likewise.
6240 (*cond_arith): Likewise.
6241 (*cond_sub): Likewise.
6242 (*cmp_ite0): Likewise.
6243 (*cmp_ite1): Likewise.
6244 (*cmp_and): Likewise.
6245 (*cmp_ior): Likewise.
6246 (*ior_scc_scc): Likewise.
6247 (*ior_scc_scc_cmp): Likewise.
6248 (*and_scc_scc): Likewise.
6249 (*and_scc_scc_cmp): Likewise.
6250 (*and_scc_scc_nod): Likewise.
6251 (*negscc): Likewise.
6252 (movcond_addsi): Likewise.
6253 (movcond): Likewise.
6254 (*ifcompare_plus_move): Likewise.
6255 (*if_plus_move): Likewise.
6256 (*ifcompare_move_plus): Likewise.
6257 (*if_move_plus): Likewise.
6258 (*ifcompare_arith_arith): Likewise.
6259 (*if_arith_arith): Likewise.
6260 (*ifcompare_arith_move): Likewise.
6261 (*if_arith_move): Likewise.
6262 (*ifcompare_move_arith): Likewise.
6263 (*if_move_arith): Likewise.
6264 (*ifcompare_move_not): Likewise.
6265 (*if_move_not): Likewise.
6266 (*ifcompare_not_move): Likewise.
6267 (*if_not_move): Likewise.
6268 (*ifcompare_shift_move): Likewise.
6269 (*if_shift_move): Likewise.
6270 (*ifcompare_move_shift): Likewise.
6271 (*if_move_shift): Likewise.
6272 (*ifcompare_shift_shift): Likewise.
6273 (*ifcompare_not_arith): Likewise.
6274 (*ifcompare_arith_not): Likewise.
6275 (*if_arith_not): Likewise.
6276 (*ifcompare_neg_move): Likewise.
6277 (*if_neg_move): Likewise.
6278 (*ifcompare_move_neg): Likewise.
6279 (*if_move_neg): Likewise.
6280 (prologue_thumb1_interwork): Likewise.
6281 (*cond_move_not): Likewise.
6282 (*sign_extract_onebit): Likewise.
6283 (*not_signextract_onebit): Likewise.
6284 (stack_tie): Likewise.
6285 (align_4): Likewise.
6286 (align_8): Likewise.
6287 (consttable_end): Likewise.
6288 (consttable_1): Likewise.
6289 (consttable_2): Likewise.
6290 (consttable_4): Likewise.
6291 (consttable_8): Likewise.
6292 (consttable_16): Likewise.
6293 (*thumb1_tablejump): Likewise.
6294 (prefetch): Likewise.
6295 (force_register_use): Likewise.
6296 (thumb_eh_return): Likewise.
6297 (load_tp_hard): Likewise.
6298 (load_tp_soft): Likewise.
6299 (tlscall): Likewise.
6300 (*arm_movtas_ze): Likewise.
6301 (*arm_rev): Likewise.
6302 (*arm_revsh): Likewise.
6303 (*arm_rev16): Likewise.
6304 * config/arm/thumb2.md
6305 (*thumb2_smaxsi3): Likewise.
6306 (*thumb2_sminsi3): Likewise.
6307 (*thumb32_umaxsi3): Likewise.
6308 (*thumb2_uminsi3): Likewise.
6309 (*thumb2_negdi2): Likewise.
6310 (*thumb2_abssi2): Likewise.
6311 (*thumb2_neg_abss): Likewise.
6312 (*thumb2_movsi_insn): Likewise.
6313 (tls_load_dot_plus_four): Likewise.
6314 (*thumb2_movhi_insn): Likewise.
6315 (*thumb2_mov_scc): Likewise.
6316 (*thumb2_mov_negs): Likewise.
6317 (*thumb2_mov_negs): Likewise.
6318 (*thumb2_mov_nots): Likewise.
6319 (*thumb2_mov_nots): Likewise.
6320 (*thumb2_movsicc_): Likewise.
6321 (*thumb2_movsfcc_soft_insn): Likewise.
6322 (*thumb2_indirect_jump): Likewise.
6323 (*thumb2_and_scc): Likewise.
6324 (*thumb2_ior_scc): Likewise.
6325 (*thumb2_ior_scc_strict_it): Likewise.
6326 (*thumb2_cond_move): Likewise.
6327 (*thumb2_cond_arith): Likewise.
6328 (*thumb2_cond_ari): Likewise.
6329 (*thumb2_cond_sub): Likewise.
6330 (*thumb2_negscc): Likewise.
6331 (*thumb2_movcond): Likewise.
6332 (thumb2_casesi_internal): Likewise.
6333 (thumb2_casesi_internal_pic): Likewise.
6334 (*thumb2_alusi3_short): Likewise.
6335 (*thumb2_mov<mode>_shortim): Likewise.
6336 (*thumb2_addsi_short): Likewise.
6337 (*thumb2_subsi_short): Likewise.
6338 (thumb2_addsi3_compare0): Likewise.
6339 (*thumb2_cbz): Likewise.
6340 (*thumb2_cbnz): Likewise.
6341 (*thumb2_one_cmplsi2_short): Likewise.
6342 (*thumb2_negsi2_short): Likewise.
6343 (*orsi_notsi_si): Likewise.
6344 * config/arm/arm1020e.md: Update with new attributes.
6345 * config/arm/arm1026ejs.md: Update with new attributes.
6346 * config/arm/arm1136jfs.md: Update with new attributes.
6347 * config/arm/arm926ejs.md: Update with new attributes.
6348 * config/arm/cortex-a15.md: Update with new attributes.
6349 * config/arm/cortex-a5.md: Update with new attributes.
6350 * config/arm/cortex-a53.md: Update with new attributes.
6351 * config/arm/cortex-a7.md: Update with new attributes.
6352 * config/arm/cortex-a8.md: Update with new attributes.
6353 * config/arm/cortex-a9.md: Update with new attributes.
6354 * config/arm/cortex-m4.md: Update with new attributes.
6355 * config/arm/cortex-r4.md: Update with new attributes.
6356 * config/arm/fa526.md: Update with new attributes.
6357 * config/arm/fa606te.md: Update with new attributes.
6358 * config/arm/fa626te.md: Update with new attributes.
6359 * config/arm/fa726te.md: Update with new attributes.
6360
1c83b673
JG		 63612013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6362
6363 * config/aarch64/aarch64-simd.md
6364 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6365 <vwx> iterator to ensure correct register choice.
6366 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6367 (aarch64_sqdmull_n<mode>): Likewise.
6368 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6369 * config/aarch64/arm_neon.h
6370 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6371 (vml<as><q>_n_<su>16): Likewise.
6372 (vml<as>l_high_lane<q>_<su>16): Likewise.
6373 (vml<as>l_high_n_<su>16): Likewise.
6374 (vml<as>l_lane<q>_<su>16): Likewise.
6375 (vml<as>l_n_<su>16): Likewise.
6376 (vmul<q>_lane<q>_<su>16): Likewise.
6377 (vmul<q>_n_<su>16): Likewise.
6378 (vmull_lane<q>_<su>16): Likewise.
6379 (vmull_n_<su>16): Likewise.
6380 (vmull_high_lane<q>_<su>16): Likewise.
6381 (vmull_high_n_<su>16): Likewise.
6382 (vqrdmulh<q>_n_s16): Likewise.
6383
bb1ae543
TB		 63842013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6385
6386 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6387 have the correct lane parameter.
6388
f23c0742
RB		 63892013-09-06 Richard Biener <rguenther@suse.de>
6390
6391 * cfganal.c (control_dependences::~control_dependences):
6392 Properly free all of the vector.
6393
c1d49770
KY		 63942013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6395
6396 PR target/58269
6397 * config/i386/i386.c (ix86_conditional_register_usage):
6398 Proper initialize extended SSE registers.
6399
6da47f52
JH		 64002013-09-06 Jan Hubicka <jh@suse.cz>
6401
6402 PR tree-optimization/58311
6403 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6404
64052013-09-06 Jan Hubicka <jh@suse.cz>
6406
6407 * Makefile.in (tree-sra.o): Update dependencies.
6408 * tree-sra.c: Include ipa-utils.h
6409 (scan_function): Use recursive_call_p.
6410 (has_caller_p): New function.
6411 (cgraph_for_node_and_aliases): Count also callers of aliases.
6412
fc11f321
JH		 64132013-09-06 Jan Hubicka <jh@suse.cz>
6414
6415 PR middle-end/58094
6416 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6417 * tree-tailcall.c: Include ipa-utils.h.
6418 (find_tail_calls): Use it.
6419 * ipa-pure-const.c (check_call): Likewise.
6420 * ipa-utils.c (recursive_call_p): New function.
6421 * ipa-utils.h (recursive_call_p): Dclare.
6422 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6423 (symtab_semantically_equivalent_p): New function.
6424 * Makefile.in (tree-tailcall.o): Update dependencies.
6425
9a6af450
EB		 64262013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6427
6428 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6429 non-inlinable part.
6430
35f32ad4
RB		 64312013-09-06 Richard Biener <rguenther@suse.de>
6432
6433 * lto-streamer.h (lto_global_var_decls): Remove.
6434 * Makefile.in (OBJS): Remove lto-symtab.o.
6435 (lto-symtab.o): Remove.
6436 (GTFILES): Remove lto-symtab.c
6437 * lto-symtab.c: Move to lto/
6438
d12a76f3
AK		 64392013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6440
6441 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6442 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6443 (UNSPEC_FPINT_RINT): New constant definitions.
6444 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6445 definition with 2 attributes.
6446 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6447 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6448 definitions.
6449
64502013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6451
f43245d1 6452 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK		 6453 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6454 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6455 attribute to "z196_alone".
6456 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6457 "zEC12_simple".
6458
c8e9d8c3
RB		 64592013-09-06 Richard Biener <rguenther@suse.de>
6460
6461 * basic-block.h (class control_dependences): New.
6462 * tree-ssa-dce.c (control_dependence_map): Remove.
6463 (cd): New global.
6464 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6465 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6466 find_pdom, find_control_dependence, find_all_control_dependences):
6467 Move to cfganal.c.
f43245d1
UB		 6468 (mark_control_dependent_edges_necessary,
6469 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6470 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB		 6471 * cfganal.c (set_control_dependence_map_bit,
6472 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6473 find_all_control_dependences): Move from tree-ssa-dce.c and
6474 implement as methods of control_dependences class.
6475 (control_dependences::control_dependences): New.
6476 (control_dependences::~control_dependences): Likewise.
6477 (control_dependences::get_edges_dependent_on): Likewise.
6478 (control_dependences::get_edge): Likewise.
6479
7d2ba471
JH		 64802013-09-04 Jan Hubicka <jh@suse.cz>
6481
6482 * tree.c (types_same_for_odr): Drop overactive check.
6483 * ipa-devirt.c (hash_type_name): Likewise.
6484
e18412fc
JH		 64852013-09-04 Jan Hubicka <jh@suse.cz>
6486
6487 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6488 (analyze_functions): ... here.
6489
d352b245
JH		 64902013-09-04 Jan Hubicka <jh@suse.cz>
6491
6492 PR middle-end/58201
6493 * cgraphunit.c (analyze_functions): Clear AUX fields
6494 after processing; initialize assembler name has.
6495
770da076
JL		 64962013-09-05 Jeff Law <law@redhat.com>
6497
6498 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6499 from thread_around_empty_block. Record threading path into PATH.
6500 Recurse if threading through the initial block is successful.
6501 (thread_across_edge): Corresponding changes to slightly simplify.
6502
fe6f68e2
JG		 65032013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6504
6505 * config/aarch64/aarch64.md
6506 (type): Remove frecpe, frecps, frecpx.
6507 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6508 fix to be a TARGET_SIMD instruction.
6509 (aarch64_frecps): Remove.
6510 * config/aarch64/aarch64-simd.md
6511 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6512 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6513
6e4150e1
JG		 65142013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6515 Sofiane Naci <sofiane.naci@arm.com>
6516
f43245d1 6517 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6518 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB		 6519 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6520 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6521 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6522 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6523 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6524 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6525 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6526 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6527 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG		 6528 shift* types.
6529 * config/arm/arm-fixed.md: Update for attribute changes
6530 all occurrences of arlo_* types.
6531 * config/arm/thumb2.md: Update for attribute changes all occurrences
6532 of arlo_* types.
6533 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6534 (cortexa7_older_only): Likewise.
6535 (cortexa7_younger): Likewise.
6536 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6537 (1020alu_shift_op): Likewise.
6538 (1020alu_shift_reg_op): Likewise.
6539 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6540 (alu_shift_op): Likewise.
6541 (alu_shift_reg_op): Likewise.
f43245d1 6542 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG		 6543 (11_alu_shift_op): Likewise.
6544 (11_alu_shift_reg_op): Likewise.
6545 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6546 (9_alu_shift_reg_op): Likewise.
6547 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6548 attribute changes.
6549 (cortex_a15_alu_shift): Likewise.
6550 (cortex_a15_alu_shift_reg): Likewise.
6551 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6552 attribute changes.
6553 (cortex_a5_alu_shift): Likewise.
f43245d1
UB		 6554 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6555 attribute changes.
6e4150e1 6556 (cortex_a53_alu_shift): Likewise.
f43245d1
UB		 6557 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6558 attribute changes.
6e4150e1
JG		 6559 (cortex_a7_alu_reg): Likewise.
6560 (cortex_a7_alu_shift): Likewise.
f43245d1
UB		 6561 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6562 attribute changes.
6e4150e1
JG		 6563 (cortex_a8_alu_shift): Likewise.
6564 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6565 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6566 (cortex_a9_dp_shift): Likewise.
f43245d1
UB		 6567 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6568 attribute changes.
6e4150e1
JG		 6569 * config/arm/cortex-r4.md
6570 (cortex_r4_alu): Update for attribute changes.
6571 (cortex_r4_mov): Likewise.
6572 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6573 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6574 (526_alu_shift_op): Likewise.
f43245d1
UB		 6575 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6576 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6577 (626te_alu_shift_op): Likewise.
f43245d1 6578 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG		 6579 (726te_alu_shift_op): Likewise.
6580 (726te_alu_shift_reg_op): Likewise.
6581 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6582 (mp626_alu_shift_op): Likewise.
6583 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6584 (pj4_alu_conds): Likewise.
6585 (pj4_shift): Likewise.
6586 (pj4_shift_conds): Likewise.
6587 (pj4_alu_shift): Likewise.
6588 (pj4_alu_shift_conds): Likewise.
6589 * config/aarch64/aarch64.md: Update for attribute change
6590 all occurrences of arlo_* and shift* types.
6591
e6523306
MS		 65922013-09-05 Mike Stump <mikestump@comcast.net>
6593
6594 * tree.h: Move documentation for tree_function_decl to tree-core.h
6595 with the declaration.
6596
80a18161
PB		 65972013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6598
6599 PR target/58139
6600 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6601 looking for widest mode.
6602
6994430a
EB		 66032013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6604
6605 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6606
e67f39f7
RB		 66072013-09-05 Richard Biener <rguenther@suse.de>
6608
6609 PR tree-optimization/58137
6610 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6611 Do not create vectors of pointers.
6612 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6613 types for the components of the vector initializer.
6614 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6615 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6616
568cda29
MJ		 66172013-09-05 Martin Jambor <mjambor@suse.cz>
6618
6619 * ipa-prop.c (remove_described_reference): Accept missing references,
6620 return false if that hppens, otherwise return true.
6621 (cgraph_node_for_jfunc): New function.
6622 (try_decrement_rdesc_refcount): Likewise.
6623 (try_make_edge_direct_simple_call): Use them.
6624 (ipa_edge_removal_hook): Remove references from rdescs.
6625 (ipa_edge_duplication_hook): Clone rdescs and their references
6626 when the new edge has the same caller as the old one.
6627 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6628 reference before removing any edges.
6629
5a200acb
RE		 66302013-09-05 Richard Earnshaw <rearnsha@arm.com>
6631
6632 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6633 initial store.
6634 * thumb2.md (thumb2_storewb_parisi): New pattern.
6635
5922847b
YZ		 66362013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6637
6638 * config/aarch64/aarch64-option-extensions.def: Add
6639 AARCH64_OPT_EXTENSION of 'crc'.
6640 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6641 (AARCH64_ISA_CRC): Ditto.
6642 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6643 description of the CRC extension.
6644
996746aa
AI		 66452013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6646
6647 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6648 * config/rs6000/linux.h: Ditto.
6649 * alpha/linux.h: Ditto.
6650 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6651 no_c99_libc_has_function.
6652 * config/c6x/uclinux-elf.h: Ditto.
6653 * config/lm32/uclinux-elf.h: Ditto.
6654 * config/m68k/uclinux.h: Ditto.
6655 * config/moxie/uclinux.h: Ditto.
6656 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6657 (crisv32-*-linux*, cris-*-linux*): Ditto.
6658 * config/bfin/bfin.c: Include "tm_p.h".
6659
c370d99f
RB		 66602013-09-05 Richard Biener <rguenther@suse.de>
6661
6662 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6663 check for a definition without a basic-block.
6664
003bb7f3
JG		 66652013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6666 Sofiane Naci <sofiane.naci@arm.com>
6667
6668 * config/aarch64/aarch64.md
6669 (*movti_aarch64): Rename r_2_f and f_2_r.
6670 (*movsf_aarch64): Likewise.
6671 (*movdf_aarch64): Likewise.
6672 (*movtf_aarch64): Likewise.
6673 (aarch64_movdi_<mode>low): Likewise.
6674 (aarch64_movdi_<mode>high): Likewise.
6675 (aarch64_mov<mode>high_di): Likewise.
6676 (aarch64_mov<mode>low_di): Likewise.
6677 (aarch64_movtilow_tilow): Likewise.
6678 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6679 values to config/arm/types.md
6680 (attribute "conds"): Update for attribute change.
6681 (anddi3_insn): Likewise.
6682 (iordi3_insn): Likewise.
6683 (xordi3_insn): Likewise.
6684 (one_cmpldi2): Likewise.
6685 * config/arm/types.md (type): Add Neon types.
6686 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6687 use "type" attribute.
6688 (movmisalign<mode>_neon_store): Likewise.
6689 (movmisalign<mode>_neon_load): Likewise.
6690 (vec_set<mode>_internal): Likewise.
6691 (vec_setv2di_internal): Likewise.
6692 (vec_extract<mode>): Likewise.
6693 (vec_extractv2di): Likewise.
6694 (add<mode>3_neon): Likewise.
6695 (adddi3_neon): Likewise.
6696 (sub<mode>3_neon): Likewise.
6697 (subdi3_neon): Likewise.
6698 (mul<mode>3_neon): Likewise.
6699 (mul<mode>3add<mode>_neon): Likewise.
6700 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6701 (fma<VCVTF:mode>4)): Likewise.
6702 (fma<VCVTF:mode>4_intrinsic): Likewise.
6703 (fmsub<VCVTF:mode>4)): Likewise.
6704 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6705 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6706 (ior<mode>3): Likewise.
6707 (and<mode>3): Likewise.
6708 (anddi3_neon): Likewise.
6709 (orn<mode>3_neon): Likewise.
6710 (orndi3_neon): Likewise.
6711 (bic<mode>3_neon): Likewise.
6712 (bicdi3_neon): Likewise.
6713 (xor<mode>3): Likewise.
6714 (one_cmpl<mode>2): Likewise.
6715 (abs<mode>2): Likewise.
6716 (neg<mode>2): Likewise.
6717 (umin<mode>3_neon): Likewise.
6718 (umax<mode>3_neon): Likewise.
6719 (smin<mode>3_neon): Likewise.
6720 (smax<mode>3_neon): Likewise.
6721 (vashl<mode>3): Likewise.
6722 (vashr<mode>3_imm): Likewise.
6723 (vlshr<mode>3_imm): Likewise.
6724 (ashl<mode>3_signed): Likewise.
6725 (ashl<mode>3_unsigned): Likewise.
6726 (neon_load_count): Likewise.
6727 (ashldi3_neon_noclobber): Likewise.
6728 (signed_shift_di3_neon): Likewise.
6729 (unsigned_shift_di3_neon): Likewise.
6730 (ashrdi3_neon_imm_noclobber): Likewise.
6731 (lshrdi3_neon_imm_noclobber): Likewise.
6732 (widen_ssum<mode>3): Likewise.
6733 (widen_usum<mode>3): Likewise.
6734 (quad_halves_<code>v4si): Likewise.
6735 (quad_halves_<code>v4sf): Likewise.
6736 (quad_halves_<code>v8hi): Likewise.
6737 (quad_halves_<code>v16qi): Likewise.
6738 (reduc_splus_v2di): Likewise.
6739 (neon_vpadd_internal<mode>): Likewise.
6740 (neon_vpsmin<mode>): Likewise.
6741 (neon_vpsmax<mode>): Likewise.
6742 (neon_vpumin<mode>): Likewise.
6743 (neon_vpumax<mode>): Likewise.
6744 (ss_add<mode>_neon): Likewise.
6745 (us_add<mode>_neon): Likewise.
6746 (ss_sub<mode>_neon): Likewise.
6747 (us_sub<mode>_neon): Likewise.
6748 (neon_vadd<mode>_unspec): Likewise.
6749 (neon_vaddl<mode>): Likewise.
6750 (neon_vaddw<mode>): Likewise.
6751 (neon_vhadd<mode>): Likewise.
6752 (neon_vqadd<mode>): Likewise.
6753 (neon_vaddhn<mode>): Likewise.
6754 (neon_vmul<mode>): Likewise.
6755 (neon_vmla<mode>): Likewise.
6756 (neon_vmlal<mode>): Likewise.
6757 (neon_vmls<mode>): Likewise.
6758 (neon_vmlsl<mode>): Likewise.
6759 (neon_vqdmulh<mode>): Likewise.
6760 (neon_vqdmlal<mode>): Likewise.
6761 (neon_vqdmlsl<mode>): Likewise.
6762 (neon_vmull<mode>): Likewise.
6763 (neon_vqdmull<mode>): Likewise.
6764 (neon_vsub<mode>_unspec): Likewise.
6765 (neon_vsubl<mode>): Likewise.
6766 (neon_vsubw<mode>): Likewise.
6767 (neon_vqsub<mode>): Likewise.
6768 (neon_vhsub<mode>): Likewise.
6769 (neon_vsubhn<mode>): Likewise.
6770 (neon_vceq<mode>): Likewise.
6771 (neon_vcge<mode>): Likewise.
6772 (neon_vcgeu<mode>): Likewise.
6773 (neon_vcgt<mode>): Likewise.
6774 (neon_vcgtu<mode>): Likewise.
6775 (neon_vcle<mode>): Likewise.
6776 (neon_vclt<mode>): Likewise.
6777 (neon_vcage<mode>): Likewise.
6778 (neon_vcagt<mode>): Likewise.
6779 (neon_vtst<mode>): Likewise.
6780 (neon_vabd<mode>): Likewise.
6781 (neon_vabdl<mode>): Likewise.
6782 (neon_vaba<mode>): Likewise.
6783 (neon_vabal<mode>): Likewise.
6784 (neon_vmax<mode>): Likewise.
6785 (neon_vmin<mode>): Likewise.
6786 (neon_vpaddl<mode>): Likewise.
6787 (neon_vpadal<mode>): Likewise.
6788 (neon_vpmax<mode>): Likewise.
6789 (neon_vpmin<mode>): Likewise.
6790 (neon_vrecps<mode>): Likewise.
6791 (neon_vrsqrts<mode>): Likewise.
6792 (neon_vqabs<mode>): Likewise.
6793 (neon_vqneg<mode>): Likewise.
6794 (neon_vcls<mode>): Likewise.
6795 (clz<mode>2): Likewise.
6796 (popcount<mode>2): Likewise.
6797 (neon_vrecpe): Likewise.
6798 (neon_vrsqrte): Likewise.
6799 (neon_vget_lane<mode>_sext_internal): Likewise.
6800 (neon_vget_lane<mode>_zext_internal): Likewise.
6801 (neon_vdup_n<mode>): Likewise.
6802 (neon_vdup_nv2di): Likewise.
6803 (neon_vdpu_lane<mode>_internal): Likewise.
6804 (neon_vswp<mode>): Likewise.
6805 (float<mode><V_cvtto>2): Likewise.
6806 (floatuns<mode><V_cvtto>2): Likewise.
6807 (fix_trunc<mode><V_cvtto>)2): Likewise
6808 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6809 (neon_vcvt<mode>): Likewise.
6810 (neon_vcvtv4sfv4hf): Likewise.
6811 (neon_vcvtv4hfv4sf): Likewise.
6812 (neon_vcvt_n<mode>): Likewise.
6813 (neon_vmovn<mode>): Likewise.
6814 (neon_vqmovn<mode>): Likewise.
6815 (neon_vqmovun<mode>): Likewise.
6816 (neon_vmovl<mode>): Likewise.
6817 (neon_vmul_lane<mode>): Likewise.
6818 (neon_vmull_lane<mode>): Likewise.
6819 (neon_vqdmull_lane<mode>): Likewise.
6820 (neon_vqdmulh_lane<mode>): Likewise.
6821 (neon_vmla_lane<mode>): Likewise.
6822 (neon_vmlal_lane<mode>): Likewise.
6823 (neon_vqdmlal_lane<mode>): Likewise.
6824 (neon_vmls_lane<mode>): Likewise.
6825 (neon_vmlsl_lane<mode>): Likewise.
6826 (neon_vqdmlsl_lane<mode>): Likewise.
6827 (neon_vext<mode>): Likewise.
6828 (neon_vrev64<mode>): Likewise.
6829 (neon_vrev32<mode>): Likewise.
6830 (neon_vrev16<mode>): Likewise.
6831 (neon_vbsl<mode>_internal): Likewise.
6832 (neon_vshl<mode>): Likewise.
6833 (neon_vqshl<mode>): Likewise.
6834 (neon_vshr_n<mode>): Likewise.
6835 (neon_vshrn_n<mode>): Likewise.
6836 (neon_vqshrn_n<mode>): Likewise.
6837 (neon_vqshrun_n<mode>): Likewise.
6838 (neon_vshl_n<mode>): Likewise.
6839 (neon_vqshl_n<mode>): Likewise.
6840 (neon_vqshlu_n<mode>): Likewise.
6841 (neon_vshll_n<mode>): Likewise.
6842 (neon_vsra_n<mode>): Likewise.
6843 (neon_vsri_n<mode>): Likewise.
6844 (neon_vsli_n<mode>): Likewise.
6845 (neon_vtbl1v8qi): Likewise.
6846 (neon_vtbl2v8qi): Likewise.
6847 (neon_vtbl3v8qi): Likewise.
6848 (neon_vtbl4v8qi): Likewise.
6849 (neon_vtbx1v8qi): Likewise.
6850 (neon_vtbx2v8qi): Likewise.
6851 (neon_vtbx3v8qi): Likewise.
6852 (neon_vtbx4v8qi): Likewise.
6853 (neon_vtrn<mode>_internal): Likewise.
6854 (neon_vzip<mode>_internal): Likewise.
6855 (neon_vuzp<mode>_internal): Likewise.
6856 (neon_vld1<mode>): Likewise.
6857 (neon_vld1_lane<mode>): Likewise.
6858 (neon_vld1_dup<mode>): Likewise.
6859 (neon_vld1_dupv2di): Likewise.
6860 (neon_vst1<mode>): Likewise.
6861 (neon_vst1_lane<mode>): Likewise.
6862 (neon_vld2<mode>): Likewise.
6863 (neon_vld2_lane<mode>): Likewise.
6864 (neon_vld2_dup<mode>): Likewise.
6865 (neon_vst2<mode>): Likewise.
6866 (neon_vst2_lane<mode>): Likewise.
6867 (neon_vld3<mode>): Likewise.
6868 (neon_vld3qa<mode>): Likewise.
6869 (neon_vld3qb<mode>): Likewise.
6870 (neon_vld3_lane<mode>): Likewise.
6871 (neon_vld3_dup<mode>): Likewise.
6872 (neon_vst3<mode>): Likewise.
6873 (neon_vst3qa<mode>): Likewise.
6874 (neon_vst3qb<mode>): Likewise.
6875 (neon_vst3_lane<mode>): Likewise.
6876 (neon_vld4<mode>): Likewise.
6877 (neon_vld4qa<mode>): Likewise.
6878 (neon_vld4qb<mode>): Likewise.
6879 (neon_vld4_lane<mode>): Likewise.
6880 (neon_vld4_dup<mode>): Likewise.
6881 (neon_vst4<mode>): Likewise.
6882 (neon_vst4qa<mode>): Likewise.
6883 (neon_vst4qb<mode>): Likewise.
6884 (neon_vst4_lane<mode>): Likewise.
6885 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6886 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6887 (neon_vec_<US>mult_lo_<mode>): Likewise.
6888 (neon_vec_<US>mult_hi_<mode>): Likewise.
6889 (neon_vec_<US>shiftl_<mode>): Likewise.
6890 (neon_unpack<US>_<mode>): Likewise.
6891 (neon_vec_<US>mult_<mode>): Likewise.
6892 (vec_pack_trunc_<mode>): Likewise.
6893 (neon_vec_pack_trunk_<mode>): Likewise.
6894 (neon_vabd<mode>_2): Likewise.
6895 (neon_vabd<mode>_3): Likewise.
6896 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6897 (thumb2_movsi_vfp): Likewise.
6898 (movdi_vfp): Likewise.
6899 (movdi_vfp_cortexa8): Likewise.
6900 (movhf_vfp_neon): Likewise.
6901 (movhf_vfp): Likewiwse.
6902 (movsf_vfp): Likewiwse.
6903 (thumb2_movsf_vfp): Likewiwse.
6904 (movdf_vfp): Likewise.
6905 (thumb2_movdf_vfp): Likewise.
6906 (movsfcc_vfp): Likewise.
6907 (thumb2_movsfcc_vfp): Likewise.
6908 (movdfcc_vfp): Likewise.
6909 (thumb2_movdfcc_vfp): Likewise.
6910 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6911 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6912 (v10_v2c): Likewise.
6913 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6914 attribute change.
6915 (cortex_a15_neon_int_2): Likewise.
6916 (cortex_a15_neon_int_3): Likewise.
6917 (cortex_a15_neon_int_4): Likewise.
6918 (cortex_a15_neon_int_5): Likewise.
6919 (cortex_a15_neon_vqneg_vqabs): Likewise.
6920 (cortex_a15_neon_vmov): Likewise.
6921 (cortex_a15_neon_vaba): Likewise.
6922 (cortex_a15_neon_vaba_qqq): Likewise.
6923 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6924 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6925 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6926 scalar_64_32_long_scalar): Likewise.
6927 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6928 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6929 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6930 lotype_qdd_64_32_long): Likewise.
6931 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6932 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6933 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6934 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6935 (cortex_a15_neon_shift_1): Likewise.
6936 (cortex_a15_neon_shift_2): Likewise.
6937 (cortex_a15_neon_shift_3): Likewise.
6938 (cortex_a15_neon_vshl_ddd): Likewise.
6939 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6940 (cortex_a15_neon_vsra_vrsra): Likewise.
6941 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6942 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6943 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6944 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6945 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6946 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6947 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6948 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6949 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6950 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6951 (cortex_a15_neon_bp_simple): Likewise.
6952 (cortex_a15_neon_bp_2cycle): Likewise.
6953 (cortex_a15_neon_bp_3cycle): Likewise.
6954 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6955 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6956 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6957 (cortex_a15_neon_vld2_4_regs): Likewise.
6958 (cortex_a15_neon_vld3_vld4): Likewise.
6959 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6960 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6961 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6962 (cortex_a15_neon_vst3_vst4): Likewise.
6963 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6964 (cortex_a15_neon_vld3_vld4_lane" 10
6965 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6966 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6967 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6968 (cortex_a15_neon_ldm_2): Likewise.0
6969 (cortex_a15_neon_stm_2): Likewise.
6970 (cortex_a15_neon_mcr): Likewise.
6971 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6972 (cortex_a15_neon_mrc): Likewise.
6973 (cortex_a15_neon_mrrc): Likewise.
6974 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6975 change.
6976 (cortex_a15_alu_shift): Likewise.
6977 (cortex_a15_alu_shift_reg): Likewise.
6978 (cortex_a15_mult32): Likewise.
6979 (cortex_a15_mult64): Likewise.
6980 (cortex_a15_block): Likewise.
6981 (cortex_a15_branch): Likewise.
6982 (cortex_a15_load1): Likewise.
6983 (cortex_a15_load3): Likewise.
6984 (cortex_a15_store1): Likewise.
6985 (cortex_a15_store3): Likewise.
6986 (cortex_a15_call): Likewise.
f43245d1 6987 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG		 6988 (cortex_a5_f2r): Likewise.
6989 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6990 change.
6991 (cortex_a53_f2r): Likewise.
6992 * config/arm/cortex-a7.md
6993 (cortex_a7_branch): Update for attribute change.
6994 (cortex_a7_call): Likewise.
6995 (cortex_a7_alu_imm): Likewise.
6996 (cortex_a7_alu_reg): Likewise.
6997 (cortex_a7_alu_shift): Likewise.
6998 (cortex_a7_mul): Likewise.
6999 (cortex_a7_load1): Likewise.
7000 (cortex_a7_store1): Likewise.
7001 (cortex_a7_load2): Likewise.
7002 (cortex_a7_store2): Likewise.
7003 (cortex_a7_load3): Likewise.
7004 (cortex_a7_store3): Likewise.
7005 (cortex_a7_load4): Likewise.
7006 (cortex_a7_store4): Likewise.
7007 (cortex_a7_fpalu): Likewise.
7008 (cortex_a7_fconst): Likewise.
7009 (cortex_a7_fpmuls): Likewise.
7010 (cortex_a7_neon_mul): Likewise.
7011 (cortex_a7_fpmacs): Likewise.
7012 (cortex_a7_neon_mla: Likewise.
7013 (cortex_a7_fpmuld: Likewise.
7014 (cortex_a7_fpmacd: Likewise.
7015 (cortex_a7_fpfmad: Likewise.
7016 (cortex_a7_fdivs: Likewise.
7017 (cortex_a7_fdivd: Likewise.
7018 (cortex_a7_r2f: Likewise.
7019 (cortex_a7_f2r: Likewise.
7020 (cortex_a7_f_flags: Likewise.
7021 (cortex_a7_f_loads: Likewise.
7022 (cortex_a7_f_loadd: Likewise.
7023 (cortex_a7_f_stores: Likewise.
7024 (cortex_a7_f_stored: Likewise.
7025 (cortex_a7_neon): Likewise.
7026 * config/arm/cortex-a8-neon.md
7027 (cortex_a8_neon_mrc): Update for attribute change.
7028 (cortex_a8_neon_mrrc): Likewise.
7029 (cortex_a8_neon_int_1): Likewise.
7030 (cortex_a8_neon_int_2): Likewise.
7031 (cortex_a8_neon_int_3): Likewise.
7032 (cortex_a8_neon_int_4): Likewise.
7033 (cortex_a8_neon_int_5): Likewise.
7034 (cortex_a8_neon_vqneg_vqabs): Likewise.
7035 (cortex_a8_neon_vmov): Likewise.
7036 (cortex_a8_neon_vaba): Likewise.
7037 (cortex_a8_neon_vaba_qqq): Likewise.
7038 (cortex_a8_neon_vsma): Likewise.
7039 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7040 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7041 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7042 long_scalar): Likewise.
003bb7f3
JG		 7043 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7044 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7045 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7046 long_scalar_qdd_64_32_long): Likewise.
7047 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7048 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7049 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7050 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7051 (cortex_a8_neon_shift_1): Likewise.
7052 (cortex_a8_neon_shift_2): Likewise.
7053 (cortex_a8_neon_shift_3): Likewise.
7054 (cortex_a8_neon_vshl_ddd): Likewise.
7055 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7056 (cortex_a8_neon_vsra_vrsra): Likewise.
7057 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7058 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7059 (cortex_a8_neon_fp_vsum): Likewise.
7060 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7061 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7062 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7063 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7064 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7065 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7066 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7067 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7068 (cortex_a8_neon_bp_simple): Likewise.
7069 (cortex_a8_neon_bp_2cycle): Likewise.
7070 (cortex_a8_neon_bp_3cycle): Likewise.
7071 (cortex_a8_neon_ldr): Likewise.
7072 (cortex_a8_neon_str): Likewise.
7073 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7074 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7075 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7076 (cortex_a8_neon_vld2_4_regs): Likewise.
7077 (cortex_a8_neon_vld3_vld4): Likewise.
7078 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7079 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7080 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7081 (cortex_a8_neon_vst3_vst4): Likewise.
7082 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7083 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7084 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7085 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7086 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7087 (cortex_a8_neon_mcr): Likewise.
7088 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 7089 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG		 7090 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7091 change.
7092 (ca9_neon_mrrc): Likewise.
7093 (cortex_a9_neon_int_1): Likewise.
7094 (cortex_a9_neon_int_2): Likewise.
7095 (cortex_a9_neon_int_3): Likewise.
7096 (cortex_a9_neon_int_4): Likewise.
7097 (cortex_a9_neon_int_5): Likewise.
7098 (cortex_a9_neon_vqneg_vqabs): Likewise.
7099 (cortex_a9_neon_vmov): Likewise.
7100 (cortex_a9_neon_vaba): Likewise.
7101 (cortex_a9_neon_vaba_qqq): Likewise.
7102 (cortex_a9_neon_vsma): Likewise.
7103 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7104 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7105 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7106 long_scalar): Likewise.
003bb7f3
JG		 7107 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7108 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7109 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7110 long_scalar_qdd_64_32_long): Likewise.
7111 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7112 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7113 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7114 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7115 (cortex_a9_neon_shift_1): Likewise.
7116 (cortex_a9_neon_shift_2): Likewise.
7117 (cortex_a9_neon_shift_3): Likewise.
7118 (cortex_a9_neon_vshl_ddd): Likewise.
7119 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7120 (cortex_a9_neon_vsra_vrsra): Likewise.
7121 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7122 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7123 (cortex_a9_neon_fp_vsum): Likewise.
7124 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7125 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7126 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7127 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7128 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7129 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7130 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7131 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7132 (cortex_a9_neon_bp_simple): Likewise.
7133 (cortex_a9_neon_bp_2cycle): Likewise.
7134 (cortex_a9_neon_bp_3cycle): Likewise.
7135 (cortex_a9_neon_ldr): Likewise.
7136 (cortex_a9_neon_str): Likewise.
7137 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7138 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7139 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7140 (cortex_a9_neon_vld2_4_regs): Likewise.
7141 (cortex_a9_neon_vld3_vld4): Likewise.
7142 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7143 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7144 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7145 (cortex_a9_neon_vst3_vst4): Likewise.
7146 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7147 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7148 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7149 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7150 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7151 (cortex_a9_neon_mcr): Likewise.
7152 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7153 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7154 (cortex_a9_fps): Likewise.
7155 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7156 change.
7157 (cortex_m4_fmuls): Likewise.
7158 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7159 change.
7160 (cortex_r4_mrc): Likewise.
7161 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 7162 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 7163 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB		 7164 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7165 attribute change.
003bb7f3
JG		 7166 (pj4_core_to_vfp): Likewise.
7167 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7168 attribute change.
7169 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7170 (vfp_fstore): Likewise.
7171 * doc/md.texi: Change references to neon_type to refer to type.
7172
fbe0dc78
DS		 71732013-09-04 Dodji Seketeli <dodji@redhat.com>
7174
7175 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7176
c167bc5b
DE		 71772013-09-04 David Edelsohn <dje.gcc@gmail.com>
7178
7179 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7180 lglobl if not weak.
7181
035cb59f
ER		 71822013-09-04 Easwaran Raman <eraman@google.com>
7183
7184 PR middle-end/57370
7185 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7186 (build_and_add_sum): Use it.
7187 (appears_later_in_bb): Simplify code.
7188
71191083
TJ		 71892013-09-04 Teresa Johnson <tejohnson@google.com>
7190
7191 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7192
0c488213
JG		 71932013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7194
7195 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7196
61a74079
JH		 71972013-09-04 Jan Hubicka <jh@suse.cz>
7198
7199 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7200 * ipa-devirt.c: Include diganostic.h
7201 (odr_type_d): Add types and types_set.
7202 (hash_type_name): Work for types with vtables during LTO.
7203 (odr_hasher::remove): Fix comment; destroy types_set.
7204 (add_type_duplicate): New function,
7205 (get_odr_type): Use it.
7206 (dump_type_inheritance_graph): Dump type duplicates.
7207 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7208 graph.
7209 * tree.c (types_same_for_odr): Give exact answers on types with
7210 virtual tables.
7211
24ba8195
DS		 72122013-09-04 Dodji Seketeli <dodji@redhat.com>
7213
7214 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 7215 explaining their differences.
24ba8195 7216
0a04869e
SKS		 72172013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7218
7219 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7220
3b18bc42
JL		 72212013-09-03 Jeff Law <law@redhat.com>
7222
7223 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7224 when not threading through a joiner block. Pass joiner/no joiner
7225 state to register_jump_thread.
7226 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7227 state from argument rather than implying on path length.
7228 Dump the entire jump thread path into debugging dump.
7229 * tree-flow.h (register_jump_thread): Update prototype.
7230
ae84e151
XDL		 72312013-08-29 Xinliang David Li <davidxl@google.com>
7232
7233 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7234 Remove a trivial gcc_assert.
7235
c716e67f
XDL		 72362013-08-29 Xinliang David Li <davidxl@google.com>
7237
7238 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7239 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7240 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7241 Delay base decl alignment adjustment.
7242 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7243 * tree-vectorizer.h: New data structure.
7244 (set_dr_misalignment): New function.
7245 (dr_misalignment): Ditto.
7246 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7247 (vectorizable_load): Ditto.
7248 (ensure_base_align): New function.
7249 (vectorize_loops): Add dbg_cnt support.
7250 (execute_vect_slp): Ditto.
7251 * dbgcnt.def: New debug counter.
7252 * Makefile: New dependency.
7253
db8800bc
MI		 72542013-09-03 Meador Inge <meadori@codesourcery.com>
7255
7256 Revert:
7257
7258 2013-08-30 Meador Inge <meadori@codesourcery.com>
7259
7260 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7261
0d9b4f5c
DE		 72622013-09-03 David Edelsohn <dje.gcc@gmail.com>
7263
7264 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7265 function descriptor.
7266
29e810b9
RB		 72672013-09-03 Richard Biener <rguenther@suse.de>
7268
7269 * tree-affine.c (add_elt_to_tree): Fix association issue,
7270 avoid useless converts and make sure to always return a
7271 properly typed result.
7272
dbfc2894
RB		 72732013-09-03 Richard Biener <rguenther@suse.de>
7274
7275 PR middle-end/57656
7276 * fold-const.c (negate_expr_p): Fix division case.
7277 (negate_expr): Likewise.
7278
fa3bf4ce
RB		 72792013-09-03 Richard Biener <rguenther@suse.de>
7280
7281 PR lto/58285
7282 * tree-streamer-out.c: Include tm.h.
7283 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7284
a64bbb3f
JH		 72852013-09-03 Jan Hubicka <jh@suse.cz>
7286
7287 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7288
93ed07e2
AM		 72892013-09-03 Alan Modra <amodra@gmail.com>
7290
7291 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7292 multilibs to big-endian target and vice versa.
7293 * config/rs6000/t-linux64: Use := assignment on all vars.
7294 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7295 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7296 * config/rs6000/t-linux64le: New file.
7297 * config/rs6000/t-linux64bele: New file.
7298 * config/rs6000/t-linux64lebe: New file.
7299
07d6cd64
JH		 73002013-09-02 Jan Hubicka <jh@suse.cz>
7301
7302 * ipa-inline-transform.c (inline_transform): Do not
7303 optimize_inline_calls when not optimizing.
7304
8e4c9a10
JH		 73052013-09-02 Jan Hubicka <jh@suse.cz>
7306
7307 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7308 duplicated nodes for assembler names.
7309 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7310 hard registers from assembler name hash.
7311
c91061e6
JH		 73122013-09-02 Jan Hubicka <jh@suse.cz>
7313
7314 * ipa-split.c (execute_split_functions): Split externally visible
7315 functions called once.
7316
2fd0985c
MJ		 73172013-09-02 Martin Jambor <mjambor@suse.cz>
7318
7319 PR ipa/58106
7320 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7321 linked list. When finding the correct duplicate, also consider also
7322 the caller in additon to its inlined_to node.
7323
91bd4114
JG		 73242013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7325
7326 * config/aarch64/aarch64-simd-builtins.def
7327 (dup_lane_scalar): Remove.
7328 * config/aarch64/aarch64-simd.md
7329 (aarch64_simd_dup): Add 'w->w' alternative.
7330 (aarch64_dup_lane<mode>): Allow for VALL.
7331 (aarch64_dup_lane_scalar<mode>): Remove.
7332 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7333 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7334 (aarch64_get_lane_unsigned<mode>): Likewise.
7335 (aarch64_get_lane<mode>): Likewise.
7336 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7337 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7338 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7339 (VCON): Change container of V2SF.
7340 (vswap_width_name): Likewise.
7341 * config/aarch64/arm_neon.h
7342 (__aarch64_vdup_lane_any): New.
7343 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7344 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7345 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7346
d617d2d8
EB		 73472013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7348
7349 PR middle-end/56382
7350 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7351 the source or the destination is a single hard register.
7352
63fdb7be
RB		 73532013-09-02 Richard Biener <rguenther@suse.de>
7354
7355 PR middle-end/57511
7356 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7357 non-linear SCEVs.
7358
78de2333
RB		 73592013-09-02 Richard Biener <rguenther@suse.de>
7360
7361 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7362 arithmetic to sizetype.
7363
85ff4ec6
BC		 73642013-09-02 Bin Cheng <bin.cheng@arm.com>
7365
7366 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7367 Find auto-increment use both before and after candidate.
7368
576016fe
MP		 73692013-09-02 Marek Polacek <polacek@redhat.com>
7370
7371 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7372
08f835dc
JH		 73732013-09-01 Jan Hubicka <jh@suse.cz>
7374
7375 * Makefile.in: Add ipa-profile.o
7376 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7377 * cgraph.c (struct cgraph_propagate_frequency_data,
7378 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7379 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7380 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB		 7381 * ipa-inline-analysis.c: Include ipa-utils.h;
7382 drop duplicated cfgloop.h.
08f835dc
JH		 7383 (inline_update_callee_summaries): Update.
7384 * ipa-profile.c: New file.
7385 * ipa-utils.h (ipa_propagate_frequency): Declare.
7386 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 7387 data-streamer.h, value-prof.h.
08f835dc
JH		 7388 (symtab_remove_unreachable_nodes): Update profile.
7389 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7390 account_time_size, cmp_counts, dump_histogram,
7391 ipa_profile_generate_summary, ipa_profile_write_summary,
7392 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7393 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 7394 Move to ipa-profile.c.
08f835dc 7395
cec8583c
JDA		 73962013-09-01 John David Anglin <danglin@gcc.gnu.org>
7397
7398 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7399
bbc9396b
JH		 74002013-09-01 Jan Hubicka <jh@suse.cz>
7401
7402 * common.opt (fdevirtualize-speculatively): New function.
7403 * invoke.texi (fdevirtualize-speculatively): Document.
7404 * ipa-devirt.c: Include ipa-inline.h
7405 (likely_target_p): New function.
7406 (ipa_devirt): New function.
7407 (gate_ipa_devirt): New function.
7408 (pass_data_ipa_devirt): New static var.
7409 (pass_ipa_devirt): Likewise.
7410 (make_pass_ipa_devirt): New function.
7411 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7412 (common_handle_option): Disable devirtualization when
7413 value range profiling is available.
7414 * passes.def (pass_ipa_devirt): Add.
7415 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 7416 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 7417
0cea211e
IS		 74182013-09-01 Iain Sandoe <iain@codesourcery.com>
7419
7420 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7421 include sanitize(undefined).
7422
5e351e96
DN		 74232013-08-31 Diego Novillo <dnovillo@google.com>
7424
7425 * Makefile.in (TREE_CORE_H): Define.
7426 (TREE_H): Use.
7427 (GTFILES): Add tree-core.h.
7428 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7429 size the array.
7430 * tree-core.h: New file.
7431 Move all data structures, enum, typedefs, global
7432 declarations and constants from ...
7433 * tree.h: ... here.
7434
63bf9a90
JH		 74352013-08-31 Jan Hubicka <jh@suse.cz>
7436
7437 * bulitins.c (expand_builtin): Do not early exit for gcov
7438 instrumented functions.
7439
5c0abd6a
MP		 74402013-08-31 Marek Polacek <polacek@redhat.com>
7441
7442 * ubsan.c: Include tm_p.h.
7443
4843f032
JH		 74442013-08-31 Jan Hubicka <jh@suse.cz>
7445
8356c89c
JH		 7446 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7447 warning.
7448
4843f032
JH		 7449 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7450 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7451 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 7452 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7453 (ipa_merge_profiles): New function.
7454 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7455 (lto_input_function_body): Likewise.
5e581212
JH		 7456 * ipa-utils.h (ipa_merge_profiles): Declare.
7457 * lto-streamer.h (lto_input_function_body): Update prototype.
7458 (emit_label_in_global_context_p): Remove.
7459 * lto-symtab.c: Include ipa-utils.h
7460 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7461
57292ce9
JH		 74622013-08-31 Jan Hubicka <jh@suse.cz>
7463
7464 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7465
b9cbfeeb
JH		 74662013-08-31 Jan Hubicka <jh@suse.cz>
7467
7468 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7469
41dedebd
UB		 74702013-08-31 Uros Bizjak <ubizjak@gmail.com>
7471
7472 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7473 "cmp" RTX before signed_comparison_operator check to account
7474 for "code" changes.
7475
06d65050
JH		 74762013-08-30 Jan Hubicka <jh@suse.cz>
7477
41dedebd 7478 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 7479 (detect_type_change_1): Rename to ...
7480 (detect_type_change): ... this one; early return on non-polymorphic
7481 types.
41dedebd 7482 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 7483 use of detect_type_change.
7484 (compute_complex_assign_jump_func): Add param_type parameter;
7485 update use of detect_type_change_ssa.
7486 (compute_complex_ancestor_jump_func): Likewise.
7487 (ipa_get_callee_param_type): New function.
7488 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7489 update calls to the jump function computation functions.
7490
600b5b1d
TJ		 74912013-08-30 Teresa Johnson <tejohnson@google.com>
7492 Steven Bosscher <steven@gcc.gnu.org>
7493
7494 * cfgrtl.c (fixup_new_cold_bb): New routine.
7495 (commit_edge_insertions): Invoke fixup_partitions.
7496 (find_partition_fixes): New routine.
7497 (fixup_partitions): Ditto.
7498 (verify_hot_cold_block_grouping): Update comments.
7499 (rtl_verify_edges): Invoke find_partition_fixes.
7500 (rtl_verify_bb_pointers): Update comments.
7501 (rtl_verify_bb_layout): Ditto.
7502 * basic-block.h (probably_never_executed_edge_p): Declare.
7503 (fixup_partitions): Ditto.
7504 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7505 * bb-reorder.c (sanitize_hot_paths): New function.
7506 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7507 sanitize_hot_paths.
7508 * predict.c (probably_never_executed_edge_p): New routine.
7509 * cfg.c (check_bb_profile): Add partition insanity warnings.
7510
0a6ea5c9
MI		 75112013-08-30 Meador Inge <meadori@codesourcery.com>
7512
7513 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7514
de5a5fa1
MP		 75152013-08-30 Marek Polacek <polacek@redhat.com>
7516
7517 * Makefile.in (ubsan.o): Add.
7518 (c-family/c-ubsan.o): Add.
7519 (builtins.o): Add ubsan.h dependency.
7520 * ubsan.h: New file.
7521 * ubsan.c: New file.
7522 * common.opt: Add -fsanitize=undefined option.
7523 (flag_sanitize): Add variable.
7524 (fsanitize=): Add option. Add Driver.
7525 (fsanitize=thread): Remove option.
7526 (fsanitize=address): Likewise.
7527 (static-libubsan): New option.
7528 * doc/invoke.texi: Document the new flag and -static-libubsan.
7529 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7530 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7531 * builtin-attrs.def (ATTR_COLD): Define.
7532 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7533 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7534 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7535 * flag-types.h (sanitize_code): New enum.
7536 * opts.c (common_handle_option): Parse command line arguments
7537 of -fsanitize=. Add -fsanitize=unreachable option.
7538 * varasm.c (get_variable_section): Adjust.
7539 (assemble_noswitch_variable): Likewise.
7540 (assemble_variable): Likewise.
7541 (output_constant_def_contents): Likewise.
7542 (categorize_decl_for_section): Likewise.
7543 (place_block_symbol): Likewise.
7544 (output_object_block): Likewise.
7545 * builtins.def: Likewise.
7546 * toplev.c (compile_file): Likewise.
7547 (process_options): Likewise.
7548 * cppbuiltin.c: Likewise.
7549 * tsan.c (tsan_pass): Likewise.
7550 (tsan_gate): Likewise.
7551 (tsan_gate_O0): Likewise.
7552 * cfgexpand.c (partition_stack_vars): Likewise.
7553 (expand_stack_vars): Likewise.
7554 (defer_stack_allocation): Likewise.
7555 (expand_used_vars): Likewise.
7556 * cfgcleanup.c (old_insns_match_p): Likewise.
7557 * asan.c (asan_finish_file): Likewise.
7558 (asan_instrument): Likewise.
7559 (gate_asan): Likewise.
7560 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7561 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7562 (asan_global_struct): Use pointer_sized_int_node instead
7563 calling build_nonstandard_integer_type.
7564 (initialize_sanitizer_builtins): Likewise.
7565 (asan_finish_file): Likewise.
7566 * gcc.c: Document %{%:function(args):X}.
7567 (static_spec_functions): Add sanitize.
7568 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7569 store funcval != NULL there.
7570 (do_spec_1): Adjust handle_spec_function caller.
7571 (handle_braces): Allow %:function(args) as condition.
7572 (sanitize_spec_function): New function.
7573 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7574 (LIBUBSAN_SPEC): Likewise.
7575 (LIBUBSAN_EARLY_SPEC): Likewise.
7576 (SANITIZER_SPEC): Handle libubsan.
7577 (SANITIZER_EARLY_SPEC): Likewise.
7578 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7579 instead of fsanitize=address.
7580 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7581 instead of fsanitize=address*.
7582 * builtins.c: Include ubsan.h.
7583 (fold_builtin_0): Instrument __builtin_unreachable.
7584 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7585 instead of flag_asan.
7586 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7587 (pointer_sized_int_node): Define.
41dedebd 7588 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7589
f07f30cf
MS		 75902013-08-30 Mike Stump <mikestump@comcast.net>
7591
7592 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7593 with RE patterns.
7594
8a41354f
JH		 75952013-08-29 Jan Hubicka <jh@suse.cz>
7596
7597 * cgraph.c (cgraph_function_body_availability): Handle weakref
7598 correctly.
7599 * passes.def: Remove pass_fixup_cfg.
7600 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7601 track when we need to remove functions.
7602 (gate_ipa_inline): Execute inlining always; add comment why.
7603 (pass_data_ipa_inline): Remove TODO_remove_functions.
7604 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7605 do not produce summaries.
7606 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7607 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7608 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7609 constant pool and vtable.
7610
dc8d7a0f
TB		 76112013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7612
41dedebd
UB		 7613 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7614 New arm_neon.h's internal macros to specify 64-bit constants.
7615 Avoid using stdint.h's macros.
dc8d7a0f 7616
ac7eacd2
JR		 76172013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7618
7619 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7620 before applying REG_P.
7621
8b29fd4e
JJ		 76222013-08-30 Jakub Jelinek <jakub@redhat.com>
7623
7624 PR tree-optimization/58277
7625 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7626 after seeing too many stmts with vdef in between dombb and current
7627 bb, invalidate everything.
7628
26d75703
RB		 76292013-08-30 Richard Biener <rguenther@suse.de>
7630
7631 * fold-const.c (fold_single_bit_test): Fix overflow test.
7632
f7e088e7
EB		 76332013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7634
7635 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7636 and which can live in a register, always retrieve the value on entry.
7637 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7638 passed by invisible reference specially.
7639 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7640 (vt_add_function_parameter): Correctly deal with a parameter passed by
7641 invisible reference.
7642
76432013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 7644
7645 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7646
6e8dad05
RB		 76472013-08-30 Richard Biener <rguenther@suse.de>
7648
7649 PR tree-optimization/58228
7650 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7651 allow invariant loads in nested loop vectorization.
7652
062ef2c8
RB		 76532013-08-30 Richard Biener <rguenther@suse.de>
7654
7655 PR tree-optimization/58223
7656 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7657 (has_anti_or_output_dependence): ... this and adjust to also
7658 look for output dependences.
7659 (mark_nodes_having_upstream_mem_writes): Adjust.
7660 (rdg_flag_uses): Likewise.
7661
7a764c60
RB		 76622013-08-30 Richard Biener <rguenther@suse.de>
7663
7664 PR tree-optimization/58010
7665 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7666 assert that we have a loop-closed PHI.
7667
82e9d642
JH		 76682013-08-29 Jan Hubicka <jh@suse.cz>
7669
7670 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7671 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7672 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7673 (lto_free_function_in_decl_state_for_node): New function.
7674
9cc1fb4b
XDL		 76752013-08-29 Xinliang David Li <davidxl@google.com>
7676
41dedebd 7677 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 7678 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7679 Emit alignment peeling message with default -fopt-info.
7680 (vect_loop_versioning): Emit loop version info message.
41dedebd 7681 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 7682 (execute_vect_slp): Ditto.
7683
440a5082
EB		 76842013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7685
7686 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7687 of the clone from the DECL_NAME of the original function.
7688
3fa3690d
OE		 76892013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7690
7691 * passes.c (register_pass): Add overload.
7692 * tree-pass.h (register_pass): Forward declare it. Add comment.
7693
0170f33c
JH		 76942013-08-29 Jan Hubicka <jh@suse.cz>
7695
41dedebd
UB		 7696 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7697 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7698 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7699 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7700 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7701 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 7702 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7703 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7704
103ff0d6
TJ		 77052013-08-29 Teresa Johnson <tejohnson@google.com>
7706
7707 * dumpfile.c (dump_loc): Output column number.
7708 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7709 * doc/invoke.texi: Document optall -fopt-info flag.
7710 * profile.c (read_profile_edge_counts): Use new dump framework.
7711 (compute_branch_probabilities): Ditto.
7712 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7713 when pass not in any opt group.
7714 * pass_manager.h (pass_manager::get_pass_profile): New method.
7715 * value-prof.c (check_counter): Use new dump framework.
7716 (check_ic_target): Ditto.
7717 * coverage.c (get_coverage_counts): Ditto.
7718 (coverage_init): Setup new dump framework.
7719
301bbc16
RB		 77202013-08-29 Richard Biener <rguenther@suse.de>
7721
7722 PR tree-optimization/58246
7723 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7724 handle the dominance check inside a basic-block.
7725
1b275000
RB		 77262013-08-29 Richard Biener <rguenther@suse.de>
7727
7728 PR middle-end/57287
7729 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7730 of default defs that appear in abnormal PHI nodes.
7731
6b1184ba
RB		 77322013-08-29 Richard Biener <rguenther@suse.de>
7733
7734 PR tree-optimization/57685
7735 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7736 single-use operands to avoid exponential complexity.
7737
1b08b734
DC		 77382013-08-28 Dehao Chen <dehao@google.com>
7739
7740 * ipa-inline.c (edge_badness): Fix integer underflow.
7741
48a3fa69
UB		 77422013-08-28 Uros Bizjak <ubizjak@gmail.com>
7743
7744 * gtm-builtins.def (_ITM_free): Declare leaf.
7745
215f73e6
JJ		 77462013-08-28 Jakub Jelinek <jakub@redhat.com>
7747
7748 PR target/58067
7749 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7750 (*tls_local_dynamic_base_64_largepic): Likewise.
7751 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7752 Remove predicate from call operand.
7753 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7754 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7755
34554d1a
JL		 77562013-08-28 Jeff Law <law@redhat.com>
7757
7758 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7759 checks for the number of predecessors and successors allowed.
7760 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7761 which require copying a joiner block if there is a request which
7762 is a subpath that requires no joiner block copying.
7763
c01c111b
JH		 77642013-08-28 Jan Hubicka <jh@suse.cz>
7765
7766 * lto-streamer-out.c (DFS_write_tree_body): Drop
7767 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7768 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7769 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7770 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7771 Do not read DECL_ERROR_ISSUED.
7772 (unpack_ts_decl_with_vis_value_fields): Do not read
7773 DECL_DEFER_OUTPUT.
48a3fa69
UB		 7774 (lto_input_ts_binfo_tree_pointers): Do not read
7775 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 7776 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7777 write DECL_ERROR_ISSUED..
7778 (pack_ts_decl_with_vis_value_fields): Do not write
7779 DECL_DEFER_OUTPUT.
7780 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 7781 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 7782 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7783 * tree.h (tree_decl_common): Update comment.
7784 (DECL_ERROR_ISSUED): Remove.
7785
4b128ece
JJ		 77862013-08-28 Jakub Jelinek <jakub@redhat.com>
7787
7788 PR middle-end/58257
7789 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7790
12211b99 77912013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 7792
7793 * builtins.def (free): Declare leaf.
7794
8dce4dbc
DM		 77952013-08-27 David Malcolm <dmalcolm@redhat.com>
7796
7797 * gdbhooks.py: New.
7798 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7799 * configure: Regenerate.
7800
b8f6e610
MJ		 78012013-08-27 Martin Jambor <mjambor@suse.cz>
7802
7803 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7804 (ipa_ancestor_jf_data): Likewise.
7805 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7806 (ipa_get_jf_pass_through_type_preserved): New function.
7807 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7808 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 7809 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 7810 (ipa_get_jf_ancestor_result): Likewise.
7811 (propagate_vals_accross_pass_through): Use
7812 ipa_get_jf_pass_through_result to do all the value mappings.
7813 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7814 type_preserved flag.
7815 (ipa_set_jf_cst_copy): New function.
7816 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7817 (ipa_set_jf_arith_pass_through): Likewise.
7818 (ipa_set_ancestor_jf): Likewise.
7819 (compute_complex_assign_jump_func): Set type_preserved instead of
7820 punting.
7821 (ipa_compute_jump_functions_for_edge): Likewise.
7822 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7823 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 7824 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 7825 (ipa_write_jump_function): Stream the type_preserved flags.
7826 (ipa_read_jump_function): Likewise.
7827
74bf76ed
JJ		 78282013-08-27 Jakub Jelinek <jakub@redhat.com>
7829 Aldy Hernandez <aldyh@redhat.com>
7830
7831 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7832 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7833 * function.h (struct function): Add has_force_vect_loops and
7834 has_simduid_loops.
12211b99 7835 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 7836 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7837 handle it.
7838 * gimple.def: Update CLAUSES comments.
7839 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7840 (gimple_build_omp_for): Add argument to prototype.
7841 (gimple_omp_for_kind): New.
7842 (gimple_omp_for_set_kind): New.
7843 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7844 GOVD_DATA_SHARE_CLASS.
7845 (enum omp_region_type): Add ORT_SIMD.
7846 (gimple_add_tmp_var): Handle ORT_SIMD.
7847 (gimplify_var_or_parm_decl): Same.
7848 (is_gimple_stmt): Same.
7849 (omp_firstprivatize_variable): Same.
7850 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7851 (omp_notice_variable): Handle ORT_SIMD.
12211b99 7852 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 7853 (omp_check_private): Handle ORT_SIMD.
7854 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7855 OMP_CLAUSE_SAFELEN.
7856 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7857 Handle OMP_CLAUSE_LASTPRIVATE.
7858 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7859 OMP_CLAUSE_SAFELEN.
7860 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7861 (gimplify_expr): Handle OMP_SIMD.
7862 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7863 (expand_GOMP_SIMD_VF): New.
7864 (expand_GOMP_SIMD_LAST_LANE): New.
7865 * internal-fn.def (GOMP_SIMD_LANE): New.
7866 (GOMP_SIMD_VF): New.
7867 (GOMP_SIMD_LAST_LANE): New.
7868 * omp-low.c: Include target.h.
7869 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7870 OMP_CLAUSE_SAFELEN.
7871 (check_omp_nesting_restrictions): Same.
7872 (omp_max_vf): New.
7873 (lower_rec_simd_input_clauses): New.
7874 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7875 OMP_CLAUSE_LINEAR.
7876 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7877 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7878 (expand_omp_build_assign): New.
7879 (expand_omp_for_init_counts): New.
7880 (expand_omp_for_init_vars): New.
7881 (extract_omp_for_update_vars): New.
7882 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7883 and rewrite accordingly.
7884 (expand_omp_simd): New.
7885 (expand_omp_for): Use expand_omp_simd.
7886 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7887 (lower_omp_for): Do not lower the body.
7888 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7889 in their own loops.
7890 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 7891 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7892 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 7893 (gate_tree_if_conversion): Similarly.
7894 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7895 gimple_build_omp_for.
7896 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7897 * tree-parloops (create_parallel_loop): Pass kind argument to
7898 gimple_build_omp_for.
7899 * tree-pretty-print.c (dump_omp_clause): Add cases for
7900 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7901 OMP_CLAUSE__SIMDUID_.
7902 (dump_generic_node): Handle OMP_SIMD.
7903 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7904 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7905 unroll OMP_SIMD loops here.
12211b99 7906 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 7907 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 7908 loop->safelen.
74bf76ed
JJ		 7909 (vect_analyze_data_refs): Handle simd loops.
7910 * tree-vect-loop.c (vectorizable_live_operation): Handle
7911 IFN_GOMP_SIMD*.
12211b99 7912 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 7913 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7914 (vectorizable_load): Same.
12211b99 7915 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 7916 (struct simduid_to_vf): New.
7917 (simduid_to_vf::hash): New.
7918 (simduid_to-vf::equal): New.
7919 (struct simd_array_to_simduid): New.
7920 (simd_array_to_simduid::hash): New.
7921 (simd_array_to_simduid::equal): New.
7922 (adjust_simduid_builtins): New.
7923 (struct note_simd_array_uses_struct): New.
7924 (note_simd_array_uses_cb): New.
7925 (note_simd_array_uses): New.
7926 (vectorize_loops): Handle simd hints and adjust simd builtins
7927 accordingly.
7928 * tree-vectorizer.h (struct _stmt_vec_info): Add
7929 simd_lane_access_p field.
7930 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7931 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7932 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7933 (omp_clause_code_name): Same.
7934 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7935 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7936 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 7937 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7938 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 7939 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7940 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7941 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7942 (OMP_CLAUSE_LINEAR_STEP): New.
7943 (OMP_CLAUSE_SAFELEN_EXPR): New.
7944 (OMP_CLAUSE__SIMDUID__DECL): New.
7945 (find_omp_clause): New prototype.
7946
d0cf4e84
L		 79472013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7948
7949 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7950 Haswell processor detection.
7951
992592ec
CW		 79522013-08-27 Christian Widmer <shadow@umbrox.de>
7953
7954 PR target/57927
7955 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7956 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7957 AVX2 capable processors.
7958
9116eb22
TB		 79592013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7960
7961 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7962 of vget_low_* with implementations in terms of other intrinsics.
7963
1c5abb53
MG		 79642013-08-27 Marc Glisse <marc.glisse@inria.fr>
7965
7966 PR middle-end/57219
7967 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7968 values to -1, 0 and 1.
7969
2e100703
VP		 79702013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7971
7972 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7973 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7974 (<optab><mode>3_insn): Remove.
7975 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7976 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7977 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7978 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7979 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7980 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7981 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7982 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7983 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7984 (ror<mode>3_insn): Likewise.
7985 * config/aarch64/predicates.md (aarch64_simd_register): New.
7986
4ded8276
RB		 79872013-08-27 Richard Biener <rguenther@suse.de>
7988
7989 PR tree-optimization/57521
7990 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7991 one edge is non-critical.
7992 (find_phi_replacement_condition): Make sure to use a non-critical
7993 edge. Cleanup and remove old bug workarounds.
7994 (bb_postdominates_preds): Remove.
7995 (if_convertible_loop_p_1): Do not compute post-dominators.
7996 (combine_blocks): Do not free post-dominators.
7997 (main_tree_if_conversion): Likewise.
7998 (pass_data_if_conversion): Add TODO_verify_ssa.
7999
5aa11061
DD		 80002013-08-27 DJ Delorie <dj@redhat.com>
8001
8002 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8003
fd91cfe3
YZ		 80042013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8005
8006 * function.c (assign_parm_find_data_types): Set passed_mode and
8007 nominal_mode to the TYPE_MODE of nominal_type for the built
8008 pointer type in case of the struct-pass-by-reference.
8009
907555ce
JR		 80102013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8011
8012 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8013 (UINT16_TYPE): Change default to "unsigned int".
8014
67518c93
JR		 8015 * config/avr/avr.opt (mfract-convert-truncate): New option.
8016 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8017 is set, round negative fractional integers according to n1169
8018 when converting to integer types.
8019
d6d989fc
JH		 80202013-08-26 Jan Hubicka <jh@suse.cz>
8021
8022 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8023 methods can not be called indirectly when their address is not taken.
8024
2aa3da06
JH		 80252013-08-26 Jan Hubicka <jh@suse.cz>
8026
12211b99
UB		 8027 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8028 ctor_for_folding.
2aa3da06 8029
0987ffe7
JH		 80302013-08-26 Jan Hubicka <jh@suse.cz>
8031
8032 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8033 can be unshared.
8034
befe8647
JR		 80352013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8036
8037 * reload.c (find_valid_class): Allow classes that do not include
8038 FIRST_PSEUDO_REGISTER - 1.
8039
a21e735e
JH		 80402013-08-26 Jan Hubicka <jh@suse.cz>
8041
8042 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8043 fix edge count/frequency when speculation failed; fix type check
8044 for the direct call.
8045
e067bd43
JH		 80462013-08-26 Jan Hubicka <jh@suse.cz>
8047
8048 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8049
e8aec975
JH		 80502013-08-26 Jan Hubicka <jh@suse.cz>
8051
8052 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8053 to be changed by edge redirection.
8054
d0b66480
JH		 80552013-08-26 Jan Hubicka <jh@suse.cz>
8056
12211b99
UB		 8057 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8058 formating; add sanity check.
d0b66480
JH		 8059 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8060 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8061 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8062 (resolve_noninline_speculation): Update callee keys, too.
8063
0f9aaac7
JH		 80642013-08-26 Jan Hubicka <jh@suse.cz>
8065
8066 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8067 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8068
4b37444e
JR		 80692013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8070
8071 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8072 into proper place.
8073
5339fc0c
UB		 80742013-08-26 Uros Bizjak <ubizjak@gmail.com>
8075
8076 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8077 (x86_64_elf_select_section): Ditto.
8078 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8079 arguments.
8080 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 8081 (output_set_got): Ditto.
8082 (ix86_unary_operator_ok): Ditto.
8083 (ix86_expand_builtin): Ditto.
8084
d5c3d3ef
JH		 80852013-08-23 Jan Hubicka <jh@suse.cz>
8086
5339fc0c 8087 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 8088
5b1e7435
JH		 80892013-08-23 Jan Hubicka <jh@suse.cz>
8090
8091 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8092 (tree_decl_with_vis): Add FINAL field.
8093
ead69dac
JL		 80942013-08-23 Jeff Law <law@redhat.com>
8095
8096 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8097 the debugging dump when the expression is fully redundant.
8098
025311c4
GDR		 80992013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8100
8101 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8102 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8103 * pretty-print.c (pp_formatted_text_data): Likewise.
8104 (pp_write_text_to_stream): Likewise.
8105 (pp_write_text_as_dot_label_to_stream): Likewise.
8106 (pp_append_r): Likewise.
8107 (pp_format): Likewise.
8108 (pp_flush): Likewise.
8109 (pp_clear_output_area): Likewise.
8110 (pp_append_text): Likewise.
8111 (pp_formatted_text): Likewise.
8112 (pp_remaining_character_count_for_line): Likewise.
8113 (pp_newline): Likewise.
8114 (pp_character): Likewise.
8115 (output_buffer::~output_buffer): Define.
8116 (pretty_printer::~pretty_printer): Destruct output buffer.
8117 * pretty-print.h (output_buffer::~output_buffer): Declare.
8118 (pretty_printer::~pretty_printer): Declare virtual.
8119
0cadbfaa
MG		 81202013-08-24 Marc Glisse <marc.glisse@inria.fr>
8121
8122 PR other/57324
8123 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8124 HOST_WIDE_INT_M1U): New macros.
8125 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8126 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8127 unsigned -1 for lshift.
8128 * cse.c (cse_insn): Likewise.
8129 * double-int.c (rshift_double, lshift_double): Likewise.
8130 * builtins.c (fold_builtin_bitop): Likewise.
8131 * combine.c (force_to_mode): Likewise.
8132 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8133 * simplify-rtx.c (simplify_const_unary_operation,
8134 simplify_const_binary_operation): Likewise.
8135 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8136 check_va_list_escapes): Likewise.
8137 * rtlanal.c (nonzero_bits1): Likewise.
8138 * expmed.c (expand_smod_pow2): Likewise.
8139 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8140
009e5353
JH		 81412013-08-23 Jan Hubicka <jh@suse.cz>
8142
8143 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8144 as having address taken.
8145
c4be6568
JH		 81462013-08-23 Jan Hubicka <jh@suse.cz>
8147
64cbf23d
JH		 8148 * ipa-utils.h (method_class_type): Declare.
8149 * ipa-devirt.c (method_class_type): Export.
8150
c4be6568
JH		 8151 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8152 do not walk base classes of anonymous types.
8153
97aba8e9
KK		 81542013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8155
8156 PR rtl-optimization/58220
8157 PR regression/58221
8158 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8159 handle SEQUENCE insns properly.
8160
0fc80001
GDR		 81612013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8162
8163 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8164 definition.
8165 (pp_newline_and_indent): Likewise.
8166 (pp_separate_with): Likewise.
8167 * pretty-print.c (pp_newline_and_flush): Define.
8168 (pp_newline_and_indent): Likewise.
8169 (pp_separate_with): Likewise.
8170
520a5868
JJ		 81712013-08-23 Jakub Jelinek <jakub@redhat.com>
8172
8173 PR target/58218
8174 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8175 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8176
e500c62a
KY		 81772013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8178
f43245d1
UB		 8179 * config/i386/predicates.md (ext_sse_reg_operand): New.
8180 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY		 8181 predicate to determine if EVEX is needed.
8182 (*movsi_internal): Ditto.
8183 (*movdf_internal): Ditto.
8184 (*movsf_internal): Ditto.
f43245d1 8185 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 8186
6b00d7dd
JJ		 81872013-08-23 Jakub Jelinek <jakub@redhat.com>
8188
8189 PR tree-optimization/58209
8190 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8191 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8192 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8193 emit POINTER_PLUS_EXPR.
8194 (create_tailcall_accumulator): For pointer result type accumulate in
8195 sizetype type.
8196
4f2a9f90
PC		 81972013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8198
8199 * configure.ac: Add backslashes missing from the last change.
8200 * configure: Regenerate.
8201
7bf4274e
JH		 82022013-08-22 Jan Hubicka <jh@susue.cz>
8203
4f2a9f90
PC		 8204 * ipa.c (function_and_variable_visibility): First remember function
8205 was global and then make it local.
7bf4274e 8206
1bea243d
JB		 82072013-08-22 Julian Brown <julian@codesourcery.com>
8208
8209 * configure.ac: Add aarch64 to list of arches which use "nop" in
8210 debug_line test.
8211 * configure: Regenerate.
8212
bcba7ecc
AK		 82132013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8214
8215 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8216 gnu_libc_has_function.
8217 * config/s390/tpf.h: Likewise.
8218
4c495b0d
JH		 82192013-08-22 Jan Hubicka <jh@susue.cz>
8220
8221 * timevar.c (validate_phases): Add cast.
8222
5e302bca
JH		 82232013-08-22 Jan Hubicka <jh@susue.cz>
8224
8225 * timevar.c (validate_phases): Use size_t for memory.
8226 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8227
da6ca2b5
GDR		 82282013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8229
8230 * pretty-print.h (output_buffer::output_buffer): Declare.
8231 (pretty_printer::pretty_printer): Likewise.
8232 (pp_construct): Remove.
8233 * pretty-print.c (output_buffer::output_buffer): Define.
8234 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8235 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8236 (print_gimple_expr): Likewise.
8237 (print_gimple_seq): Likewise.
8238 (gimple_dump_bb): Likewise.
8239 * sched-vis.c (dump_value_slim): Likewise.
8240 (dump_insn_slim): Likewise.
8241 (dump_rtl_slim): Likewise.
8242 (str_pattern_slim): Likewise.
8243 * tree-mudflap.c (mf_varname_tree): Likewise.
8244 * graph.c (print_graph_cfg): Likewise.
8245 (start_graph_dump): Likewise.
8246 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8247 placement-new.
8248 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8249 pretty printer initialization.
8250 * coretypes.h (diagnostic_context): Remove superflous type alias
8251 declaration.
8252 (pretty_printer): Likewise. Declare directly as a class.
8253 (pretty_print_info): Remove declaration as class.
8254 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8255 and pp_clear_output_area.
8256 (asan_add_global): Likewise.
8257
0e1474e5
JH		 82582013-08-22 Jan Hubicka <jh@suse.cz>
8259
8260 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8261 * ipa-utils.h (update_type_inheritance_graph): Declare.
8262 (possible_polymorphic_call_target_p): Declare.
8263 (possible_polymorphic_call_target_p): New.
8264 * ipa-devirt.c: Update toplevel comments.
8265 (cached_polymorphic_call_targets): Move up.
8266 (odr_type_d): Move ID down.
8267 (polymorphic_type_binfo_p): Update comment.
8268 (odr_hasher::remove): Likewise;
8269 (get_odr_type): Set anonymous_namespace.
8270 (dump_odr_type): Dump it.
8271 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8272 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8273 (record_binfo): Add comment.
5339fc0c
UB		 8274 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8275 built.
0e1474e5
JH		 8276 (devirt_node_removal_hook): Do not iCE when cache is freed.
8277 (possible_polymorphic_call_target_p): New predicate.
8278 (update_type_inheritance_graph): New function.
8279
3f97cb0b
AI		 82802013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8281 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8282 Sergey Lega <sergey.s.lega@intel.com>
8283 Anna Tikhonova <anna.tikhonova@intel.com>
8284 Ilya Tocar <ilya.tocar@intel.com>
8285 Andrey Turetskiy <andrey.turetskiy@intel.com>
8286 Ilya Verbin <ilya.verbin@intel.com>
8287 Kirill Yukhin <kirill.yukhin@intel.com>
8288 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8289
8290 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8291 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8292 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8293 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8294 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8295 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8296 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8297 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8298 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8299 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8300 OPT_mavx512pf, OPT_mavx512er cases.
8301 * config/i386/constraints.md (v): New constraint.
8302 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8303 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8304 (bit_AVX512CD): New.
8305 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8306 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8307 * config/i386/i386-c.c (ix86_target_macros_internal):
8308 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8309 __AVX512PF__.
8310 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8311 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8312 * config/i386/i386.c (regclass_map, dbx_register_map)
8313 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8314 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8315 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8316 -mavx512pf options.
8317 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8318 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8319 -mavx512pf options. Fix formatting.
8320 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8321 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8322 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8323 -mavx512cd, -mavx512pf options.
8324 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8325 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8326 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8327 ALL_SSE_REGS.
8328 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8329 (ix86_set_reg_reg_cost): Ditto.
8330 (x86_order_regs_for_local_alloc): Ditto.
8331 (MAX_VECT_LEN): Extend to 64-byte.
8332 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8333 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8334 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8335 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8336 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8337 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8338 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8339 (SSE_REG_MODE_P): Support new modes.
8340 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8341 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8342 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8343 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8344 (REG_CLASS_CONTENTS): Add new registers.
8345 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8346 (EXT_REX_SSE_REGNO_P): New.
8347 (HI_REGISTER_NAMES): Add new registers.
8348 * config/i386/i386.md: Define constants for new registers.
8349 (mode): Add new 512-bit modes.
8350 (prefix): Support evex prefix.
8351 (isa): Support avx512f, noavx512f, fma_avx512f.
8352 (ssemodesuffix): Add new 512-bit modes.
8353 (movxi): New.
8354 (*movxi_internal_avx512f): Ditto.
8355 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8356 Support MODE_XI.
8357 (*movsi_internal): Likewise.
8358 (*movdf_internal): Likewise.
8359 (*movsf_internal): Likewise.
8360 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8361 (<code><mode>3): Likewise.
5339fc0c
UB		 8362 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8363 New.
3f97cb0b
AI		 8364 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8365 with the new constraint "v".
8366 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8367 modes.
8368 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8369 with the new constraint "v".
8370 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8371 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8372 (*<plusminus_insn><mode>3): Likewise.
8373 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8374 (*mul<mode>3): Likewise.
8375 (<sse>_vmmul<mode>3): Likewise.
8376 (<sse>_div<mode>3): Likewise.
8377 (<sse>_vmdiv<mode>3): Likewise.
8378 (<sse>_sqrt<mode>2): Likewise.
8379 (<sse>_vmsqrt<mode>2): Likewise.
8380 (*<code><mode>3_finite): Likewise.
8381 (*<code><mode>3) <smaxmin>: Likewise.
8382 (<sse>_vm<code><mode>3): Likewise.
8383 (*<code><mode>3) <any_logic>: Likewise.
8384 (*fma_fmadd_<mode>): Likewise.
8385 (*fma_fmsub_<mode>): Likewise.
8386 (*fma_fnmadd_<mode>): Likewise.
8387 (*fma_fnmsub_<mode>): Likewise.
8388 (*fma_fmaddsub_<mode>): Likewise.
8389 (*fma_fmsubadd_<mode>): Likewise.
8390 (*fmai_fmadd_<mode>): Likewise.
8391 (*fmai_fmsub_<mode>): Likewise.
8392 (*fmai_fnmadd_<mode>): Likewise.
8393 (*fmai_fnmsub_<mode>): Likewise.
8394 (sse_cvtsi2ss): Likewise.
8395 (sse_cvtsi2ssq): Likewise.
8396 (sse_cvtss2si): Likewise.
8397 (sse_cvtss2si_2): Likewise.
8398 (sse_cvtss2siq): Likewise.
8399 (sse_cvtss2siq_2): Likewise.
8400 (sse_cvttss2si): Likewise.
8401 (sse_cvtss2siq_2): Likewise.
8402 (float<sseintvecmodelower><mode>2): Likewise.
8403 (sse2_cvtsd2si_2): Likewise.
8404 (sse2_cvtsd2siq_2): Likewise.
8405 (*<plusminus_insn><mode>3): Likewise.
8406 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8407 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8408 (ashr<mode>3): Likewise.
8409 (<shift_insn><mode>3): Likewise.
8410 (avx2_<code><mode>3): Likewise.
8411 (*avx2_<code><mode>3): Likewise.
8412 (*andnot<mode>3): Likewise.
8413 (*<code><mode>3) <any_logic>: Likewise.
8414 (abs<mode>2): Likewise.
8415 (avx2_permvar<mode>): Likewise.
8416 (avx2_perm<mode>_1): Likewise.
8417 (*avx_vpermilp<mode>): Likewise.
8418 (avx_vpermilvar<mode>3): Likewise.
8419 (avx2_ashrv<mode>): Likewise.
8420 (avx2_<shift_insn>v<mode>): Likewise.
8421 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8422 -mavx512cd.
8423 * doc/rtl.texi: Document XImode.
8424
dd5e8423
JL		 84252013-08-21 Jeff Law <law@redhat.com>
8426
b9ebee5d
JL		 8427 * tree-flow.h (register_jump_thread): Pass vector of edges
8428 instead of each important edge.
8429 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8430 thread path into a vector and pass that to register_jump_thread.
8431 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8432 passed in edge vector to the current 3-edge form.
8433
dd5e8423
JL		 8434 Revert:
8435 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8436
9bb6628e 8437 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 8438 delete_unmarked_insns removed anything.
8439
445dc8df
JR		 84402013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8441
8442 * reload.h (struct reg_equivs): Rename to ..
8443 (struct reg_equivs_s): .. this.
8444
e0df53dd
ML		 84452013-08-20 Martin Liska <marxin.liska@gmail.com>
8446
8447 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8448
e7d1d3eb
RO		 84492013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8450
5339fc0c 8451 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8452
5562e26e
JL		 84532013-08-21 Jeff Law <law@redhat.com>
8454
8455 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8456 simplify assignments too. If the RHS collapses to a singleton
8457 range, then return the value for the range.
8458
c7ecdec6
KY		 84592013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8460
7a5911d3
UB		 8461 * config/i386/sse.md (V16): Rename to...
8462 (VMOVE): this.
8463 (mov<mode>): Update iterator name.
8464 (*mov<mode>_internal): Ditto.
8465 (push<mode>1): Ditto.
8466 (movmisalign<mode>): Ditto.
c7ecdec6 8467
bfa3b50a
JH		 84682013-08-20 Jan Hubicka <jh@suse.cz>
8469
8470 PR bootstrap/58186
8471 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8472 entry for direct edges.
8473 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8474
6cd4d135
DM		 84752013-08-20 David Malcolm <dmalcolm@redhat.com>
8476
8477 Revert my last two changes, r201865 and r201864:
8478
8479 Revert r201865:
8480 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8481
8482 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8483 instances can own GC refs.
8484
8485 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8486 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8487 (gcc::context::gt_pch_nx): Likewise.
8488 (gcc::context::gt_pch_nx): Likewise.
8489 * ggc.h (gt_ggc_mx <T>): New.
8490 (gt_pch_nx_with_op <T>): New.
8491 (gt_pch_nx <T>): New.
8492 * passes.c (opt_pass::gt_ggc_mx): New.
8493 (opt_pass::gt_pch_nx): New.
8494 (opt_pass::gt_pch_nx_with_op): New.
8495 (pass_manager::gt_ggc_mx): New.
8496 (pass_manager::gt_pch_nx): New.
8497 (pass_manager::gt_pch_nx_with_op): New.
8498 (pass_manager::operator new): Use
8499 ggc_internal_cleared_alloc_stat rather than xcalloc.
8500 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8501 (pass_manager::gt_ggc_mx): New.
8502 (pass_manager::gt_pch_nx): New.
8503 (pass_manager::gt_pch_nx_with_op): New.
8504 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8505 (opt_pass::operator new): New.
8506 (opt_pass::gt_ggc_mx): New.
8507 (opt_pass::gt_pch_nx): New.
8508 (opt_pass::gt_pch_nx_with_op): New.
8509
8510 Revert r201864:
8511 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8512
8513 * Makefile.in (GTFILES): Add context.h.
8514 * context.c (gcc::context::operator new): New.
8515 (gcc::context::gt_ggc_mx): New.
8516 (gcc::context::gt_pch_nx): New.
8517 (gcc::context::gt_pch_nx): New.
8518 * context.h (gcc::context): Add GTY((user)) marking.
8519 (gcc::context::operator new): New.
8520 (gcc::context::gt_ggc_mx): New.
8521 (gcc::context::gt_pch_nx): New.
8522 (gcc::context::gt_pch_nx): New.
8523 (g): Add GTY marking.
8524 (gt_ggc_mx (gcc::context *)): New.
8525 (gt_pch_nx (gcc::context *)): New.
8526 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8527 void *cookie)): New.
8528 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8529
a046136a
AM		 85302013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8531
8532 * dce.c (fini_dce): Call df_analyze again just in case
8533 delete_unmarked_insns removed anything.
8534
67598720
TJ		 85352013-08-20 Teresa Johnson <tejohnson@google.com>
8536
8537 PR rtl-optimizations/57451
8538 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8539 from crossing split section boundaries.
8540
f8693faf
MGD		 85412013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8542
8543 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8544 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8545 with MULTLIB_DEFAULTS.
8546
7d0b9a9c
NC		 85472013-08-20 Nick Clifton <nickc@redhat.com>
8548
8549 * target.def (narrow_volatile_bitfield): Note that the default
8550 value is false, not !TARGET_STRICT_ALIGN.
8551 * doc/tm.texi: Regenerate.
8552
8edb8dc8
PC		 85532013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8554
8555 Fix LIB_SPEC for systems without libpthread.
8556
8557 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8558 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8559 for Android.
8560 * config/i386/linux-common.h: Likewise.
8561 * config/mips/linux-common.h: Likewise.
8562
12211b99 85632013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 8564
8565 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8566 checks.
8567
e42c64cb
DM		 85682013-08-20 David Malcolm <dmalcolm@redhat.com>
8569
8570 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8571 instances can own GC refs.
8572
8573 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8574 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8575 (gcc::context::gt_pch_nx): Likewise.
8576 (gcc::context::gt_pch_nx): Likewise.
8577 * ggc.h (gt_ggc_mx <T>): New.
8578 (gt_pch_nx_with_op <T>): New.
8579 (gt_pch_nx <T>): New.
8580 * passes.c (opt_pass::gt_ggc_mx): New.
8581 (opt_pass::gt_pch_nx): New.
8582 (opt_pass::gt_pch_nx_with_op): New.
8583 (pass_manager::gt_ggc_mx): New.
8584 (pass_manager::gt_pch_nx): New.
8585 (pass_manager::gt_pch_nx_with_op): New.
8586 (pass_manager::operator new): Use
8587 ggc_internal_cleared_alloc_stat rather than xcalloc.
8588 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8589 (pass_manager::gt_ggc_mx): New.
8590 (pass_manager::gt_pch_nx): New.
8591 (pass_manager::gt_pch_nx_with_op): New.
8592 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8593 (opt_pass::operator new): New.
8594 (opt_pass::gt_ggc_mx): New.
8595 (opt_pass::gt_pch_nx): New.
8596 (opt_pass::gt_pch_nx_with_op): New.
8597
910c02a0
DM		 85982013-08-20 David Malcolm <dmalcolm@redhat.com>
8599
8600 * Makefile.in (GTFILES): Add context.h.
8601 * context.c (gcc::context::operator new): New.
8602 (gcc::context::gt_ggc_mx): New.
8603 (gcc::context::gt_pch_nx): New.
8604 (gcc::context::gt_pch_nx): New.
8605 * context.h (gcc::context): Add GTY((user)) marking.
8606 (gcc::context::operator new): New.
8607 (gcc::context::gt_ggc_mx): New.
8608 (gcc::context::gt_pch_nx): New.
8609 (gcc::context::gt_pch_nx): New.
8610 (g): Add GTY marking.
8611 (gt_ggc_mx (gcc::context *)): New.
8612 (gt_pch_nx (gcc::context *)): New.
8613 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8614 void *cookie)): New.
8615 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8616
af4b4236
AM		 86172013-08-20 Alan Modra <amodra@gmail.com>
8618
8619 PR target/57865
8620 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8621 (rs6000_emit_epilogue): Likewise.
8622
12211b99 86232013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 8624
8625 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8626
e2323f5b
PB		 86272013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8628 Jakub Jelinek <jakub@redhat.com>
8629
8630 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8631 (BUILT_IN_FABSD64): Likewise.
8632 (BUILT_IN_FABSD128): Likewise.
8633 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8634 (fold_builtin_1): Likewise.
5339fc0c
UB		 8635 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8636 destination and source operands.
e2323f5b
PB		 8637 (*abstd2_fpr): Likewise.
8638 (*nabstd2_fpr): Likewise.
8639
cfbf3ee8
RS		 86402013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8641
8642 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8643 JUMP_P and INSN_P.
8644
d3136aeb
AH		 86452013-08-19 Aldy Hernandez <aldyh@redhat.com>
8646
8647 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8648 incomplete.
8649
d33d9e47
AI		 86502013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8651
8652 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8653 * builtins.c (default_libc_has_function): New.
8654 (gnu_libc_has_function): Ditto.
8655 (no_c99_libc_has_function): Ditto.
8656 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8657 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8658 (fold_builtin_sincos): Likewise.
8659 (fold_builtin_cexp): Likewise.
8660 * builtins.def (DEF_C94_BUILTIN): Likewise.
8661 (DEF_C99_BUILTIN): Likewise.
8662 (DEF_C99_C90RES_BUILTIN): Likewise.
8663 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8664 definitions to using this define.
8665 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 8666 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 8667 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8668 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8669 * config/darwin.h: Ditto.
8670 * config/elfos.h: Ditto.
8671 * config/freebsd.h: Ditto.
8672 * config/i386/cygming.h: Ditto.
8673 * config/i386/djgpp.h: Ditto.
8674 * config/i386/i386-interix.h: Ditto.
8675 * config/microblaze/microblaze.h: Ditto.
8676 * config/mmix/mmix.h: Ditto.
8677 * config/gnu-user.h: Ditto.
8678 * config/ia64/hpux.h: Ditto.
8679 * config/pa/pa-hpux.h: Ditto.
8680 * config/pdp11/pdp11.h: Ditto.
8681 * config/picochip/picochip.h: Ditto.
8682 * config/linux.h: Ditto.
8683 * config/netbsd.h: Ditto.
8684 * config/openbsd.h: Ditto.
8685 * config/rs6000/aix43.h: Ditto.
8686 * config/rs6000/aix51.h: Ditto.
8687 * config/rs6000/aix52.h: Ditto.
8688 * config/rs6000/aix53.h: Ditto.
8689 * config/rs6000/aix61.h: Ditto.
8690 * config/rs6000/darwin.h: Ditto.
8691 * config/rs6000/linux.h: Ditto.
8692 * config/rs6000/linux64.h: Ditto.
8693 * config/s390/tpf.h: Ditto.
8694 * config/sol2-10.h: Ditto.
8695 * config/sol2.h: Ditto.
8696 * config/vms/vms.h: Ditto.
8697 * config/vxworks.h: Ditto.
8698 * config/linux-android.c (linux_android_libc_has_function):
8699 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8700 * config/linux-protos.h (linux_android_libc_has_function):
8701 New declaration.
8702 * config/i386/i386.c (ix86_libc_has_function): New.
8703 * config/i386/i386-protos.h
8704 (ix86_libc_has_function): New declaration.
8705 * config/i386/i386.md
8706 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8707 ("isinf<mode>2): Likewise.
8708 * convert.c (convert_to_integer): Using new target hook
8709 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8710 TARGET_C99_FUNCTIONS.
8711 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8712 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8713 * coretypes.h (function_class): New enum for different
8714 classes of functions.
8715 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8716 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8717 (TARGET_HAS_SINCOS): Likewise.
8718 (TARGET_LIBC_HAS_FUNCTION): New.
8719 * doc/tm.texi: Regenerated.
8720 * targhooks.h (default_libc_has_function): New declaration.
8721 (no_c99_libc_has_function): Ditto.
8722 (gnu_libc_has_function): Ditto.
8723 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8724 and TARGET_HAS_SINCOS.
8725
eefe9a99
JH		 87262013-08-18 Jan Hubicka <jh@suse.cz>
8727
8728 * Makeifle-in (ipa-devirt.o): New.
8729 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8730 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8731 (analyze_functions): Look into possible targets of polymorphic call.
8732 * dumpfile.c (dump_files): Add type-inheritance dump.
8733 * dumpfile.h (TDI_inheritance): New.
8734 * ipa-devirt.c: New file.
8735 * ipa-utils.h (odr_type_d): Forward declare.
8736 (odr_type): New type.
8737 (build_type_inheritance_graph): Declare.
8738 (possible_polymorphic_call_targets): Declare and introduce inline
8739 variant when only edge is pased.
8740 (dump_possible_polymorphic_call_targets): Likewise.
8741 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8742 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8743 (types_same_for_odr): ... here.
8744 * tree.h (type_in_anonymous_namespace_p): Declare.
8745
29b89442
JJ		 87462013-08-18 Jakub Jelinek <jakub@redhat.com>
8747
8748 PR tree-optimization/58006
8749 * tree-parloops.c (take_address_of): Don't ICE if get_name
8750 returns NULL.
8751 (eliminate_local_variables_stmt): Remove clobber stmts.
8752
4f219961
EB		 87532013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8754
8755 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8756 error message is issued for an alias to undefined symbol.
8757
1d5755ef
JH		 87582013-08-18 Jan Hubicka <jh@suse.cz>
8759
8760 * cgraph.c (cgraph_create_indirect_edge): Discover
8761 polymorphic calls and record basic info into indirect_info.
8762 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8763 devirtualization, ignore objc function calls.
8764 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8765 call with no parm index info.
8766 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8767 * tree.c (virtual_method_call_p): New function.
8768 * tree.h (virtual_method_call_p): Declare.
8769
5945bebf
JH		 87702013-08-16 Jan Hubicka <jh@suse.cz>
8771
8772 PR middle-end/58179
8773 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8774
47fa5d23
DE		 87752013-08-16 David Edelsohn <dje.gcc@gmail.com>
8776
8777 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8778 attribute.
8779
c74f54a0
DM		 87802013-08-16 David Malcolm <dmalcolm@redhat.com>
8781
8782 * gengtype.c (type_for_name): Add special-case support for
8783 locating types within the "gcc::" namespace.
8784 (open_base_files): Emit a "using namespace gcc" directive.
8785
f758f299
MM		 87862013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8787
8788 PR target/58160
8789 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8790 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8791
8792 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8793 array instead of each individual operand as a separate argument.
8794 (emit_fusion_gpr_load): Likewise.
8795 (expand_fusion_gpr_load): Add new function declaration.
8796
8797 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8798 signature to have the operands passed as an array, instead of as
8799 separate arguments. Allow ZERO_EXTEND to be in the memory
8800 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8801 depend on the register live/dead flags when peepholes are run.
8802 (expand_fusion_gpr_load): New function to be called from the
8803 peephole2 pass, to change the register that addis sets to be the
8804 target register.
8805 (emit_fusion_gpr_load): Change the calling signature to have the
8806 operands passed as an array, instead of as separate arguments.
8807 Allow ZERO_EXTEND to be in the memory address, and also
8808 SIGN_EXTEND if -mpower8-fusion-sign.
8809
8810 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8811 unspec enumeration.
8812 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8813 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 8814 depend on the register live/dead state when the peephole pass is done.
f758f299 8815
158f4e4f
DM		 88162013-08-16 David Malcolm <dmalcolm@redhat.com>
8817
8818 * gengtype.c (create_user_defined_type): Ensure that the kind
8819 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8820 declaration is seen before the GTY((user)) marking.
8821
22f8cea5
BE		 88222013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8823
8824 PR target/58105
8825 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8826
c49bdb2e
JH		 88272013-08-16 Jan Hubicka <jh@suse.cz>
8828
8829 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8830 arugment expected_type.
8831 (gimple_fold_call): Use it.
8832 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8833 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 8834 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 8835 (try_make_edge_direct_virtual_call): Likewise.
8836 * tree.c (obj_type_ref_class): New.
8837 * tree.h (obj_type_ref_class): Use it.
8838
4042dca9
GDR		 88392013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8840
8841 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8842 (rtl_slim_pp): Likewise.
8843 (init_rtl_slim_pretty_print): Likewise.
8844 (dump_value_slim): Don't call it. Use local pretty printer.
8845 (dump_insn_slim): Likewise.
8846 (dump_rtl_slim): Likewise.
8847 (str_pattern_slim): Likewise.
8848 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8849 Simplify.
8850
7a460594
JJ		 88512013-08-16 Jakub Jelinek <jakub@redhat.com>
8852
639dc669
JJ		 8853 PR tree-optimization/58164
8854 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8855 walk gimple_goto_dest of GIMPLE_GOTO.
8856
7a460594
JJ		 8857 PR tree-optimization/58165
8858 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8859 bi_call must be the last stmt in a bb, don't split_block, instead
8860 use fallthru edge from it and give up if there is none.
8861 Release conds vector when returning early.
8862
e3c02cc7
XDL		 88632013-08-14 Xinliang David Li <davidxl@google.com>
8864
8865 * config/i386/i386.c (ix86_option_override_internal):
8866 Remove unused variable and field.
8867
6469da2c
BS		 88682013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8869
8870 PR target/57949
22f8cea5 8871 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 8872 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8873 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8874 and Linux, correct BLKmode alignment when 128-bit alignment is
8875 required and compatibility flag is not set.
22f8cea5
BE		 8876 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8877 for zero-size arguments when compatibility flag is not set.
6469da2c 8878
4ca890e2
JJ		 88792013-08-14 Jakub Jelinek <jakub@redhat.com>
8880
8881 PR tree-optimization/58145
8882 * tree-sra.c (build_ref_for_offset): If prev_base has
8883 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8884
daec1759 88852013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 8886
daec1759 8887 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 8888 Fix uninitialized variable error.
daec1759 8889
3ad20bd4
XDL		 88902013-08-14 Xinliang David Li <davidxl@google.com>
8891
8892 * config/i386/i386.opt: Define two new options.
8893 * config/i386/x86-tune.def: Add arch selector field in macros.
8894 * config/i386/i386.h: Adjust macro definition.
8895 * config/i386/i386.c (ix86_option_override_internal):
8896 Refactor the code.
8897 (parse_mtune_ctrl_str): New function.
8898 (set_ix86_tune_features): New function.
8899 (ix86_function_specific_restore): Call the new helper function.
8900
e839e2a9
AB		 89012013-08-14 Andrey Belevantsev <abel@ispras.ru>
8902
8903 PR rtl-optimization/57662
8904 * sel-sched.c (code_motion_process_successors): When the current insn
8905 is removed after the recursive traversal, break from the loop.
8906 Add comments and debug printouts.
8907
27ed665f
JJ		 89082013-08-14 Jakub Jelinek <jakub@redhat.com>
8909 Alexandre Oliva <aoliva@redhat.com>
8910
8911 PR target/58067
8912 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8913 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8914 there also UNSPEC_PLTOFF.
8915
61c1a609
MP		 89162013-08-14 Marek Polacek <polacek@redhat.com>
8917
8918 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8919 NUM_CONDITIONS bit positions.
8920
0e901c67
CC		 89212013-08-13 Cary Coutant <ccoutant@google.com>
8922
8923 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8924 (attr_checksum): Hash vector contents instead of pointer.
8925 (attr_checksum_ordered): Likewise.
8926
061eff6d
UB		 89272013-08-13 Uros Bizjak <ubizjak@gmail.com>
8928
8929 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8930 when Pmode != word_mode. Add length_address attribute.
8931 (sse3_monitor_<mode>): Merge from sse3_monitor and
8932 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8933 Pmode != word_mode. Update insn length attribute.
8934 * config/i386/i386.c (ix86_option_override_internal): Update
8935 ix86_gen_monitor selection for merged sse3_monitor insn.
8936
8ec9249b
JB		 89372013-08-13 Julian Brown <julian@codesourcery.com>
8938
8939 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8940 perform invalid legitimization on greater-than-word-size modes for
8941 TARGET_E500_DOUBLE.
8942
761a8eb7
VM		 89432013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8944
8945 * ira.c (setup_class_translate_array): Use aclass instead of cl
8946 for classes not fully covered by allocno classes.
8947
99c2bd54
JJ		 89482013-08-13 Jakub Jelinek <jakub@redhat.com>
8949
4029a5e0
JJ		 8950 PR tree-optimization/57661
8951 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8952 * tree-inline.c (tree_function_versioning): Initialize it.
8953 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8954 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8955 that is not being copied.
8956
99c2bd54
JJ		 8957 PR sanitizer/56417
8958 * asan.c (instrument_strlen_call): Fix typo in comment.
8959 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8960
b5f8f063
SE		 89612013-08-13 Steve Ellcey <sellcey@mips.com>
8962
8963 * config/mips/mips.md (prefetch): Use lw instead of ld on
8964 loongson in 32bit mode.
8965
4de80584
NC		 89662013-08-13 Nick Clifton <nickc@redhat.com>
8967
8968 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8969
09ce3660
JH		 89702013-08-13 Jan Hubicka <jh@suse.cz>
8971
8972 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8973 introduced edge; fix typo in sanity check.
8974 (cgraph_resolve_speculation): Export; improve diagnostic.
8975 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8976 speculation at type mismatch.
8977 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8978 (cgraph_resolve_speculation): Declare.
8979 (symtab_can_be_discarded): New function.
8980 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8981 * ipa-inline-transform.c (speculation_removed): New global var.
8982 (clone_inlined_nodes): See if speculation can be removed.
8983 (inline_call): If speculations was removed, we growths may not match.
8984 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8985 (speculation_useful_p): New function.
8986 (resolve_noninline_speculation): New function.
8987 (inline_small_functions): Resolve useless speculations.
8988 * ipa-inline.h (speculation_useful_p): Declare
8989 * ipa.c (can_replace_by_local_alias): Simplify.
8990 (ipa_profile): Produce speculative calls in non-lto, too;
8991 add simple cost model; produce local aliases.
8992
f971dc24
DM		 89932013-08-13 David Malcolm <dmalcolm@redhat.com>
8994
8995 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 8996 PASS_MANAGER_H.
f971dc24 8997
2e507b9a
PC		 89982013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8999
9000 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9001
c97fd4b8
UB		 90022013-08-12 Uros Bizjak <ubizjak@gmail.com>
9003
9004 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9005 iterator instead of X87MODEF.
9006
90072013-08-12 Perez Read <netfirewall@gmail.com>
9008
9009 PR target/58132
9010 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9011 operand 0 for intel asm alternative.
9012 (*movabs<mode>_2): Ditto for operand 1.
9013
f4f4204c
JG		 90142013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9015
9016 * config/aarch64/arm_none.h
9017 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9018
133b1a8e
NC		 90192013-08-12 Nick Clifton <nickc@redhat.com>
9020
9021 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9022
15dd8b3a
YR		 90232013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9024
9025 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9026 expand for QI/HImode operand to produce more effictive code for
9027 unsigned char(short) --> float(double) conversion.
9028
94109a6a
AM		 90292013-08-12 Alexander Monakov <amonakov@ispras.ru>
9030
9031 * doc/invoke.texi: Mention that -ftls-model does not force the final
9032 model.
9033
4e115102
MP		 90342013-08-12 Marek Polacek <polacek@redhat.com>
9035 Marc Glisse <marc.glisse@inria.fr>
9036
9037 PR tree-optimization/57980
9038 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9039 when creating -1 constant.
9040
748d71f3
JH		 90412013-08-10 Jan Hubicka <jh@suse.cz>
9042
1f97e193 9043 Workaround binutils PR14342.
748d71f3
JH		 9044 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9045 (gimple_init_edge_profiler): Likewise.
9046 (gimple_gen_ic_func_profiler): Likewise.
9047
5979aa54
JH		 90482013-08-09 Jan Hubicka <jh@suse.cz>
9049
9050 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9051
ad83025e
XDL		 90522013-08-09 Xinliang David Li <davidxl@google.com>
9053
9054 * config/i386/stringop.def: New file.
9055 * config/i386/stringop.opt: New file.
9056 * config/i386/i386-opts.h: Include stringopt.def.
9057 * config/i386/i386.opt: Include stringopt.opt.
9058 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 9059 Override default size based stringop inline strategies with options.
ad83025e
XDL		 9060 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9061 New function.
9062
e59df5fd
JH		 90632013-08-09 Jan Hubicka <jh@suse.cz>
9064
c97fd4b8 9065 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 9066
634ab819
JH		 90672013-08-09 Jan Hubicka <jh@suse.cz>
9068
9069 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9070 CGRAPH_FREQ_MAX.
9071 (dump_cgraph_node): Dump profile-id.
9072 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9073 and common_target_probability.
9074 * lto-cgraph.c (lto_output_edge): Stream common targets.
9075 (lto_output_node): Stream profile ids.
9076 (input_node): Stream profile ids.
9077 (input_edge): Stream common targets.
9078 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9079 * ipa.c: Include value-prof.h
9080 (ipa_profile_generate_summary): Turn indirect call statement histograms
9081 into common targets.
9082 (ipa_profile): Turn common targets into speculative edges.
9083
2fa3d31b
JH		 90842013-08-09 Jan Hubicka <jh@suse.cz>
9085
9086 * cgraph.h (cgraph_node): Add profile_id.
9087 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 9088 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 9089 (del_node_map): Update.
9090 (find_func_by_funcdef_no): Replace by ...
9091 (find_func_by_profile_id): ... this one.
9092 (gimple_ic_transform): Do not remove useful histograms when
9093 speculation is not done; dump info when indirect call removal
9094 can happen at LTO.
9095 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9096 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9097 (__gcov_indirect_call_profiler_v2): .. this one.
9098 * profile.h (init_node_map): Update.
9099 * coverage.c (coverage_compute_profile_id): New function.
9100 * coverage.h (coverage_compute_profile_id): Declare.
9101 * tree-profile.c (init_ic_make_global_vars): Make
9102 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9103 (gimple_init_edge_profiler): Update prototype of
9104 __gcov_indirect_call_profiler.
9105 (gimple_gen_ic_func_profiler): Simplify.
9106 (tree_profiling): Use init_node_map
9107
042ae7d2
JH		 91082013-08-09 Jan Hubicka <jh@suse.cz>
9109
c97fd4b8
UB		 9110 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9111 non-speculative refs.
042ae7d2
JH		 9112 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9113 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9114 (cgraph_set_call_stmt): Likewise.
9115 (cgraph_create_edge_1): Fix release checking compilatoin;
9116 clear lto_stmt_uid.
9117 (cgraph_free_edge): Free indirect info.
9118 (cgraph_turn_edge_to_speculative): New function.
9119 (cgraph_speculative_call_info): New function.
9120 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 9121 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 9122 (dump_cgraph_node): Dump speculation.
9123 (verify_edge_count_and_frequency): Accept speculative edges.
9124 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9125 (verify_cgraph_node): Handle speculation.
9126 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9127 (cgraph_set_call_stmt): Update prototype.
9128 (cgraph_make_edge_direct): Update prototype.
9129 (cgraph_speculative_call_info): Declare.
9130 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9131 to change; update call of ipa_find_references.
9132 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9133 lto_stmt_uid and speculative flags.
9134 (ipa_dump_references): Dump speculation.
9135 (ipa_clone_references): Clone speculative flag.
9136 (ipa_clone_referring): Likewise.
9137 (ipa_clone_ref): New function.
9138 (ipa_find_reference): Look into lto_stmt_uids
9139 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9140 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9141 (ipa_find_reference): Update declaration.
9142 (ipa_clone_ref): Declare.
9143 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9144 stream speculative flag.
9145 (lto_output_ref): Stream statements uids and speculation.
9146 (input_ref): Likewise.
9147 (input_edge): Stream speuclation.
9148 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9149 (cgraph_set_call_stmt_including_clones): Handle speculation.
9150 * ipa-inline.c (heap_edge_removal_hook): New function.
9151 (inline_small_functions): Register it.
9152 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9153 also initialize refs.
9154 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9155 edge to change.
9156 (try_make_edge_direct_simple_call): Likewise.
9157 (try_make_edge_direct_simple_call): Likewise.
9158 (update_indirect_edges_after_inlining): Likewise.
9159 (remove_described_reference): Look proper lto_stmt_uid.
9160 (propagate_controlled_uses): Likewise.
9161 (propagate_controlled_uses): Liekwise.
9162 * tree-inline.c (copy_bb): Copy speculative edges.
9163 (redirect_all_calls): New function.
c97fd4b8
UB		 9164 (copy_cfg_body): Do redirection after loop info is updated.
9165 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 9166
aa1e10cc
JH		 91672013-08-09 Jan Hubicka <jh@suse.cz>
9168
9169 * lto-streamer-out.c (output_function): Renumber PHIs.
9170 * lto-streamer-in.c (input_function): Likewise.
9171
66adb8eb
JG		 91722013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9173
9174 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9175 (get_lane_unsigned): Likewise.
9176 (dup_lane_scalar): Likewise.
9177 (get_lane): enable for VALL.
9178 * config/aarch64/aarch64-simd.md
9179 (aarch64_dup_lane_scalar<mode>): Remove.
9180 (aarch64_get_lane_signed<mode>): Likewise.
9181 (aarch64_get_lane_unsigned<mode>): Likewise.
9182 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9183 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9184 (aarch64_get_lane<mode>): Enable for all vector modes.
9185 (aarch64_get_lanedi): Remove misleading constraints.
9186 * config/aarch64/arm_neon.h
9187 (__aarch64_vget_lane_any): Define.
9188 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9189 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9190 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9191 * config/aarch64/iterators.md (VDQQH): New.
9192 (VDQQHS): Likewise.
9193 (vwcore): Likewise.
9194
1f65ae7a
EB		 91952013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9196
9197 * configure.ac: Add GAS check for LEON instructions on SPARC.
9198 * configure: Regenerate.
9199 * config.in: Likewise.
9200 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9201 sparc*-*-* block.
9202 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9203 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9204 for LEON or LEON3.
9205 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9206 (AS_LEON_FLAG): New macro.
9207 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9208 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9209 Deal with LEON and LEON3 for the memory model.
23a6cb78 9210 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 9211 (atomic_compare_and_swap<mode>_1): Likewise.
9212 (*atomic_compare_and_swap<mode>_1): Likewise.
9213
ff522f7f
ZC		 92142013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9215
9216 * config/arm/neon.md (vcond): Fix floating-point vector
9217 comparisons against 0.
9218
cb1cca12
VM		 92192013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9220
9221 * lra-constraints.c (emit_spill_move): Remove assert.
9222 (process_alt_operands): Add more debugging
9223 output. Increase reject for spilling into memory. Decrease
9224 reject for reloading scratch.
9225 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9226
a3719e31
SE		 92272013-08-08 Steve Ellcey <sellcey@mips.com>
9228
9229 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9230 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9231 micromips incompatible. Add nan2008.
9232 (MULTILIB_DIRNAMES): Add nan2008.
9233 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9234 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9235 and micromips incompatible. Add nan2008.
9236 (MULTILIB_DIRNAMES): Add nan2008.
9237 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9238
f40423e2
RS		 92392013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9240
9241 PR rtl-optimization/58079
9242 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9243 simplify_comparison has widened a comparison with an integer.
9244
869b9125
KT		 92452013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9246
9247 * config/arm/neon.md (movmisalign<mode>): Disable when we
9248 don't allow unaligned accesses.
9249 (*movmisalign<mode>_neon_store): Likewise.
9250 (*movmisalign<mode>_neon_load): Likewise.
9251 (*movmisalign<mode>_neon_store): Likewise.
9252 (*movmisalign<mode>_neon_load): Likewise.
9253
71cafea9
JH		 92542013-08-08 Jan Hubicka <jh@suse.cz>
9255
9256 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9257 (make_pass_rebuild_cgraph_edges): Also clear references.
9258 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9259 * ipa-inline-transform.c (inline_transform): Remove all references
9260 after inlining.
c97fd4b8
UB		 9261 * cgraphunit.c (expand_function): Remove all references after
9262 expansion.
71cafea9
JH		 9263 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9264 (ipa_find_reference): Rewrite to iterator.
9265 (remove_stmt_references): Likewise.
9266 (ipa_clear_stmts_in_references): New function.
9267 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 9268 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9269 clear references.
71cafea9
JH		 9270 * ipa-split.c (split_function): Remove references in split function.
9271
27d2e612
RE		 92722013-08-08 Richard Earnshaw <rearnsha@arm.com>
9273
9274 PR target/57431
1f65ae7a 9275 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 9276 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9277
5efc661c
RE		 92782013-08-08 Richard Earnshaw <rearnsha@arm.com>
9279
9280 PR target/56979
1f65ae7a 9281 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 9282 suggested mode for the assignment isn't compatible with the
9283 registers required.
9284
f276d31d
BE		 92852013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9286
9287 PR target/58065
9288 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9289
4b8bc035
XDL		 92902013-08-07 Xinliang David Li <davidxl@google.com>
9291
9292 * config/i386/i386.opt: New option -mtune-ctrl=.
9293 * config/i386/x86-tune.def: New file.
9294 * config/i386/i386.h: include x86-tune.def.
9295 * config/i386/i386.c (ix86_option_override_internal):
9296 Parsing -mtune-ctrl= option and set tune features.
9297
12211b99 92982013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 9299
9300 PR other/12081
03b0ee0a 9301 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 9302 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9303 to gen_2arg_fn_t.
9304
5cf6635b
EB		 93052013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9306
9307 * rtl.h (update_alignments): Declare.
9308 * final.c (grow_label_align): New function extracted from...
9309 (shorten_branches): ...here. Call it.
9310 (update_alignments): New function.
9311 * reorg.c (sibling_labels): New variable.
9312 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9313 the new label along with it onto the sibling_labels vector.
9314 (fill_simple_delay_slots): Adjust call to get_label_before.
9315 (fill_slots_from_thread): Likewise.
9316 (relax_delay_slots): Likewise.
9317 (make_return_insns): Likewise.
9318 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9319
3c8ca1ab
EB		 93202013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9321
9322 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9323 document its semantics.
9324 (diagnostic_report_diagnostic): Adjust accordingly.
9325
5ee5b32c
DM		 93262013-08-07 David Malcolm <dmalcolm@redhat.com>
9327
03b0ee0a 9328 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 9329 (sparc_option_override): ...and port to new C++ pass API.
9330 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9331
abd566fa
PB		 93322013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9333
9334 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9335
2077db1b
CT		 93362013-08-06 Caroline Tice <cmtice@google.com>
9337
9338 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9339 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9340 * tree-pass.h: Add pass_vtable_verify.
9341 * varasm.c (assemble_variable): Add code to properly set the comdat
9342 section and name for the .vtable_map_vars section.
9343 (assemble_vtyv_preinit_initializer): New function.
9344 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9345 LINK_ONCE flag.
9346 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9347 * vtable-verify.c: New file.
9348 * vtable-verify.h: New file.
9349 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9350 initialiation levels.
9351 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9352 * passes.def: Insert pass_vtable_verify.
9353 * aclocal.m4: Reorder includes.
9354 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9355 -fvtv-counts options.
9356 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9357 as appropriate, if -fvtable-verify=... is used.
9358 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9359 -fvtable-verify=... is used.
9360 * Makefile.in (OBJS): Add vtable-verify.o to list.
9361 (vtable-verify.o): Add new build rule.
9362 (GTFILES): Add vtable-verify.c to list.
9363 * common.opt (fvtable-verify=): New flag.
9364 (vtv_priority): Values for fvtable-verify= flag.
9365 (fvtv-counts): New flag.
9366 (fvtv-debug): New flag.
9367 * tree.h (save_vtable_map_decl): New extern function decl.
9368
03085d1c
DM		 93692013-08-07 David Malcolm <dmalcolm@redhat.com>
9370
9371 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9372 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9373 (pass_data_rl78_devirt): ...new pass_data instance and...
9374 (make_pass_rl78_devirt): ...new function.
9375 (rl78_asm_file_start): Port pass registration to new C++ API.
9376
05555c4a
DM		 93772013-08-07 David Malcolm <dmalcolm@redhat.com>
9378
9379 * coretypes.h (rtl_opt_pass): Add.
9380 (gcc::context): Add.
9381 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9382 (epiphany_init): Port to new C++ pass API.
9383 (epiphany_optimize_mode_switching): Likewise.
9384 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9385 (pass_manager::get_pass_mode_switching): New.
9386 (pass_manager::get_pass_peephole2): New.
9387 * mode-switching.c (pass_mode_switching): Add clone method.
9388 * recog.c (pass_peephole2): Add clone method.
9389 (pass_split_all_insns): Add clone method.
9390
c5a12e26
DM		 93912013-08-06 David Malcolm <dmalcolm@redhat.com>
9392
03b0ee0a
UB		 9393 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9394 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 9395
a2e2a668
JH		 93962013-08-06 Jan Hubicka <jh@suse.cz>
9397
9398 * cgraph.c (cgraph_get_body): New function based on lto.c
9399 implementation.
9400 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 9401 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9402 LTO paths.
a2e2a668
JH		 9403 * cgraphunit.c (expand_function): Get body prior expanding.
9404 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 9405 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9406 really need.
a2e2a668
JH		 9407 * passes.c (do_per_function_toporder): Get body.
9408 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9409 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9410
15682f24
MJ		 94112013-08-06 Martin Jambor <mjambor@suse.cz>
9412
9413 PR fortran/57987
9414 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9415 is not re-finalized. Rename second parameter to no_collect.
9416
78f6dd68
MJ		 94172013-08-06 Martin Jambor <mjambor@suse.cz>
9418
9419 PR middle-end/58041
9420 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9421 MEM_REF has proper alignment information.
9422
12211b99 94232013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 9424
9425 PR other/12081
9426 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9427 class insn_gen_fn.
9428 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9429 rtx (*) (rtx, ...) with insn_gen_fn.
9430 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9431 insn_gen_fn::stored_funcptr. Add initializer braces.
9432
8ac69a6c
DM		 94332013-08-05 David Malcolm <dmalcolm@redhat.com>
9434
9435 Rewrite how instances of passes are cloned to remove assumptions
9436 about their sizes (thus allowing pass subclasses to have
9437 additional data fields, albeit non-GC-managed ones at this point).
9438
9439 * passes.c (make_pass_instance): Now that passes have clone
9440 methods, rewrite this function to eliminate XNEW and memcpy
9441 calls that used hardcoded sizes. Since this function no longer
9442 creates pass instances, rename it to...
03b0ee0a
UB		 9443 (add_pass_instance): ...this. Document the old way that passes were
9444 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 9445 (next_pass_1): Add an initial_pass argument for use by
9446 add_pass_instance.
9447 (position_pass): When adding multiple instances of a pass, use
9448 the pass's clone method, rather than relying on the XNEW/memcpy
9449 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 9450 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9451 the initial instance of the current pass within the pass manager.
8ac69a6c 9452
27a4cd48
DM		 94532013-08-05 David Malcolm <dmalcolm@redhat.com>
9454
9455 This is the automated part of the conversion of passes from C
9456 structs to C++ classes.
9457
9458 Patch autogenerated by refactor_passes.py from
9459 https://github.com/davidmalcolm/gcc-refactoring-scripts
9460 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9461
9462 * asan.c (pass_asan): Convert from a global struct to a subclass of
9463 gimple_opt_pass along with...
9464 (pass_data_asan): ...new pass_data instance and...
9465 (make_pass_asan): ...new function.
9466 (pass_asan_O0): Convert from a global struct to a subclass of
9467 gimple_opt_pass along with...
9468 (pass_data_asan_O0): ...new pass_data instance and...
9469 (make_pass_asan_O0): ...new function.
9470 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9471 subclass of rtl_opt_pass along with...
9472 (pass_data_inc_dec): ...new pass_data instance and...
9473 (make_pass_inc_dec): ...new function.
9474 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9475 a subclass of rtl_opt_pass along with...
9476 (pass_data_reorder_blocks): ...new pass_data instance and...
9477 (make_pass_reorder_blocks): ...new function.
9478 (pass_duplicate_computed_gotos): Convert from a global struct to a
9479 subclass of rtl_opt_pass along with...
9480 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9481 (make_pass_duplicate_computed_gotos): ...new function.
9482 (pass_partition_blocks): Convert from a global struct to a subclass of
9483 rtl_opt_pass along with...
9484 (pass_data_partition_blocks): ...new pass_data instance and...
9485 (make_pass_partition_blocks): ...new function.
9486 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9487 struct to a subclass of rtl_opt_pass along with...
9488 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9489 and...
9490 (make_pass_branch_target_load_optimize1): ...new function.
9491 (pass_branch_target_load_optimize2): Convert from a global struct to a
9492 subclass of rtl_opt_pass along with...
9493 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9494 and...
9495 (make_pass_branch_target_load_optimize2): ...new function.
9496 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9497 of rtl_opt_pass along with...
9498 (pass_data_jump): ...new pass_data instance and...
9499 (make_pass_jump): ...new function.
9500 (pass_jump2): Convert from a global struct to a subclass of
9501 rtl_opt_pass along with...
9502 (pass_data_jump2): ...new pass_data instance and...
9503 (make_pass_jump2): ...new function.
9504 * cfgexpand.c (pass_expand): Convert from a global struct to a
9505 subclass of rtl_opt_pass along with...
9506 (pass_data_expand): ...new pass_data instance and...
9507 (make_pass_expand): ...new function.
9508 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9509 of rtl_opt_pass along with...
9510 (pass_data_free_cfg): ...new pass_data instance and...
9511 (make_pass_free_cfg): ...new function.
9512 (pass_into_cfg_layout_mode): Convert from a global struct to a
9513 subclass of rtl_opt_pass along with...
9514 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9515 (make_pass_into_cfg_layout_mode): ...new function.
9516 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9517 subclass of rtl_opt_pass along with...
9518 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9519 (make_pass_outof_cfg_layout_mode): ...new function.
9520 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9521 struct to a subclass of gimple_opt_pass along with...
9522 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9523 (make_pass_build_cgraph_edges): ...new function.
9524 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9525 subclass of gimple_opt_pass along with...
9526 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9527 (make_pass_rebuild_cgraph_edges): ...new function.
9528 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9529 subclass of gimple_opt_pass along with...
9530 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9531 and...
9532 (make_pass_remove_cgraph_callee_edges): ...new function.
9533 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9534 struct to a subclass of rtl_opt_pass along with...
9535 (pass_data_stack_adjustments): ...new pass_data instance and...
9536 (make_pass_stack_adjustments): ...new function.
9537 * combine.c (pass_combine): Convert from a global struct to a subclass
9538 of rtl_opt_pass along with...
9539 (pass_data_combine): ...new pass_data instance and...
9540 (make_pass_combine): ...new function.
9541 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9542 global struct to a subclass of rtl_opt_pass along with...
9543 (pass_data_compare_elim_after_reload): ...new pass_data instance
9544 and...
9545 (make_pass_compare_elim_after_reload): ...new function.
9546 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9547 of rtl_opt_pass along with...
9548 (pass_data_rtl_cprop): ...new pass_data instance and...
9549 (make_pass_rtl_cprop): ...new function.
9550 * cse.c (pass_cse): Convert from a global struct to a subclass of
9551 rtl_opt_pass along with...
9552 (pass_data_cse): ...new pass_data instance and...
9553 (make_pass_cse): ...new function.
9554 (pass_cse2): Convert from a global struct to a subclass of
9555 rtl_opt_pass along with...
9556 (pass_data_cse2): ...new pass_data instance and...
9557 (make_pass_cse2): ...new function.
9558 (pass_cse_after_global_opts): Convert from a global struct to a
9559 subclass of rtl_opt_pass along with...
9560 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9561 (make_pass_cse_after_global_opts): ...new function.
9562 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9563 of rtl_opt_pass along with...
9564 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9565 (make_pass_ud_rtl_dce): ...new function.
9566 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9567 rtl_opt_pass along with...
9568 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9569 (make_pass_fast_rtl_dce): ...new function.
9570 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9571 a subclass of rtl_opt_pass along with...
9572 (pass_data_df_initialize_opt): ...new pass_data instance and...
9573 (make_pass_df_initialize_opt): ...new function.
9574 (pass_df_initialize_no_opt): Convert from a global struct to a
9575 subclass of rtl_opt_pass along with...
9576 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9577 (make_pass_df_initialize_no_opt): ...new function.
9578 (pass_df_finish): Convert from a global struct to a subclass of
9579 rtl_opt_pass along with...
9580 (pass_data_df_finish): ...new pass_data instance and...
9581 (make_pass_df_finish): ...new function.
9582 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9583 rtl_opt_pass along with...
9584 (pass_data_rtl_dse1): ...new pass_data instance and...
9585 (make_pass_rtl_dse1): ...new function.
9586 (pass_rtl_dse2): Convert from a global struct to a subclass of
9587 rtl_opt_pass along with...
9588 (pass_data_rtl_dse2): ...new pass_data instance and...
9589 (make_pass_rtl_dse2): ...new function.
9590 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9591 subclass of rtl_opt_pass along with...
9592 (pass_data_dwarf2_frame): ...new pass_data instance and...
9593 (make_pass_dwarf2_frame): ...new function.
9594 * except.c (pass_set_nothrow_function_flags): Convert from a global
9595 struct to a subclass of rtl_opt_pass along with...
9596 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9597 and...
9598 (make_pass_set_nothrow_function_flags): ...new function.
9599 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9600 subclass of rtl_opt_pass along with...
9601 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9602 and...
9603 (make_pass_convert_to_eh_region_ranges): ...new function.
9604 * final.c (pass_compute_alignments): Convert from a global struct to a
9605 subclass of rtl_opt_pass along with...
9606 (pass_data_compute_alignments): ...new pass_data instance and...
9607 (make_pass_compute_alignments): ...new function.
9608 (pass_final): Convert from a global struct to a subclass of
9609 rtl_opt_pass along with...
9610 (pass_data_final): ...new pass_data instance and...
9611 (make_pass_final): ...new function.
9612 (pass_shorten_branches): Convert from a global struct to a subclass of
9613 rtl_opt_pass along with...
9614 (pass_data_shorten_branches): ...new pass_data instance and...
9615 (make_pass_shorten_branches): ...new function.
9616 (pass_clean_state): Convert from a global struct to a subclass of
9617 rtl_opt_pass along with...
9618 (pass_data_clean_state): ...new pass_data instance and...
9619 (make_pass_clean_state): ...new function.
9620 * function.c (pass_instantiate_virtual_regs): Convert from a global
9621 struct to a subclass of rtl_opt_pass along with...
9622 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9623 (make_pass_instantiate_virtual_regs): ...new function.
9624 (pass_leaf_regs): Convert from a global struct to a subclass of
9625 rtl_opt_pass along with...
9626 (pass_data_leaf_regs): ...new pass_data instance and...
9627 (make_pass_leaf_regs): ...new function.
9628 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9629 subclass of rtl_opt_pass along with...
9630 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9631 and...
9632 (make_pass_thread_prologue_and_epilogue): ...new function.
9633 (pass_match_asm_constraints): Convert from a global struct to a
9634 subclass of rtl_opt_pass along with...
9635 (pass_data_match_asm_constraints): ...new pass_data instance and...
9636 (make_pass_match_asm_constraints): ...new function.
9637 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9638 subclass of rtl_opt_pass along with...
9639 (pass_data_rtl_fwprop): ...new pass_data instance and...
9640 (make_pass_rtl_fwprop): ...new function.
9641 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9642 rtl_opt_pass along with...
9643 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9644 (make_pass_rtl_fwprop_addr): ...new function.
9645 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9646 rtl_opt_pass along with...
9647 (pass_data_rtl_pre): ...new pass_data instance and...
9648 (make_pass_rtl_pre): ...new function.
9649 (pass_rtl_hoist): Convert from a global struct to a subclass of
9650 rtl_opt_pass along with...
9651 (pass_data_rtl_hoist): ...new pass_data instance and...
9652 (make_pass_rtl_hoist): ...new function.
9653 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9654 subclass of gimple_opt_pass along with...
9655 (pass_data_lower_cf): ...new pass_data instance and...
9656 (make_pass_lower_cf): ...new function.
9657 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9658 from a global struct to a subclass of gimple_opt_pass along with...
9659 (pass_data_strength_reduction): ...new pass_data instance and...
9660 (make_pass_strength_reduction): ...new function.
9661 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9662 of rtl_opt_pass along with...
9663 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9664 (make_pass_rtl_ifcvt): ...new function.
9665 (pass_if_after_combine): Convert from a global struct to a subclass of
9666 rtl_opt_pass along with...
9667 (pass_data_if_after_combine): ...new pass_data instance and...
9668 (make_pass_if_after_combine): ...new function.
9669 (pass_if_after_reload): Convert from a global struct to a subclass of
9670 rtl_opt_pass along with...
9671 (pass_data_if_after_reload): ...new pass_data instance and...
9672 (make_pass_if_after_reload): ...new function.
9673 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9674 a subclass of rtl_opt_pass along with...
9675 (pass_data_initialize_regs): ...new pass_data instance and...
9676 (make_pass_initialize_regs): ...new function.
9677 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9678 of ipa_opt_pass_d along with...
9679 (pass_data_ipa_cp): ...new pass_data instance and...
9680 (make_pass_ipa_cp): ...new function.
9681 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9682 global struct to a subclass of gimple_opt_pass along with...
9683 (pass_data_inline_parameters): ...new pass_data instance and...
9684 (make_pass_inline_parameters): ...new function.
9685 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9686 subclass of gimple_opt_pass along with...
9687 (pass_data_early_inline): ...new pass_data instance and...
9688 (make_pass_early_inline): ...new function.
9689 (pass_ipa_inline): Convert from a global struct to a subclass of
9690 ipa_opt_pass_d along with...
9691 (pass_data_ipa_inline): ...new pass_data instance and...
9692 (make_pass_ipa_inline): ...new function.
9693 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9694 struct to a subclass of gimple_opt_pass along with...
9695 (pass_data_local_pure_const): ...new pass_data instance and...
9696 (make_pass_local_pure_const): ...new function.
9697 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9698 ipa_opt_pass_d along with...
9699 (pass_data_ipa_pure_const): ...new pass_data instance and...
9700 (make_pass_ipa_pure_const): ...new function.
9701 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9702 to a subclass of ipa_opt_pass_d along with...
9703 (pass_data_ipa_reference): ...new pass_data instance and...
9704 (make_pass_ipa_reference): ...new function.
9705 * ipa-split.c (pass_split_functions): Convert from a global struct to
9706 a subclass of gimple_opt_pass along with...
9707 (pass_data_split_functions): ...new pass_data instance and...
9708 (make_pass_split_functions): ...new function.
9709 (pass_feedback_split_functions): Convert from a global struct to a
9710 subclass of gimple_opt_pass along with...
9711 (pass_data_feedback_split_functions): ...new pass_data instance and...
9712 (make_pass_feedback_split_functions): ...new function.
9713 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9714 global struct to a subclass of simple_ipa_opt_pass along with...
9715 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9716 instance and...
9717 (make_pass_ipa_function_and_variable_visibility): ...new function.
9718 (pass_ipa_free_inline_summary): Convert from a global struct to a
9719 subclass of simple_ipa_opt_pass along with...
9720 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9721 (make_pass_ipa_free_inline_summary): ...new function.
9722 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9723 subclass of ipa_opt_pass_d along with...
9724 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9725 and...
9726 (make_pass_ipa_whole_program_visibility): ...new function.
9727 (pass_ipa_profile): Convert from a global struct to a subclass of
9728 ipa_opt_pass_d along with...
9729 (pass_data_ipa_profile): ...new pass_data instance and...
9730 (make_pass_ipa_profile): ...new function.
9731 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9732 ipa_opt_pass_d along with...
9733 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9734 (make_pass_ipa_cdtor_merge): ...new function.
9735 * ira.c (pass_ira): Convert from a global struct to a subclass of
9736 rtl_opt_pass along with...
9737 (pass_data_ira): ...new pass_data instance and...
9738 (make_pass_ira): ...new function.
9739 (pass_reload): Convert from a global struct to a subclass of
9740 rtl_opt_pass along with...
9741 (pass_data_reload): ...new pass_data instance and...
9742 (make_pass_reload): ...new function.
9743 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9744 subclass of rtl_opt_pass along with...
9745 (pass_data_cleanup_barriers): ...new pass_data instance and...
9746 (make_pass_cleanup_barriers): ...new function.
9747 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9748 of rtl_opt_pass along with...
9749 (pass_data_loop2): ...new pass_data instance and...
9750 (make_pass_loop2): ...new function.
9751 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9752 rtl_opt_pass along with...
9753 (pass_data_rtl_loop_init): ...new pass_data instance and...
9754 (make_pass_rtl_loop_init): ...new function.
9755 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9756 rtl_opt_pass along with...
9757 (pass_data_rtl_loop_done): ...new pass_data instance and...
9758 (make_pass_rtl_loop_done): ...new function.
9759 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9760 subclass of rtl_opt_pass along with...
9761 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9762 (make_pass_rtl_move_loop_invariants): ...new function.
9763 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9764 rtl_opt_pass along with...
9765 (pass_data_rtl_unswitch): ...new pass_data instance and...
9766 (make_pass_rtl_unswitch): ...new function.
9767 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9768 subclass of rtl_opt_pass along with...
9769 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9770 and...
9771 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9772 (pass_rtl_doloop): Convert from a global struct to a subclass of
9773 rtl_opt_pass along with...
9774 (pass_data_rtl_doloop): ...new pass_data instance and...
9775 (make_pass_rtl_doloop): ...new function.
9776 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9777 a subclass of rtl_opt_pass along with...
9778 (pass_data_lower_subreg): ...new pass_data instance and...
9779 (make_pass_lower_subreg): ...new function.
9780 (pass_lower_subreg2): Convert from a global struct to a subclass of
9781 rtl_opt_pass along with...
9782 (pass_data_lower_subreg2): ...new pass_data instance and...
9783 (make_pass_lower_subreg2): ...new function.
9784 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9785 struct to a subclass of ipa_opt_pass_d along with...
9786 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9787 (make_pass_ipa_lto_gimple_out): ...new function.
9788 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9789 of ipa_opt_pass_d along with...
9790 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9791 (make_pass_ipa_lto_finish_out): ...new function.
9792 * mode-switching.c (pass_mode_switching): Convert from a global struct
9793 to a subclass of rtl_opt_pass along with...
9794 (pass_data_mode_switching): ...new pass_data instance and...
9795 (make_pass_mode_switching): ...new function.
9796 * modulo-sched.c (pass_sms): Convert from a global struct to a
9797 subclass of rtl_opt_pass along with...
9798 (pass_data_sms): ...new pass_data instance and...
9799 (make_pass_sms): ...new function.
9800 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9801 subclass of gimple_opt_pass along with...
9802 (pass_data_expand_omp): ...new pass_data instance and...
9803 (make_pass_expand_omp): ...new function.
9804 (pass_lower_omp): Convert from a global struct to a subclass of
9805 gimple_opt_pass along with...
9806 (pass_data_lower_omp): ...new pass_data instance and...
9807 (make_pass_lower_omp): ...new function.
9808 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9809 of gimple_opt_pass along with...
9810 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9811 (make_pass_diagnose_omp_blocks): ...new function.
9812 * passes.c (pass_early_local_passes): Convert from a global struct to
9813 a subclass of simple_ipa_opt_pass along with...
9814 (pass_data_early_local_passes): ...new pass_data instance and...
9815 (make_pass_early_local_passes): ...new function.
9816 (pass_all_early_optimizations): Convert from a global struct to a
9817 subclass of gimple_opt_pass along with...
9818 (pass_data_all_early_optimizations): ...new pass_data instance and...
9819 (make_pass_all_early_optimizations): ...new function.
9820 (pass_all_optimizations): Convert from a global struct to a subclass
9821 of gimple_opt_pass along with...
9822 (pass_data_all_optimizations): ...new pass_data instance and...
9823 (make_pass_all_optimizations): ...new function.
9824 (pass_all_optimizations_g): Convert from a global struct to a subclass
9825 of gimple_opt_pass along with...
9826 (pass_data_all_optimizations_g): ...new pass_data instance and...
9827 (make_pass_all_optimizations_g): ...new function.
9828 (pass_rest_of_compilation): Convert from a global struct to a subclass
9829 of rtl_opt_pass along with...
9830 (pass_data_rest_of_compilation): ...new pass_data instance and...
9831 (make_pass_rest_of_compilation): ...new function.
9832 (pass_postreload): Convert from a global struct to a subclass of
9833 rtl_opt_pass along with...
9834 (pass_data_postreload): ...new pass_data instance and...
9835 (make_pass_postreload): ...new function.
9836 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9837 subclass of rtl_opt_pass along with...
9838 (pass_data_gcse2): ...new pass_data instance and...
9839 (make_pass_gcse2): ...new function.
9840 * postreload.c (pass_postreload_cse): Convert from a global struct to
9841 a subclass of rtl_opt_pass along with...
9842 (pass_data_postreload_cse): ...new pass_data instance and...
9843 (make_pass_postreload_cse): ...new function.
9844 * predict.c (pass_profile): Convert from a global struct to a subclass
9845 of gimple_opt_pass along with...
9846 (pass_data_profile): ...new pass_data instance and...
9847 (make_pass_profile): ...new function.
9848 (pass_strip_predict_hints): Convert from a global struct to a subclass
9849 of gimple_opt_pass along with...
9850 (pass_data_strip_predict_hints): ...new pass_data instance and...
9851 (make_pass_strip_predict_hints): ...new function.
9852 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9853 of rtl_opt_pass along with...
9854 (pass_data_peephole2): ...new pass_data instance and...
9855 (make_pass_peephole2): ...new function.
9856 (pass_split_all_insns): Convert from a global struct to a subclass of
9857 rtl_opt_pass along with...
9858 (pass_data_split_all_insns): ...new pass_data instance and...
9859 (make_pass_split_all_insns): ...new function.
9860 (pass_split_after_reload): Convert from a global struct to a subclass
9861 of rtl_opt_pass along with...
9862 (pass_data_split_after_reload): ...new pass_data instance and...
9863 (make_pass_split_after_reload): ...new function.
9864 (pass_split_before_regstack): Convert from a global struct to a
9865 subclass of rtl_opt_pass along with...
9866 (pass_data_split_before_regstack): ...new pass_data instance and...
9867 (make_pass_split_before_regstack): ...new function.
9868 (pass_split_before_sched2): Convert from a global struct to a subclass
9869 of rtl_opt_pass along with...
9870 (pass_data_split_before_sched2): ...new pass_data instance and...
9871 (make_pass_split_before_sched2): ...new function.
9872 (pass_split_for_shorten_branches): Convert from a global struct to a
9873 subclass of rtl_opt_pass along with...
9874 (pass_data_split_for_shorten_branches): ...new pass_data instance
9875 and...
9876 (make_pass_split_for_shorten_branches): ...new function.
9877 * ree.c (pass_ree): Convert from a global struct to a subclass of
9878 rtl_opt_pass along with...
9879 (pass_data_ree): ...new pass_data instance and...
9880 (make_pass_ree): ...new function.
9881 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9882 subclass of rtl_opt_pass along with...
9883 (pass_data_stack_regs): ...new pass_data instance and...
9884 (make_pass_stack_regs): ...new function.
9885 (pass_stack_regs_run): Convert from a global struct to a subclass of
9886 rtl_opt_pass along with...
9887 (pass_data_stack_regs_run): ...new pass_data instance and...
9888 (make_pass_stack_regs_run): ...new function.
9889 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9890 subclass of rtl_opt_pass along with...
9891 (pass_data_cprop_hardreg): ...new pass_data instance and...
9892 (make_pass_cprop_hardreg): ...new function.
9893 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9894 subclass of rtl_opt_pass along with...
9895 (pass_data_reginfo_init): ...new pass_data instance and...
9896 (make_pass_reginfo_init): ...new function.
9897 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9898 of rtl_opt_pass along with...
9899 (pass_data_regmove): ...new pass_data instance and...
9900 (make_pass_regmove): ...new function.
9901 * regrename.c (pass_regrename): Convert from a global struct to a
9902 subclass of rtl_opt_pass along with...
9903 (pass_data_regrename): ...new pass_data instance and...
9904 (make_pass_regrename): ...new function.
9905 * reorg.c (pass_delay_slots): Convert from a global struct to a
9906 subclass of rtl_opt_pass along with...
9907 (pass_data_delay_slots): ...new pass_data instance and...
9908 (make_pass_delay_slots): ...new function.
9909 (pass_machine_reorg): Convert from a global struct to a subclass of
9910 rtl_opt_pass along with...
9911 (pass_data_machine_reorg): ...new pass_data instance and...
9912 (make_pass_machine_reorg): ...new function.
9913 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9914 of rtl_opt_pass along with...
9915 (pass_data_sched): ...new pass_data instance and...
9916 (make_pass_sched): ...new function.
9917 (pass_sched2): Convert from a global struct to a subclass of
9918 rtl_opt_pass along with...
9919 (pass_data_sched2): ...new pass_data instance and...
9920 (make_pass_sched2): ...new function.
9921 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9922 to a subclass of rtl_opt_pass along with...
9923 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9924 (make_pass_stack_ptr_mod): ...new function.
9925 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9926 to a subclass of rtl_opt_pass along with...
9927 (pass_data_rtl_store_motion): ...new pass_data instance and...
9928 (make_pass_rtl_store_motion): ...new function.
9929 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9930 of gimple_opt_pass along with...
9931 (pass_data_tracer): ...new pass_data instance and...
9932 (make_pass_tracer): ...new function.
9933 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9934 to a subclass of gimple_opt_pass along with...
9935 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9936 (make_pass_diagnose_tm_blocks): ...new function.
9937 (pass_lower_tm): Convert from a global struct to a subclass of
9938 gimple_opt_pass along with...
9939 (pass_data_lower_tm): ...new pass_data instance and...
9940 (make_pass_lower_tm): ...new function.
9941 (pass_tm_init): Convert from a global struct to a subclass of
9942 gimple_opt_pass along with...
9943 (pass_data_tm_init): ...new pass_data instance and...
9944 (make_pass_tm_init): ...new function.
9945 (pass_tm_mark): Convert from a global struct to a subclass of
9946 gimple_opt_pass along with...
9947 (pass_data_tm_mark): ...new pass_data instance and...
9948 (make_pass_tm_mark): ...new function.
9949 (pass_tm_edges): Convert from a global struct to a subclass of
9950 gimple_opt_pass along with...
9951 (pass_data_tm_edges): ...new pass_data instance and...
9952 (make_pass_tm_edges): ...new function.
9953 (pass_tm_memopt): Convert from a global struct to a subclass of
9954 gimple_opt_pass along with...
9955 (pass_data_tm_memopt): ...new pass_data instance and...
9956 (make_pass_tm_memopt): ...new function.
9957 (pass_ipa_tm): Convert from a global struct to a subclass of
9958 simple_ipa_opt_pass along with...
9959 (pass_data_ipa_tm): ...new pass_data instance and...
9960 (make_pass_ipa_tm): ...new function.
9961 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9962 subclass of gimple_opt_pass along with...
9963 (pass_data_call_cdce): ...new pass_data instance and...
9964 (make_pass_call_cdce): ...new function.
9965 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9966 subclass of gimple_opt_pass along with...
9967 (pass_data_build_cfg): ...new pass_data instance and...
9968 (make_pass_build_cfg): ...new function.
9969 (pass_split_crit_edges): Convert from a global struct to a subclass of
9970 gimple_opt_pass along with...
9971 (pass_data_split_crit_edges): ...new pass_data instance and...
9972 (make_pass_split_crit_edges): ...new function.
9973 (pass_warn_function_return): Convert from a global struct to a
9974 subclass of gimple_opt_pass along with...
9975 (pass_data_warn_function_return): ...new pass_data instance and...
9976 (make_pass_warn_function_return): ...new function.
9977 (pass_warn_function_noreturn): Convert from a global struct to a
9978 subclass of gimple_opt_pass along with...
9979 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9980 (make_pass_warn_function_noreturn): ...new function.
9981 (pass_warn_unused_result): Convert from a global struct to a subclass
9982 of gimple_opt_pass along with...
9983 (pass_data_warn_unused_result): ...new pass_data instance and...
9984 (make_pass_warn_unused_result): ...new function.
9985 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9986 a subclass of gimple_opt_pass along with...
9987 (pass_data_merge_phi): ...new pass_data instance and...
9988 (make_pass_merge_phi): ...new function.
9989 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9990 a subclass of gimple_opt_pass along with...
9991 (pass_data_lower_complex): ...new pass_data instance and...
9992 (make_pass_lower_complex): ...new function.
9993 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9994 gimple_opt_pass along with...
9995 (pass_data_lower_complex_O0): ...new pass_data instance and...
9996 (make_pass_lower_complex_O0): ...new function.
9997 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9998 subclass of gimple_opt_pass along with...
9999 (pass_data_lower_eh): ...new pass_data instance and...
10000 (make_pass_lower_eh): ...new function.
10001 (pass_refactor_eh): Convert from a global struct to a subclass of
10002 gimple_opt_pass along with...
10003 (pass_data_refactor_eh): ...new pass_data instance and...
10004 (make_pass_refactor_eh): ...new function.
10005 (pass_lower_resx): Convert from a global struct to a subclass of
10006 gimple_opt_pass along with...
10007 (pass_data_lower_resx): ...new pass_data instance and...
10008 (make_pass_lower_resx): ...new function.
10009 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10010 of gimple_opt_pass along with...
10011 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10012 (make_pass_lower_eh_dispatch): ...new function.
10013 (pass_cleanup_eh): Convert from a global struct to a subclass of
10014 gimple_opt_pass along with...
10015 (pass_data_cleanup_eh): ...new pass_data instance and...
10016 (make_pass_cleanup_eh): ...new function.
10017 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10018 to a subclass of simple_ipa_opt_pass along with...
10019 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10020 (make_pass_ipa_lower_emutls): ...new function.
10021 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10022 a subclass of gimple_opt_pass along with...
10023 (pass_data_if_conversion): ...new pass_data instance and...
10024 (make_pass_if_conversion): ...new function.
10025 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10026 subclass of gimple_opt_pass along with...
10027 (pass_data_build_ssa): ...new pass_data instance and...
10028 (make_pass_build_ssa): ...new function.
10029 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10030 global struct to a subclass of gimple_opt_pass along with...
10031 (pass_data_loop_distribution): ...new pass_data instance and...
10032 (make_pass_loop_distribution): ...new function.
10033 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10034 subclass of gimple_opt_pass along with...
10035 (pass_data_mudflap_1): ...new pass_data instance and...
10036 (make_pass_mudflap_1): ...new function.
10037 (pass_mudflap_2): Convert from a global struct to a subclass of
10038 gimple_opt_pass along with...
10039 (pass_data_mudflap_2): ...new pass_data instance and...
10040 (make_pass_mudflap_2): ...new function.
10041 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10042 subclass of gimple_opt_pass along with...
10043 (pass_data_mudflap_1): ...new pass_data instance and...
10044 (make_pass_mudflap_1): ...new function.
10045 (pass_mudflap_2): Convert from a global struct to a subclass of
10046 gimple_opt_pass along with...
10047 (pass_data_mudflap_2): ...new pass_data instance and...
10048 (make_pass_mudflap_2): ...new function.
10049 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10050 gimple_opt_pass along with...
10051 (pass_data_nrv): ...new pass_data instance and...
10052 (make_pass_nrv): ...new function.
10053 (pass_return_slot): Convert from a global struct to a subclass of
10054 gimple_opt_pass along with...
10055 (pass_data_return_slot): ...new pass_data instance and...
10056 (make_pass_return_slot): ...new function.
10057 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10058 to a subclass of gimple_opt_pass along with...
10059 (pass_data_object_sizes): ...new pass_data instance and...
10060 (make_pass_object_sizes): ...new function.
10061 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10062 global struct to a subclass of gimple_opt_pass along with...
10063 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10064 and...
10065 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10066 (pass_fixup_cfg): Convert from a global struct to a subclass of
10067 gimple_opt_pass along with...
10068 (pass_data_fixup_cfg): ...new pass_data instance and...
10069 (make_pass_fixup_cfg): ...new function.
10070 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10071 (make_pass_mudflap_1): ...new function.
10072 (pass_mudflap_2): Replace declaration with that of...
10073 (make_pass_mudflap_2): ...new function.
10074 (pass_asan): Replace declaration with that of...
10075 (make_pass_asan): ...new function.
10076 (pass_asan_O0): Replace declaration with that of...
10077 (make_pass_asan_O0): ...new function.
10078 (pass_tsan): Replace declaration with that of...
10079 (make_pass_tsan): ...new function.
10080 (pass_tsan_O0): Replace declaration with that of...
10081 (make_pass_tsan_O0): ...new function.
10082 (pass_lower_cf): Replace declaration with that of...
10083 (make_pass_lower_cf): ...new function.
10084 (pass_refactor_eh): Replace declaration with that of...
10085 (make_pass_refactor_eh): ...new function.
10086 (pass_lower_eh): Replace declaration with that of...
10087 (make_pass_lower_eh): ...new function.
10088 (pass_lower_eh_dispatch): Replace declaration with that of...
10089 (make_pass_lower_eh_dispatch): ...new function.
10090 (pass_lower_resx): Replace declaration with that of...
10091 (make_pass_lower_resx): ...new function.
10092 (pass_build_cfg): Replace declaration with that of...
10093 (make_pass_build_cfg): ...new function.
10094 (pass_early_tree_profile): Replace declaration with that of...
10095 (make_pass_early_tree_profile): ...new function.
10096 (pass_cleanup_eh): Replace declaration with that of...
10097 (make_pass_cleanup_eh): ...new function.
10098 (pass_sra): Replace declaration with that of...
10099 (make_pass_sra): ...new function.
10100 (pass_sra_early): Replace declaration with that of...
10101 (make_pass_sra_early): ...new function.
10102 (pass_early_ipa_sra): Replace declaration with that of...
10103 (make_pass_early_ipa_sra): ...new function.
10104 (pass_tail_recursion): Replace declaration with that of...
10105 (make_pass_tail_recursion): ...new function.
10106 (pass_tail_calls): Replace declaration with that of...
10107 (make_pass_tail_calls): ...new function.
10108 (pass_tree_loop): Replace declaration with that of...
10109 (make_pass_tree_loop): ...new function.
10110 (pass_tree_loop_init): Replace declaration with that of...
10111 (make_pass_tree_loop_init): ...new function.
10112 (pass_lim): Replace declaration with that of...
10113 (make_pass_lim): ...new function.
10114 (pass_tree_unswitch): Replace declaration with that of...
10115 (make_pass_tree_unswitch): ...new function.
10116 (pass_predcom): Replace declaration with that of...
10117 (make_pass_predcom): ...new function.
10118 (pass_iv_canon): Replace declaration with that of...
10119 (make_pass_iv_canon): ...new function.
10120 (pass_scev_cprop): Replace declaration with that of...
10121 (make_pass_scev_cprop): ...new function.
10122 (pass_empty_loop): Replace declaration with that of...
10123 (make_pass_empty_loop): ...new function.
10124 (pass_record_bounds): Replace declaration with that of...
10125 (make_pass_record_bounds): ...new function.
10126 (pass_graphite): Replace declaration with that of...
10127 (make_pass_graphite): ...new function.
10128 (pass_graphite_transforms): Replace declaration with that of...
10129 (make_pass_graphite_transforms): ...new function.
10130 (pass_if_conversion): Replace declaration with that of...
10131 (make_pass_if_conversion): ...new function.
10132 (pass_loop_distribution): Replace declaration with that of...
10133 (make_pass_loop_distribution): ...new function.
10134 (pass_vectorize): Replace declaration with that of...
10135 (make_pass_vectorize): ...new function.
10136 (pass_slp_vectorize): Replace declaration with that of...
10137 (make_pass_slp_vectorize): ...new function.
10138 (pass_complete_unroll): Replace declaration with that of...
10139 (make_pass_complete_unroll): ...new function.
10140 (pass_complete_unrolli): Replace declaration with that of...
10141 (make_pass_complete_unrolli): ...new function.
10142 (pass_parallelize_loops): Replace declaration with that of...
10143 (make_pass_parallelize_loops): ...new function.
10144 (pass_loop_prefetch): Replace declaration with that of...
10145 (make_pass_loop_prefetch): ...new function.
10146 (pass_iv_optimize): Replace declaration with that of...
10147 (make_pass_iv_optimize): ...new function.
10148 (pass_tree_loop_done): Replace declaration with that of...
10149 (make_pass_tree_loop_done): ...new function.
10150 (pass_ch): Replace declaration with that of...
10151 (make_pass_ch): ...new function.
10152 (pass_ccp): Replace declaration with that of...
10153 (make_pass_ccp): ...new function.
10154 (pass_phi_only_cprop): Replace declaration with that of...
10155 (make_pass_phi_only_cprop): ...new function.
10156 (pass_build_ssa): Replace declaration with that of...
10157 (make_pass_build_ssa): ...new function.
10158 (pass_build_alias): Replace declaration with that of...
10159 (make_pass_build_alias): ...new function.
10160 (pass_build_ealias): Replace declaration with that of...
10161 (make_pass_build_ealias): ...new function.
10162 (pass_dominator): Replace declaration with that of...
10163 (make_pass_dominator): ...new function.
10164 (pass_dce): Replace declaration with that of...
10165 (make_pass_dce): ...new function.
10166 (pass_dce_loop): Replace declaration with that of...
10167 (make_pass_dce_loop): ...new function.
10168 (pass_cd_dce): Replace declaration with that of...
10169 (make_pass_cd_dce): ...new function.
10170 (pass_call_cdce): Replace declaration with that of...
10171 (make_pass_call_cdce): ...new function.
10172 (pass_merge_phi): Replace declaration with that of...
10173 (make_pass_merge_phi): ...new function.
10174 (pass_split_crit_edges): Replace declaration with that of...
10175 (make_pass_split_crit_edges): ...new function.
10176 (pass_pre): Replace declaration with that of...
10177 (make_pass_pre): ...new function.
10178 (pass_profile): Replace declaration with that of...
10179 (make_pass_profile): ...new function.
10180 (pass_strip_predict_hints): Replace declaration with that of...
10181 (make_pass_strip_predict_hints): ...new function.
10182 (pass_lower_complex_O0): Replace declaration with that of...
10183 (make_pass_lower_complex_O0): ...new function.
10184 (pass_lower_complex): Replace declaration with that of...
10185 (make_pass_lower_complex): ...new function.
10186 (pass_lower_vector): Replace declaration with that of...
10187 (make_pass_lower_vector): ...new function.
10188 (pass_lower_vector_ssa): Replace declaration with that of...
10189 (make_pass_lower_vector_ssa): ...new function.
10190 (pass_lower_omp): Replace declaration with that of...
10191 (make_pass_lower_omp): ...new function.
10192 (pass_diagnose_omp_blocks): Replace declaration with that of...
10193 (make_pass_diagnose_omp_blocks): ...new function.
10194 (pass_expand_omp): Replace declaration with that of...
10195 (make_pass_expand_omp): ...new function.
10196 (pass_expand_omp_ssa): Replace declaration with that of...
10197 (make_pass_expand_omp_ssa): ...new function.
10198 (pass_object_sizes): Replace declaration with that of...
10199 (make_pass_object_sizes): ...new function.
10200 (pass_strlen): Replace declaration with that of...
10201 (make_pass_strlen): ...new function.
10202 (pass_fold_builtins): Replace declaration with that of...
10203 (make_pass_fold_builtins): ...new function.
10204 (pass_stdarg): Replace declaration with that of...
10205 (make_pass_stdarg): ...new function.
10206 (pass_early_warn_uninitialized): Replace declaration with that of...
10207 (make_pass_early_warn_uninitialized): ...new function.
10208 (pass_late_warn_uninitialized): Replace declaration with that of...
10209 (make_pass_late_warn_uninitialized): ...new function.
10210 (pass_cse_reciprocals): Replace declaration with that of...
10211 (make_pass_cse_reciprocals): ...new function.
10212 (pass_cse_sincos): Replace declaration with that of...
10213 (make_pass_cse_sincos): ...new function.
10214 (pass_optimize_bswap): Replace declaration with that of...
10215 (make_pass_optimize_bswap): ...new function.
10216 (pass_optimize_widening_mul): Replace declaration with that of...
10217 (make_pass_optimize_widening_mul): ...new function.
10218 (pass_warn_function_return): Replace declaration with that of...
10219 (make_pass_warn_function_return): ...new function.
10220 (pass_warn_function_noreturn): Replace declaration with that of...
10221 (make_pass_warn_function_noreturn): ...new function.
10222 (pass_cselim): Replace declaration with that of...
10223 (make_pass_cselim): ...new function.
10224 (pass_phiopt): Replace declaration with that of...
10225 (make_pass_phiopt): ...new function.
10226 (pass_forwprop): Replace declaration with that of...
10227 (make_pass_forwprop): ...new function.
10228 (pass_phiprop): Replace declaration with that of...
10229 (make_pass_phiprop): ...new function.
10230 (pass_tree_ifcombine): Replace declaration with that of...
10231 (make_pass_tree_ifcombine): ...new function.
10232 (pass_dse): Replace declaration with that of...
10233 (make_pass_dse): ...new function.
10234 (pass_nrv): Replace declaration with that of...
10235 (make_pass_nrv): ...new function.
10236 (pass_rename_ssa_copies): Replace declaration with that of...
10237 (make_pass_rename_ssa_copies): ...new function.
10238 (pass_sink_code): Replace declaration with that of...
10239 (make_pass_sink_code): ...new function.
10240 (pass_fre): Replace declaration with that of...
10241 (make_pass_fre): ...new function.
10242 (pass_check_data_deps): Replace declaration with that of...
10243 (make_pass_check_data_deps): ...new function.
10244 (pass_copy_prop): Replace declaration with that of...
10245 (make_pass_copy_prop): ...new function.
10246 (pass_vrp): Replace declaration with that of...
10247 (make_pass_vrp): ...new function.
10248 (pass_uncprop): Replace declaration with that of...
10249 (make_pass_uncprop): ...new function.
10250 (pass_return_slot): Replace declaration with that of...
10251 (make_pass_return_slot): ...new function.
10252 (pass_reassoc): Replace declaration with that of...
10253 (make_pass_reassoc): ...new function.
10254 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10255 (make_pass_rebuild_cgraph_edges): ...new function.
10256 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10257 (make_pass_remove_cgraph_callee_edges): ...new function.
10258 (pass_build_cgraph_edges): Replace declaration with that of...
10259 (make_pass_build_cgraph_edges): ...new function.
10260 (pass_local_pure_const): Replace declaration with that of...
10261 (make_pass_local_pure_const): ...new function.
10262 (pass_tracer): Replace declaration with that of...
10263 (make_pass_tracer): ...new function.
10264 (pass_warn_unused_result): Replace declaration with that of...
10265 (make_pass_warn_unused_result): ...new function.
10266 (pass_diagnose_tm_blocks): Replace declaration with that of...
10267 (make_pass_diagnose_tm_blocks): ...new function.
10268 (pass_lower_tm): Replace declaration with that of...
10269 (make_pass_lower_tm): ...new function.
10270 (pass_tm_init): Replace declaration with that of...
10271 (make_pass_tm_init): ...new function.
10272 (pass_tm_mark): Replace declaration with that of...
10273 (make_pass_tm_mark): ...new function.
10274 (pass_tm_memopt): Replace declaration with that of...
10275 (make_pass_tm_memopt): ...new function.
10276 (pass_tm_edges): Replace declaration with that of...
10277 (make_pass_tm_edges): ...new function.
10278 (pass_split_functions): Replace declaration with that of...
10279 (make_pass_split_functions): ...new function.
10280 (pass_feedback_split_functions): Replace declaration with that of...
10281 (make_pass_feedback_split_functions): ...new function.
10282 (pass_strength_reduction): Replace declaration with that of...
10283 (make_pass_strength_reduction): ...new function.
10284 (pass_ipa_lower_emutls): Replace declaration with that of...
10285 (make_pass_ipa_lower_emutls): ...new function.
10286 (pass_ipa_function_and_variable_visibility): Replace declaration with
10287 that of...
10288 (make_pass_ipa_function_and_variable_visibility): ...new function.
10289 (pass_ipa_tree_profile): Replace declaration with that of...
10290 (make_pass_ipa_tree_profile): ...new function.
10291 (pass_early_local_passes): Replace declaration with that of...
10292 (make_pass_early_local_passes): ...new function.
10293 (pass_ipa_whole_program_visibility): Replace declaration with that
10294 of...
10295 (make_pass_ipa_whole_program_visibility): ...new function.
10296 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10297 (make_pass_ipa_lto_gimple_out): ...new function.
10298 (pass_ipa_increase_alignment): Replace declaration with that of...
10299 (make_pass_ipa_increase_alignment): ...new function.
10300 (pass_ipa_inline): Replace declaration with that of...
10301 (make_pass_ipa_inline): ...new function.
10302 (pass_ipa_free_lang_data): Replace declaration with that of...
10303 (make_pass_ipa_free_lang_data): ...new function.
10304 (pass_ipa_free_inline_summary): Replace declaration with that of...
10305 (make_pass_ipa_free_inline_summary): ...new function.
10306 (pass_ipa_cp): Replace declaration with that of...
10307 (make_pass_ipa_cp): ...new function.
10308 (pass_ipa_reference): Replace declaration with that of...
10309 (make_pass_ipa_reference): ...new function.
10310 (pass_ipa_pure_const): Replace declaration with that of...
10311 (make_pass_ipa_pure_const): ...new function.
10312 (pass_ipa_pta): Replace declaration with that of...
10313 (make_pass_ipa_pta): ...new function.
10314 (pass_ipa_lto_finish_out): Replace declaration with that of...
10315 (make_pass_ipa_lto_finish_out): ...new function.
10316 (pass_ipa_tm): Replace declaration with that of...
10317 (make_pass_ipa_tm): ...new function.
10318 (pass_ipa_profile): Replace declaration with that of...
10319 (make_pass_ipa_profile): ...new function.
10320 (pass_ipa_cdtor_merge): Replace declaration with that of...
10321 (make_pass_ipa_cdtor_merge): ...new function.
10322 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10323 of...
10324 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10325 (pass_init_datastructures): Replace declaration with that of...
10326 (make_pass_init_datastructures): ...new function.
10327 (pass_fixup_cfg): Replace declaration with that of...
10328 (make_pass_fixup_cfg): ...new function.
10329 (pass_expand): Replace declaration with that of...
10330 (make_pass_expand): ...new function.
10331 (pass_instantiate_virtual_regs): Replace declaration with that of...
10332 (make_pass_instantiate_virtual_regs): ...new function.
10333 (pass_rtl_fwprop): Replace declaration with that of...
10334 (make_pass_rtl_fwprop): ...new function.
10335 (pass_rtl_fwprop_addr): Replace declaration with that of...
10336 (make_pass_rtl_fwprop_addr): ...new function.
10337 (pass_jump): Replace declaration with that of...
10338 (make_pass_jump): ...new function.
10339 (pass_jump2): Replace declaration with that of...
10340 (make_pass_jump2): ...new function.
10341 (pass_lower_subreg): Replace declaration with that of...
10342 (make_pass_lower_subreg): ...new function.
10343 (pass_cse): Replace declaration with that of...
10344 (make_pass_cse): ...new function.
10345 (pass_fast_rtl_dce): Replace declaration with that of...
10346 (make_pass_fast_rtl_dce): ...new function.
10347 (pass_ud_rtl_dce): Replace declaration with that of...
10348 (make_pass_ud_rtl_dce): ...new function.
10349 (pass_rtl_dce): Replace declaration with that of...
10350 (make_pass_rtl_dce): ...new function.
10351 (pass_rtl_dse1): Replace declaration with that of...
10352 (make_pass_rtl_dse1): ...new function.
10353 (pass_rtl_dse2): Replace declaration with that of...
10354 (make_pass_rtl_dse2): ...new function.
10355 (pass_rtl_dse3): Replace declaration with that of...
10356 (make_pass_rtl_dse3): ...new function.
10357 (pass_rtl_cprop): Replace declaration with that of...
10358 (make_pass_rtl_cprop): ...new function.
10359 (pass_rtl_pre): Replace declaration with that of...
10360 (make_pass_rtl_pre): ...new function.
10361 (pass_rtl_hoist): Replace declaration with that of...
10362 (make_pass_rtl_hoist): ...new function.
10363 (pass_rtl_store_motion): Replace declaration with that of...
10364 (make_pass_rtl_store_motion): ...new function.
10365 (pass_cse_after_global_opts): Replace declaration with that of...
10366 (make_pass_cse_after_global_opts): ...new function.
10367 (pass_rtl_ifcvt): Replace declaration with that of...
10368 (make_pass_rtl_ifcvt): ...new function.
10369 (pass_into_cfg_layout_mode): Replace declaration with that of...
10370 (make_pass_into_cfg_layout_mode): ...new function.
10371 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10372 (make_pass_outof_cfg_layout_mode): ...new function.
10373 (pass_loop2): Replace declaration with that of...
10374 (make_pass_loop2): ...new function.
10375 (pass_rtl_loop_init): Replace declaration with that of...
10376 (make_pass_rtl_loop_init): ...new function.
10377 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10378 (make_pass_rtl_move_loop_invariants): ...new function.
10379 (pass_rtl_unswitch): Replace declaration with that of...
10380 (make_pass_rtl_unswitch): ...new function.
10381 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10382 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10383 (pass_rtl_doloop): Replace declaration with that of...
10384 (make_pass_rtl_doloop): ...new function.
10385 (pass_rtl_loop_done): Replace declaration with that of...
10386 (make_pass_rtl_loop_done): ...new function.
10387 (pass_web): Replace declaration with that of...
10388 (make_pass_web): ...new function.
10389 (pass_cse2): Replace declaration with that of...
10390 (make_pass_cse2): ...new function.
10391 (pass_df_initialize_opt): Replace declaration with that of...
10392 (make_pass_df_initialize_opt): ...new function.
10393 (pass_df_initialize_no_opt): Replace declaration with that of...
10394 (make_pass_df_initialize_no_opt): ...new function.
10395 (pass_reginfo_init): Replace declaration with that of...
10396 (make_pass_reginfo_init): ...new function.
10397 (pass_inc_dec): Replace declaration with that of...
10398 (make_pass_inc_dec): ...new function.
10399 (pass_stack_ptr_mod): Replace declaration with that of...
10400 (make_pass_stack_ptr_mod): ...new function.
10401 (pass_initialize_regs): Replace declaration with that of...
10402 (make_pass_initialize_regs): ...new function.
10403 (pass_combine): Replace declaration with that of...
10404 (make_pass_combine): ...new function.
10405 (pass_if_after_combine): Replace declaration with that of...
10406 (make_pass_if_after_combine): ...new function.
10407 (pass_ree): Replace declaration with that of...
10408 (make_pass_ree): ...new function.
10409 (pass_partition_blocks): Replace declaration with that of...
10410 (make_pass_partition_blocks): ...new function.
10411 (pass_match_asm_constraints): Replace declaration with that of...
10412 (make_pass_match_asm_constraints): ...new function.
10413 (pass_regmove): Replace declaration with that of...
10414 (make_pass_regmove): ...new function.
10415 (pass_split_all_insns): Replace declaration with that of...
10416 (make_pass_split_all_insns): ...new function.
10417 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10418 (make_pass_fast_rtl_byte_dce): ...new function.
10419 (pass_lower_subreg2): Replace declaration with that of...
10420 (make_pass_lower_subreg2): ...new function.
10421 (pass_mode_switching): Replace declaration with that of...
10422 (make_pass_mode_switching): ...new function.
10423 (pass_sms): Replace declaration with that of...
10424 (make_pass_sms): ...new function.
10425 (pass_sched): Replace declaration with that of...
10426 (make_pass_sched): ...new function.
10427 (pass_ira): Replace declaration with that of...
10428 (make_pass_ira): ...new function.
10429 (pass_reload): Replace declaration with that of...
10430 (make_pass_reload): ...new function.
10431 (pass_clean_state): Replace declaration with that of...
10432 (make_pass_clean_state): ...new function.
10433 (pass_branch_prob): Replace declaration with that of...
10434 (make_pass_branch_prob): ...new function.
10435 (pass_value_profile_transformations): Replace declaration with that
10436 of...
10437 (make_pass_value_profile_transformations): ...new function.
10438 (pass_postreload_cse): Replace declaration with that of...
10439 (make_pass_postreload_cse): ...new function.
10440 (pass_gcse2): Replace declaration with that of...
10441 (make_pass_gcse2): ...new function.
10442 (pass_split_after_reload): Replace declaration with that of...
10443 (make_pass_split_after_reload): ...new function.
10444 (pass_branch_target_load_optimize1): Replace declaration with that
10445 of...
10446 (make_pass_branch_target_load_optimize1): ...new function.
10447 (pass_thread_prologue_and_epilogue): Replace declaration with that
10448 of...
10449 (make_pass_thread_prologue_and_epilogue): ...new function.
10450 (pass_stack_adjustments): Replace declaration with that of...
10451 (make_pass_stack_adjustments): ...new function.
10452 (pass_peephole2): Replace declaration with that of...
10453 (make_pass_peephole2): ...new function.
10454 (pass_if_after_reload): Replace declaration with that of...
10455 (make_pass_if_after_reload): ...new function.
10456 (pass_regrename): Replace declaration with that of...
10457 (make_pass_regrename): ...new function.
10458 (pass_cprop_hardreg): Replace declaration with that of...
10459 (make_pass_cprop_hardreg): ...new function.
10460 (pass_reorder_blocks): Replace declaration with that of...
10461 (make_pass_reorder_blocks): ...new function.
10462 (pass_branch_target_load_optimize2): Replace declaration with that
10463 of...
10464 (make_pass_branch_target_load_optimize2): ...new function.
10465 (pass_leaf_regs): Replace declaration with that of...
10466 (make_pass_leaf_regs): ...new function.
10467 (pass_split_before_sched2): Replace declaration with that of...
10468 (make_pass_split_before_sched2): ...new function.
10469 (pass_compare_elim_after_reload): Replace declaration with that of...
10470 (make_pass_compare_elim_after_reload): ...new function.
10471 (pass_sched2): Replace declaration with that of...
10472 (make_pass_sched2): ...new function.
10473 (pass_stack_regs): Replace declaration with that of...
10474 (make_pass_stack_regs): ...new function.
10475 (pass_stack_regs_run): Replace declaration with that of...
10476 (make_pass_stack_regs_run): ...new function.
10477 (pass_df_finish): Replace declaration with that of...
10478 (make_pass_df_finish): ...new function.
10479 (pass_compute_alignments): Replace declaration with that of...
10480 (make_pass_compute_alignments): ...new function.
10481 (pass_duplicate_computed_gotos): Replace declaration with that of...
10482 (make_pass_duplicate_computed_gotos): ...new function.
10483 (pass_variable_tracking): Replace declaration with that of...
10484 (make_pass_variable_tracking): ...new function.
10485 (pass_free_cfg): Replace declaration with that of...
10486 (make_pass_free_cfg): ...new function.
10487 (pass_machine_reorg): Replace declaration with that of...
10488 (make_pass_machine_reorg): ...new function.
10489 (pass_cleanup_barriers): Replace declaration with that of...
10490 (make_pass_cleanup_barriers): ...new function.
10491 (pass_delay_slots): Replace declaration with that of...
10492 (make_pass_delay_slots): ...new function.
10493 (pass_split_for_shorten_branches): Replace declaration with that of...
10494 (make_pass_split_for_shorten_branches): ...new function.
10495 (pass_split_before_regstack): Replace declaration with that of...
10496 (make_pass_split_before_regstack): ...new function.
10497 (pass_convert_to_eh_region_ranges): Replace declaration with that
10498 of...
10499 (make_pass_convert_to_eh_region_ranges): ...new function.
10500 (pass_shorten_branches): Replace declaration with that of...
10501 (make_pass_shorten_branches): ...new function.
10502 (pass_set_nothrow_function_flags): Replace declaration with that of...
10503 (make_pass_set_nothrow_function_flags): ...new function.
10504 (pass_dwarf2_frame): Replace declaration with that of...
10505 (make_pass_dwarf2_frame): ...new function.
10506 (pass_final): Replace declaration with that of...
10507 (make_pass_final): ...new function.
10508 (pass_rtl_seqabstr): Replace declaration with that of...
10509 (make_pass_rtl_seqabstr): ...new function.
10510 (pass_release_ssa_names): Replace declaration with that of...
10511 (make_pass_release_ssa_names): ...new function.
10512 (pass_early_inline): Replace declaration with that of...
10513 (make_pass_early_inline): ...new function.
10514 (pass_inline_parameters): Replace declaration with that of...
10515 (make_pass_inline_parameters): ...new function.
10516 (pass_update_address_taken): Replace declaration with that of...
10517 (make_pass_update_address_taken): ...new function.
10518 (pass_convert_switch): Replace declaration with that of...
10519 (make_pass_convert_switch): ...new function.
10520 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10521 to a subclass of simple_ipa_opt_pass along with...
10522 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10523 (make_pass_ipa_tree_profile): ...new function.
10524 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10525 subclass of gimple_opt_pass along with...
10526 (pass_data_sra_early): ...new pass_data instance and...
10527 (make_pass_sra_early): ...new function.
10528 (pass_sra): Convert from a global struct to a subclass of
10529 gimple_opt_pass along with...
10530 (pass_data_sra): ...new pass_data instance and...
10531 (make_pass_sra): ...new function.
10532 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10533 gimple_opt_pass along with...
10534 (pass_data_early_ipa_sra): ...new pass_data instance and...
10535 (make_pass_early_ipa_sra): ...new function.
10536 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10537 subclass of gimple_opt_pass along with...
10538 (pass_data_ccp): ...new pass_data instance and...
10539 (make_pass_ccp): ...new function.
10540 (pass_fold_builtins): Convert from a global struct to a subclass of
10541 gimple_opt_pass along with...
10542 (pass_data_fold_builtins): ...new pass_data instance and...
10543 (make_pass_fold_builtins): ...new function.
10544 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10545 subclass of gimple_opt_pass along with...
10546 (pass_data_copy_prop): ...new pass_data instance and...
10547 (make_pass_copy_prop): ...new function.
10548 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10549 global struct to a subclass of gimple_opt_pass along with...
10550 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10551 (make_pass_rename_ssa_copies): ...new function.
10552 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10553 subclass of gimple_opt_pass along with...
10554 (pass_data_dce): ...new pass_data instance and...
10555 (make_pass_dce): ...new function.
10556 (pass_dce_loop): Convert from a global struct to a subclass of
10557 gimple_opt_pass along with...
10558 (pass_data_dce_loop): ...new pass_data instance and...
10559 (make_pass_dce_loop): ...new function.
10560 (pass_cd_dce): Convert from a global struct to a subclass of
10561 gimple_opt_pass along with...
10562 (pass_data_cd_dce): ...new pass_data instance and...
10563 (make_pass_cd_dce): ...new function.
10564 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10565 subclass of gimple_opt_pass along with...
10566 (pass_data_dominator): ...new pass_data instance and...
10567 (make_pass_dominator): ...new function.
10568 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10569 gimple_opt_pass along with...
10570 (pass_data_phi_only_cprop): ...new pass_data instance and...
10571 (make_pass_phi_only_cprop): ...new function.
10572 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10573 subclass of gimple_opt_pass along with...
10574 (pass_data_dse): ...new pass_data instance and...
10575 (make_pass_dse): ...new function.
10576 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10577 a subclass of gimple_opt_pass along with...
10578 (pass_data_forwprop): ...new pass_data instance and...
10579 (make_pass_forwprop): ...new function.
10580 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10581 struct to a subclass of gimple_opt_pass along with...
10582 (pass_data_tree_ifcombine): ...new pass_data instance and...
10583 (make_pass_tree_ifcombine): ...new function.
10584 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10585 subclass of gimple_opt_pass along with...
10586 (pass_data_ch): ...new pass_data instance and...
10587 (make_pass_ch): ...new function.
10588 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10589 subclass of gimple_opt_pass along with...
10590 (pass_data_tree_loop): ...new pass_data instance and...
10591 (make_pass_tree_loop): ...new function.
10592 (pass_tree_loop_init): Convert from a global struct to a subclass of
10593 gimple_opt_pass along with...
10594 (pass_data_tree_loop_init): ...new pass_data instance and...
10595 (make_pass_tree_loop_init): ...new function.
10596 (pass_lim): Convert from a global struct to a subclass of
10597 gimple_opt_pass along with...
10598 (pass_data_lim): ...new pass_data instance and...
10599 (make_pass_lim): ...new function.
10600 (pass_tree_unswitch): Convert from a global struct to a subclass of
10601 gimple_opt_pass along with...
10602 (pass_data_tree_unswitch): ...new pass_data instance and...
10603 (make_pass_tree_unswitch): ...new function.
10604 (pass_predcom): Convert from a global struct to a subclass of
10605 gimple_opt_pass along with...
10606 (pass_data_predcom): ...new pass_data instance and...
10607 (make_pass_predcom): ...new function.
10608 (pass_vectorize): Convert from a global struct to a subclass of
10609 gimple_opt_pass along with...
10610 (pass_data_vectorize): ...new pass_data instance and...
10611 (make_pass_vectorize): ...new function.
10612 (pass_graphite): Convert from a global struct to a subclass of
10613 gimple_opt_pass along with...
10614 (pass_data_graphite): ...new pass_data instance and...
10615 (make_pass_graphite): ...new function.
10616 (pass_graphite_transforms): Convert from a global struct to a subclass
10617 of gimple_opt_pass along with...
10618 (pass_data_graphite_transforms): ...new pass_data instance and...
10619 (make_pass_graphite_transforms): ...new function.
10620 (pass_check_data_deps): Convert from a global struct to a subclass of
10621 gimple_opt_pass along with...
10622 (pass_data_check_data_deps): ...new pass_data instance and...
10623 (make_pass_check_data_deps): ...new function.
10624 (pass_iv_canon): Convert from a global struct to a subclass of
10625 gimple_opt_pass along with...
10626 (pass_data_iv_canon): ...new pass_data instance and...
10627 (make_pass_iv_canon): ...new function.
10628 (pass_scev_cprop): Convert from a global struct to a subclass of
10629 gimple_opt_pass along with...
10630 (pass_data_scev_cprop): ...new pass_data instance and...
10631 (make_pass_scev_cprop): ...new function.
10632 (pass_record_bounds): Convert from a global struct to a subclass of
10633 gimple_opt_pass along with...
10634 (pass_data_record_bounds): ...new pass_data instance and...
10635 (make_pass_record_bounds): ...new function.
10636 (pass_complete_unroll): Convert from a global struct to a subclass of
10637 gimple_opt_pass along with...
10638 (pass_data_complete_unroll): ...new pass_data instance and...
10639 (make_pass_complete_unroll): ...new function.
10640 (pass_complete_unrolli): Convert from a global struct to a subclass of
10641 gimple_opt_pass along with...
10642 (pass_data_complete_unrolli): ...new pass_data instance and...
10643 (make_pass_complete_unrolli): ...new function.
10644 (pass_parallelize_loops): Convert from a global struct to a subclass
10645 of gimple_opt_pass along with...
10646 (pass_data_parallelize_loops): ...new pass_data instance and...
10647 (make_pass_parallelize_loops): ...new function.
10648 (pass_loop_prefetch): Convert from a global struct to a subclass of
10649 gimple_opt_pass along with...
10650 (pass_data_loop_prefetch): ...new pass_data instance and...
10651 (make_pass_loop_prefetch): ...new function.
10652 (pass_iv_optimize): Convert from a global struct to a subclass of
10653 gimple_opt_pass along with...
10654 (pass_data_iv_optimize): ...new pass_data instance and...
10655 (make_pass_iv_optimize): ...new function.
10656 (pass_tree_loop_done): Convert from a global struct to a subclass of
10657 gimple_opt_pass along with...
10658 (pass_data_tree_loop_done): ...new pass_data instance and...
10659 (make_pass_tree_loop_done): ...new function.
10660 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10661 struct to a subclass of gimple_opt_pass along with...
10662 (pass_data_cse_reciprocals): ...new pass_data instance and...
10663 (make_pass_cse_reciprocals): ...new function.
10664 (pass_cse_sincos): Convert from a global struct to a subclass of
10665 gimple_opt_pass along with...
10666 (pass_data_cse_sincos): ...new pass_data instance and...
10667 (make_pass_cse_sincos): ...new function.
10668 (pass_optimize_bswap): Convert from a global struct to a subclass of
10669 gimple_opt_pass along with...
10670 (pass_data_optimize_bswap): ...new pass_data instance and...
10671 (make_pass_optimize_bswap): ...new function.
10672 (pass_optimize_widening_mul): Convert from a global struct to a
10673 subclass of gimple_opt_pass along with...
10674 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10675 (make_pass_optimize_widening_mul): ...new function.
10676 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10677 subclass of gimple_opt_pass along with...
10678 (pass_data_phiopt): ...new pass_data instance and...
10679 (make_pass_phiopt): ...new function.
10680 (pass_cselim): Convert from a global struct to a subclass of
10681 gimple_opt_pass along with...
10682 (pass_data_cselim): ...new pass_data instance and...
10683 (make_pass_cselim): ...new function.
10684 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10685 subclass of gimple_opt_pass along with...
10686 (pass_data_phiprop): ...new pass_data instance and...
10687 (make_pass_phiprop): ...new function.
10688 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10689 subclass of gimple_opt_pass along with...
10690 (pass_data_pre): ...new pass_data instance and...
10691 (make_pass_pre): ...new function.
10692 (pass_fre): Convert from a global struct to a subclass of
10693 gimple_opt_pass along with...
10694 (pass_data_fre): ...new pass_data instance and...
10695 (make_pass_fre): ...new function.
10696 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10697 subclass of gimple_opt_pass along with...
10698 (pass_data_reassoc): ...new pass_data instance and...
10699 (make_pass_reassoc): ...new function.
10700 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10701 subclass of gimple_opt_pass along with...
10702 (pass_data_sink_code): ...new pass_data instance and...
10703 (make_pass_sink_code): ...new function.
10704 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10705 subclass of gimple_opt_pass along with...
10706 (pass_data_strlen): ...new pass_data instance and...
10707 (make_pass_strlen): ...new function.
10708 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10709 struct to a subclass of gimple_opt_pass along with...
10710 (pass_data_build_alias): ...new pass_data instance and...
10711 (make_pass_build_alias): ...new function.
10712 (pass_build_ealias): Convert from a global struct to a subclass of
10713 gimple_opt_pass along with...
10714 (pass_data_build_ealias): ...new pass_data instance and...
10715 (make_pass_build_ealias): ...new function.
10716 (pass_ipa_pta): Convert from a global struct to a subclass of
10717 simple_ipa_opt_pass along with...
10718 (pass_data_ipa_pta): ...new pass_data instance and...
10719 (make_pass_ipa_pta): ...new function.
10720 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10721 subclass of gimple_opt_pass along with...
10722 (pass_data_uncprop): ...new pass_data instance and...
10723 (make_pass_uncprop): ...new function.
10724 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10725 global struct to a subclass of gimple_opt_pass along with...
10726 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10727 (make_pass_late_warn_uninitialized): ...new function.
10728 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10729 to a subclass of gimple_opt_pass along with...
10730 (pass_data_init_datastructures): ...new pass_data instance and...
10731 (make_pass_init_datastructures): ...new function.
10732 (pass_early_warn_uninitialized): Convert from a global struct to a
10733 subclass of gimple_opt_pass along with...
10734 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10735 (make_pass_early_warn_uninitialized): ...new function.
10736 (pass_update_address_taken): Convert from a global struct to a
10737 subclass of gimple_opt_pass along with...
10738 (pass_data_update_address_taken): ...new pass_data instance and...
10739 (make_pass_update_address_taken): ...new function.
10740 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10741 struct to a subclass of gimple_opt_pass along with...
10742 (pass_data_release_ssa_names): ...new pass_data instance and...
10743 (make_pass_release_ssa_names): ...new function.
10744 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10745 subclass of gimple_opt_pass along with...
10746 (pass_data_stdarg): ...new pass_data instance and...
10747 (make_pass_stdarg): ...new function.
10748 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10749 global struct to a subclass of gimple_opt_pass along with...
10750 (pass_data_convert_switch): ...new pass_data instance and...
10751 (make_pass_convert_switch): ...new function.
10752 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10753 to a subclass of gimple_opt_pass along with...
10754 (pass_data_tail_recursion): ...new pass_data instance and...
10755 (make_pass_tail_recursion): ...new function.
10756 (pass_tail_calls): Convert from a global struct to a subclass of
10757 gimple_opt_pass along with...
10758 (pass_data_tail_calls): ...new pass_data instance and...
10759 (make_pass_tail_calls): ...new function.
10760 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10761 struct to a subclass of gimple_opt_pass along with...
10762 (pass_data_lower_vector): ...new pass_data instance and...
10763 (make_pass_lower_vector): ...new function.
10764 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10765 gimple_opt_pass along with...
10766 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10767 (make_pass_lower_vector_ssa): ...new function.
10768 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10769 to a subclass of gimple_opt_pass along with...
10770 (pass_data_slp_vectorize): ...new pass_data instance and...
10771 (make_pass_slp_vectorize): ...new function.
10772 (pass_ipa_increase_alignment): Convert from a global struct to a
10773 subclass of simple_ipa_opt_pass along with...
10774 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10775 (make_pass_ipa_increase_alignment): ...new function.
10776 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10777 gimple_opt_pass along with...
10778 (pass_data_vrp): ...new pass_data instance and...
10779 (make_pass_vrp): ...new function.
10780 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10781 subclass of simple_ipa_opt_pass along with...
10782 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10783 (make_pass_ipa_free_lang_data): ...new function.
10784 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10785 gimple_opt_pass along with...
10786 (pass_data_tsan): ...new pass_data instance and...
10787 (make_pass_tsan): ...new function.
10788 (pass_tsan_O0): Convert from a global struct to a subclass of
10789 gimple_opt_pass along with...
10790 (pass_data_tsan_O0): ...new pass_data instance and...
10791 (make_pass_tsan_O0): ...new function.
10792 * var-tracking.c (pass_variable_tracking): Convert from a global
10793 struct to a subclass of rtl_opt_pass along with...
10794 (pass_data_variable_tracking): ...new pass_data instance and...
10795 (make_pass_variable_tracking): ...new function.
10796 * web.c (pass_web): Convert from a global struct to a subclass of
10797 rtl_opt_pass along with...
10798 (pass_data_web): ...new pass_data instance and...
10799 (make_pass_web): ...new function.
10800 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10801 declaration with that of...
10802 (make_pass_mode_switch_use): ...new function.
10803 (pass_resolve_sw_modes): Replace declaration with that of...
10804 (make_pass_resolve_sw_modes): ...new function.
10805 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10806 from a global struct to a subclass of rtl_opt_pass along with...
10807 (pass_data_mode_switch_use): ...new pass_data instance and...
10808 (make_pass_mode_switch_use): ...new function.
10809 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10810 from a global struct to a subclass of rtl_opt_pass along with...
10811 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10812 (make_pass_resolve_sw_modes): ...new function.
10813 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10814 struct to a subclass of rtl_opt_pass along with...
10815 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10816 (make_pass_insert_vzeroupper): ...new function.
10817 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10818 global struct to a subclass of rtl_opt_pass along with...
10819 (pass_data_work_around_errata): ...new pass_data instance and...
10820 (make_pass_work_around_errata): ...new function.
10821 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10822 struct to a subclass of rtl_opt_pass along with...
10823 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10824 (make_pass_mips_machine_reorg2): ...new function.
10825
6a389ed5
DM		 108262013-08-05 David Malcolm <dmalcolm@redhat.com>
10827
10828 * passes.c (pass_manager::operator new): New.
10829
f7695dbf
DM		 108302013-08-05 David Malcolm <dmalcolm@redhat.com>
10831
10832 Handwritten part of conversion of passes to C++ classes.
10833
10834 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10835 (toplev.o): Add dep on PASS_MANAGER_H.
10836 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10837 of early local pases to reflect this moving from a global to a
10838 member of gcc::pass_manager.
10839 (cgraph_add_new_function): Likewise.
10840 * lto-cgraph.c (lto_output_node): Update for conversion of
10841 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10842 * passes.c (opt_pass::clone): New.
10843 (opt_pass::gate): New.
10844 (opt_pass::execute): New.
10845 (opt_pass::opt_pass): New.
10846 (pass_manager::execute_early_local_passes): New.
10847 (pass_manager::execute_pass_mode_switching): new.
10848 (finish_optimization_passes): Convert to...
10849 (pass_manager::finish_optimization_passes): ...this.
10850 (finish_optimization_passes): Update for conversion of passes to
10851 C++ classes.
10852 (register_dump_files_1): Use has_gate since we cannot portably
10853 check a vtable entry against NULL.
10854 (dump_one_pass): Likewise.
10855 (ipa_write_summaries_2): Likewise.
10856 (ipa_write_optimization_summaries_1): Likewise.
10857 (ipa_read_summaries_1): Likewise.
10858 (ipa_read_optimization_summaries_1): Likewise.
10859 (execute_ipa_stmt_fixups): Likewise.
10860 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10861 pass-creation functions rather than wiring up globals, and
10862 storing the results in fields of pass_manager generated using
10863 pass-instances.def.
10864 (pass_manager::dump_profile_report): Update for conversion of
10865 passes to C++ classes.
10866 (pass_manager::execute_ipa_summary_passes): Likewise.
10867 (execute_one_ipa_transform_pass): Likewise.
10868 (execute_one_pass): Use has_gate and has_execute since we cannot
10869 portably check a vtable entry against NULL.
10870 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10871 (pass_manager): Use pass-instances.def to add fields for the
10872 various pass instances.
10873 * toplev.c (finalize): Update for move of
10874 finish_optimization_passes to a method of gcc::pass_manager.
10875 * toplev.h (finish_optimization_passes): Move to method of class
10876 pass_manager.
10877 * tree-pass.h (struct pass_data): New.
03b0ee0a 10878 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 10879 (opt_pass::gate): Convert to virtual function.
10880 (opt_pass::~opt_pass): New.
10881 (opt_pass::clone): New.
10882 (opt_pass::execute): Convert to virtual function.
10883 (opt_pass::opt_pass): New.
10884 (opt_pass::ctxt_): new.
10885 (gimple_opt_pass): Convert to subclass of opt_pass.
10886 (gimple_opt_pass::gimple_opt_pass): New.
10887 (rtl_opt_pass): Convert to subclass of opt_pass.
10888 (rtl_opt_pass::rtl_opt_pass): New.
10889 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10890 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10891 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10892 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10893 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10894 invocation of pass_mode_switching to reflect this moving from a
10895 global to a member of gcc::pass_manager.
10896 (ix86_option_override): Rework how pass_insert_vzeroupper is
10897 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 10898 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 10899
d0b2f831
RE		 109002013-08-05 Richard Earnshaw <rearnsha@arm.com>
10901
10902 PR rtl-optimization/57708
10903 * recog.c (peep2_find_free_register): Validate all regs in a
10904 multi-reg mode.
10905
51a5c0c2
JH		 109062013-08-05 Jan Hubicka <jh@suse.cz>
10907
10908 PR lto/57602
03b0ee0a
UB		 10909 * cgraph.c (verify_cgraph_node): Accept local flags from other
10910 partitions.
51a5c0c2
JH		 10911 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10912 (function_and_variable_visibility): Likewise.
10913 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10914
65f0a120
GDR		 109152013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10916
10917 * graph.c (init_graph_slim_pretty_print): Remove.
10918 (print_graph_cfg): Do not call it. Use local pretty printer.
10919 (start_graph_dump): Likewise.
10920
b3f80694
GDR		 109212013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10922
10923 * gimple-pretty-print.c (buffer): Remove.
10924 (initialized): Likewise.
10925 (maybe_init_pretty_print): Likewise.
10926 (print_gimple_stmt): Do not call it. Use non-static local
10927 pretty_printer variable.
10928 (print_gimple_expr): Likewise.
10929 (print_gimple_seq): Likewise.
10930 (gimple_dump_bb): Likewise.
10931
11a877b3
GDR		 109322013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10933
10934 * asan.c (asan_pp): Remove.
10935 (asan_pp_initialized): Likewise.
10936 (asan_pp_initialize): Likewise.
10937 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10938 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10939 (asan_add_global): Likewise.
10940
b066401f
GDR		 109412013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10942
10943 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10944 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10945 * pretty-print.c (pp_formatted_text_data): Likewise.
10946 (pp_indent): Rename from pp_base_indent.
10947 (pp_format): Rename from pp_base_format.
10948 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10949 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10950 (pp_flush): Rename from pp_base_flush.
10951 (pp_set_line_maximum_length): Rename from
10952 pp_base_set_line_maximum_length.
10953 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10954 (pp_set_prefix): Rename from pp_base_set_prefix.
10955 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10956 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10957 (pp_append_text): Rename from pp_base_append_text.
10958 (pp_formatted_text): Rename from pp_base_formatted_text.
10959 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10960 (pp_remaining_character_count_for_line): Rename from
10961 pp_base_remaining_character_count_for_line.
10962 (pp_newline): Rename from pp_base_newline.
10963 (pp_character): Rename from pp_base_character.
10964 (pp_string): Rename from pp_base_string.
10965 (pp_maybe_space): Rename from pp_base_maybe_space.
10966 * asan.c (asan_pp_string): Adjust.
10967 (asan_emit_stack_protection): Likewise.
10968 (asan_add_global): Likewise.
10969 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10970 * tree-mudflap.c (mf_varname_tree): Likewise.
10971 * tree-pretty-print.c (pp_tree_identifier): Rename from
10972 pp_base_tree_identifier.
10973 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10974 Declare as function.
10975
137a1a27
GDR		 109762013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10977
10978 * pretty-print.h (pp_bar_bar): New.
10979 (pp_ampersand_ampersand): Likewise.
10980 (pp_less_equal): Likewise.
10981 (pp_greater_equal): Likewise.
10982 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10983 printer functions instead of pp_string or operators and punctuators.
10984 (dump_gimple_call): Likewise.
10985 (dump_gimple_omp_for): Likewise.
10986 (dump_gimple_transaction): Likewise.
10987 (dump_gimple_phi): Likewise.
10988 (pp_gimple_stmt_1): Likewise.
10989 * sched-vis.c (print_insn): Likewise.
10990 * tree-mudflap.c (mf_varname_tree): Likewise.
10991 * tree-pretty-print.c (dump_block_node): Likewise.
10992 (dump_generic_node): Likewise.
10993
815effe1
JH		 109942013-08-02 Jan Hubicka <jh@suse.cz>
10995
10996 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10997 boundaries.
10998 * lto-streamer-out.c (tree_is_indexable): Results decls and
10999 parm decls are not indexable.
11000 (DFS_write_tree_body): Do not follow args and results.
11001 (hash_tree): Likewise.
11002 (output_functions): Rearrange so struct function is needed
11003 only when real body is output; be able to also ouptut abstract
11004 functions; output DECL_ARGUMENTS and DECL_RESULT.
11005 (lto_output): When not in WPA, ale store abstract functions.
11006 (write_symbol): Do not care about RESULT_DECL.
11007 (output_symbol_p): Handle correctly sbtract decls.
11008 * lto-streamer-in.c (input_function): Rearrange so struct
11009 function can be NULL at entry; allow streaming of
11010 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11011 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11012 sanity check during LTO.
11013 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11014 RESULT_DECl and DECL_ARGUMENTS.
11015 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11016 Likewise.
11017
07838b13
GDR		 110182013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11019
11020 * pretty-print.h (pp_underscore): New.
11021 (pp_comma): Tidy.
11022 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11023 printer functions instead of pp_character.
11024 (dump_binary_rhs): Likewise.
11025 (dump_ternary_rhs): Likewise.
11026 (dump_gimple_call_args): Likewise.
11027 (pp_points_to_solution): Likewise.
11028 (dump_gimple_call): Likewise.
11029 (dump_gimple_switch): Likewise.
11030 (dump_gimple_cond): Likewise.
11031 (dump_gimple_bind): Likewise.
11032 (dump_gimple_try): Likewise.
11033 (dump_gimple_omp_for): Likewise.
11034 (dump_gimple_omp_continue): Likewise.
11035 (dump_gimple_omp_single): Likewise.
11036 (dump_gimple_omp_sections): Likewise.
11037 (dump_gimple_omp_block): Likewise.
11038 (dump_gimple_omp_critical): Likewise.
11039 (dump_gimple_transaction): Likewise.
11040 (dump_gimple_asm): Likewise.
11041 (dump_gimple_phi): Likewise.
11042 (dump_gimple_omp_parallel): Likewise.
11043 (dump_gimple_omp_task): Likewise.
11044 (dump_gimple_omp_atomic_load): Likewise.
11045 (dump_gimple_omp_atomic_store): Likewise.
11046 (dump_gimple_mem_ops): Likewise.
11047 (pp_gimple_stmt_1): Likewise.
11048 (pp_cfg_jump): Likewise.
11049 (dump_implicit_edges): Likewise.
11050 (gimple_dump_bb_for_graph): Likewise.
11051 * graph.c (draw_cfg_node): Likewise.
11052 * langhooks.c (lhd_print_error_function): Likewise.
11053 * sched-vis.c (print_exp): Likewise.
11054 (print_value): Likewise.
11055 (print_pattern): Likewise.
11056 (print_insn): Likewise.
11057 (rtl_dump_bb_for_graph): Likewise.
11058 * tree-pretty-print.c (dump_function_declaration): Likewise.
11059 (dump_array_domain): Likewise.
11060 (dump_omp_clause): Likewise.
11061 (dump_location): Likewise.
11062 (dump_generic_node): Likewise.
11063 (print_struct_decl): Likewise.
11064 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11065
bb0d2039
BS		 110662013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11067
11068 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11069 candidate table when replacing a candidate statement.
11070 (replace_rhs_if_not_dup): Likewise.
11071 (replace_one_candidate): Likewise.
11072
bc0ec027 110732013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 11074 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 11075
11076 * cgraphunit.c (add_new_function): Fix logic when adding from
11077 late IPA pass.
11078 (assemble_thunk): Rename to ...
03b0ee0a
UB		 11079 (expand_thunk); .. this one; export; get it working with
11080 general functions; make produced gimple valid.
bc0ec027
JH		 11081 * cgraph.h (expand_thunk): Declare.
11082
0e8853ee
JH		 110832013-08-02 Jan Hubicka <jh@suse.cz>
11084
03b0ee0a
UB		 11085 * ipa-cp.c (gather_context_independent_values): Use
11086 ipa_get_param_move_cost.
11087 (get_replacement_map): Remove PARAM; move parameter folding
11088 into tree-inline.c
0e8853ee
JH		 11089 (create_specialized_node): Update.
11090 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11091 assert that we have gimple body; update move_cost.
11092 (count_formal_params): Assert that we have gimple body.
11093 (ipa_dump_param): New function.
11094 (ipa_alloc_node_params): Break out from ...
11095 (ipa_initialize_node_params): ... here.
11096 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11097 (ipa_write_node_info): Stream move costs.
11098 (ipa_read_node_info): Read move costs.
11099 (ipa_update_after_lto_read): Do not recompute node params.
11100 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11101 (ipa_get_param): Check we are not in WPA.
11102 (ipa_get_param_move_cost): New.
11103 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11104 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11105 parm numbers to be present.
11106
ecee672b
VM		 111072013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11108
11109 PR rtl-optimization/58048
11110 * lra-constraints.c (process_alt_operands): Don't check asm
11111 operand on register.
11112
13ec0527
EB		 111132013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11114
11115 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11116 the implied StoreLoad barrier for atomic operations if before.
11117
af15184a
JH		 111182013-08-02 Jan Hubicka <jh@suse.cz>
11119 Martin Liska <marxin.liska@gmail.com>
11120
03b0ee0a
UB		 11121 * cgraph.c (cgraph_function_body_availability): Do not check
11122 cgraph flags.
af15184a
JH		 11123 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11124 symtab_node_availability): Declare.
11125 * ipa.c (can_replace_by_local_alias): New.
11126 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 11127 * symtab.c (symtab_for_node_and_aliases,
11128 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 11129
01e54ef8
VM		 111302013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11131
11132 PR rtl-optimization/57963
03b0ee0a 11133 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 11134 (lra_constraints): Use them.
11135
e2fd7ca7
SN		 111362013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11137
03b0ee0a
UB		 11138 * config/arm/types.md (define_attr "type"): Add "load_acq"
11139 and "store_rel".
e2fd7ca7
SN		 11140 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11141 changes.
11142 (cortex_a53_store1): Likewise.
11143
0dcc26c3
JH		 111442013-08-01 Jan Hubicka <jh@suse.cz>
11145
03b0ee0a
UB		 11146 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11147 partitions are not needed.
0dcc26c3 11148
0bd72901
UB		 111492013-08-01 Uros Bizjak <ubizjak@gmail.com>
11150
11151 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11152 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11153 MAYBE_NON_Q_CLASS_P where appropriate.
11154
12123452
JH		 111552013-08-01 Jan Hubicka <jh@suse.cz>
11156
11157 * cgraph.h (release_function_body): Declare.
11158 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11159 of unused delcarations.
11160
ce852f9c
KT		 111612013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11162
03b0ee0a
UB		 11163 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11164 RTL form when subtracting a constant.
ce852f9c 11165
c743b246
KT		 111662013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11167
11168 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11169 Generate canonical plus rtx with negated immediate instead of minus
11170 where appropriate.
11171 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11172
c0c123ef
JH		 111732013-08-01 Jan Hubicka <jh@suse.cz>
11174
11175 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11176 (cgraph_release_function_body): Likewise.
11177 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11178 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11179 to used_as_abstract_origin.
11180 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11181 symbols used as abstract origins.
11182 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 11183 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11184 used_as_abstract_origin.
c0c123ef 11185 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 11186 used_as_abstract_origin; be ready for DECL_RESULT and
11187 DECL_ARGUMENTS to be NULL.
c0c123ef 11188
0bd72901
UB		 11189 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11190 for abstract functions.
11191 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11192 real symbols.
c0c123ef 11193
1f029433
JH		 111942013-08-01 Jan Hubicka <jh@suse.cz>
11195
11196 * profile.c (compute_value_histograms): Fix thinko.
11197
c451f4d6
SN		 111982013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11199
11200 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11201 aarch-common-protos.h to extra_headers.
11202 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11203 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11204 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11205
34b512f5
SN		 112062013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11207
11208 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11209 Include "../arm/types.md". Define "type" attribute for all patterns.
11210 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11211 attribute changes.
11212
d86e633a
MM		 112132013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11214
11215 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11216 to support power8 load fusion.
11217 (fusion_gpr_mem_load): Likewise.
11218
11219 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11220
11221 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11222 declarations for power8 load fusion.
11223 (emit_fusion_gpr_load): Likewise.
11224
11225 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11226 tuning for power8, turn on fusion mode by default. Turn on sign
11227 extending fusion mode if normal fusion mode is on, and we are at
11228 -O2 or -O3.
11229 (fusion_gpr_load_p): New function, return true if we can fuse an
11230 addis instruction with a dependent load to a GPR.
11231 (emit_fusion_gpr_load): Emit the instructions for power8 load
11232 fusion to GPRs.
11233
0bd72901 11234 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 11235 (VSX load fusion peepholes): New peepholes to fuse together an
11236 addi instruction with a VSX load instruction.
11237
11238 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11239 peepholes to fuse an addis instruction with a load to a GPR base
11240 register. If we are supporting sign extending fusions, convert
11241 sign extending loads to zero extending loads and add an explicit
11242 sign extension.
11243
c3f35647
SN		 112442013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11245
11246 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11247 aarch-common-protos.h to extra_headers.
11248 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11249 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11250 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11251 (arm_no_early_alu_shift_dep): Likewise.
11252 (arm_no_early_alu_shift_value_dep): Likewise.
11253 (arm_no_early_mul_dep): Likewise.
11254 (arm_no_early_store_addr_dep): Likewise.
11255 (arm_mac_accumulator_is_mul_result): Likewise.
11256 (arm_mac_accumulator_is_result): Likewise.
c3f35647 11257 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 11258 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11259 here to ...
c3f35647 11260 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11261 (arm_no_early_alu_shift_dep): Likewise.
11262 (arm_no_early_alu_shift_value_dep): Likewise.
11263 (arm_no_early_mul_dep): Likewise.
11264 (arm_no_early_store_addr_dep): Likewise.
11265 (arm_mac_accumulator_is_mul_result): Likewise.
11266 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN		 11267 * config/arm/aarch-common-protos.h: ... here. New file.
11268 * config/arm/t-arm (aarch-common.o): Define.
11269
112702013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11271
11272 * config/arm/arm.md: Include new file "types.md".
11273 (define_attr "type"): Move from here to ...
11274 (define_attr "mul32"): Likewise.
11275 (define_attr "mul64"): Likewise.
11276 * config/arm/types.md: ... here. New file.
11277
228c1313
SH		 112782013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11279
11280 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11281 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11282
d6e9046d 112832013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 11284
11285 * gen-pass-instances.awk: Fix offset of substr().
11286
a167b052
DM		 112872013-07-31 David Malcolm <dmalcolm@redhat.com>
11288
11289 * Makefile.in (pass-instances.def): New.
11290 (passes.o): Replace dependency on passes.def with one on
11291 pass-instances.def
11292
11293 * gen-pass-instances.awk: New.
11294
11295 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11296 rather than passes.def, updating local definition of NEXT_PASS
11297 macro to add an extra NUM parameter (currently unused).
11298
315f8c0e
DM		 112992013-07-30 David Malcolm <dmalcolm@redhat.com>
11300
11301 * Makefile.in (PASS_MANAGER_H): New.
11302 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11303 (passes.o): Likewise.
11304 (statistics.o): Likewise.
11305 (cgraphunit.o): Likewise.
11306 (context.o): Depend on PASS_MANAGER_H.
11307
11308 * pass_manager.h: New.
11309
11310 * cgraphunit.c (cgraph_add_new_function): Update for moves
11311 of globals to fields of pass_manager.
11312 (analyze_function): Likewise.
11313 (expand_function): Likewise.
11314 (ipa_passes): Likewise.
11315 (compile): Likewise.
11316
11317 * context.c (context::context): New.
11318 * context.h (context::context): New.
11319 (context::get_passes): New.
11320 (context::passes_): New.
11321
11322 * lto-cgraph.c (input_node): Update for moves of globals to
11323 fields of pass_manager.
11324
11325 * passes.c (all_passes): Remove, in favor of a field of the
11326 same name within the new class pass_manager.
11327 (all_small_ipa_passes): Likewise.
11328 (all_lowering_passes): Likewise.
11329 (all_regular_ipa_passes): Likewise.
11330 (all_late_ipa_passes): Likewise.
11331 (all_lto_gen_passes): Likewise.
11332 (passes_by_id): Likewise.
11333 (passes_by_id_size): Likewise.
11334 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11335 the new class pass_manager.
11336 (set_pass_for_id): Convert to...
11337 (pass_manager::set_pass_for_id): ...method.
11338 (get_pass_for_id): Convert to...
11339 (pass_manager::get_pass_for_id): ...method.
11340 (register_one_dump_file): Move body of implementation into...
11341 (pass_manager::register_one_dump_file): ...here.
11342 (register_dump_files_1): Convert to...
11343 (pass_manager::register_dump_files_1): ...method.
11344 (register_dump_files): Convert to...
11345 (pass_manager::register_dump_files): ...method.
11346 (create_pass_tab): Update for moves of globals to fields of
11347 pass_manager.
11348 (dump_passes): Move body of implementation into...
11349 (pass_manager::dump_passes): ...here.
11350 (register_pass): Move body of implementation into...
11351 (pass_manager::register_pass): ...here.
11352 (init_optimization_passes): Convert into...
11353 (pass_manager::pass_manager): ...constructor for new
11354 pass_manager class, and initialize the pass_lists array.
11355 (check_profile_consistency): Update for moves of globals to
11356 fields of pass_manager.
11357 (dump_profile_report): Move body of implementation into...
11358 (pass_manager::dump_profile_report): ...here.
11359 (ipa_write_summaries_1): Update for moves of pass lists from
11360 being globals to fields of pass_manager.
11361 (ipa_write_optimization_summaries): Likewise.
11362 (ipa_read_summaries): Likewise.
11363 (ipa_read_optimization_summaries): Likewise.
11364 (execute_all_ipa_stmt_fixups): Likewise.
11365
11366 * statistics.c (statistics_fini): Update for moves of globals to
11367 fields of pass_manager.
11368
11369 * toplev.c (general_init): Replace call to
11370 init_optimization_passes with construction of the pass_manager
11371 instance.
11372
11373 * tree-pass.h (all_passes): Remove, in favor of a field of the
11374 same name within the new class pass_manager.
11375 (all_small_ipa_passes): Likewise.
11376 (all_lowering_passes): Likewise.
11377 (all_regular_ipa_passes): Likewise.
11378 (all_lto_gen_passes): Likewise.
11379 (all_late_ipa_passes): Likewise.
11380 (passes_by_id): Likewise.
11381 (passes_by_id_size): Likewise.
11382 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11383 the new class pass_manager.
11384 (get_pass_for_id): Remove.
11385
02cab1c1
RE		 113862013-07-30 Richard Earnshaw <rearnsha@arm.com>
11387
11388 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11389 configs.
11390
b2d3f886
RE		 113912013-07-30 Richard Earnshaw <rearnsha@arm.com>
11392
11393 * arm.md (mulhi3): New expand pattern.
11394
605e86fa
JH		 113952013-07-30 Jan Hubicka <jh@suse.cz>
11396 Martin Liska <marxin.liska@gmail.com>
11397
11398 * profile.c (compute_value_histograms): Do not ICE when
11399 there is mismatch only on some counters.
11400
14a87636
ZC		 114012013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11402
11403 PR rtl-optimization/57637
11404 * function.c (move_insn_for_shrink_wrap): Also check the
11405 GEN set of the LIVE problem for the liveness analysis
11406 if it exists, otherwise give up.
11407
0100cd3f
BS		 114082013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11409
11410 PR tree-optimization/57993
11411 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11412 replaced statement in the candidate table.
11413 (phi_add_costs): Return infinite cost when the hidden basis does
11414 not dominate all phis on which the candidate is dependent.
11415 (replace_one_candidate): Record replaced statement in the
11416 candidate table.
11417
f0d811f7
JR		 114182013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11419
11420 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11421 (ashlv2si3): New expander.
11422 (*ashlv2si3_i): New define_insn_and_split.
11423 * predicates.md (float_operation): Allow patterns with three
11424 basic sub-patterns.
11425
bba33211
JR		 11426 PR rtl-optimization/58021
11427 * mode-switching.c (create_pre_exit): Always split off preceding
11428 insns if we are not at the basic block head.
11429
ff3f3951
MR		 114302013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11431
11432 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11433 (UCLIBC_DYNAMIC_LINKER): New macro.
11434 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11435 `-mnan=2008'.
11436 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11437 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11438 `-mnan=2008'.
11439 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11440 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11441 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11442 for SF and DF modes. Use ieee_quad_format for TF mode.
11443 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11444 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11445 (mips_option_override): Handle `-mnan=legacy'.
11446 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11447 `-mabs=2008' and `-mnan=2008'.
11448 (OPTION_DEFAULT_SPECS): Add "nan" default.
11449 (ASM_SPEC): Handle `-mnan='.
11450 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11451 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11452 comment accordingly.
11453 (neg<mode>2): Likewise.
11454 * config/mips/mips.opt (mabs, mnan): New options.
11455 * doc/install.texi (Configuration): Document `--with-nan=' option.
11456 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11457 `-mnan=' options.
11458 (MIPS Options): Document them.
11459 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11460 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11461 * configure: Regenerate.
11462 * config.in: Regenerate.
11463
576e2f37
UB		 114642013-07-29 Uros Bizjak <ubizjak@gmail.com>
11465
11466 * config/i386/i386.md (float post-reload splitters): Do not check
11467 for subregs of SSE registers.
11468
114692013-07-29 Uros Bizjak <ubizjak@gmail.com>
11470 H.J. Lu <hongjiu.lu@intel.com>
11471
11472 PR target/57954
11473 PR target/57988
11474 * config/i386/i386.md (post-reload splitter
11475 to avoid partial SSE reg dependency stalls): New pattern.
11476
a71f0749
DV		 114772013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11478
11479 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11480
4c97f1cc
JR		 114812013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11482
11483 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11484 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11485
b0c733d8
CC		 114862013-07-26 Cary Coutant <ccoutant@google.com>
11487
11488 * dwarf2out.c (die_checksum_ordered): Don't include template
11489 instantiations in signature.
11490 (is_template_parameter): New function.
11491 (is_template_instantiation): New function.
11492 (generate_skeleton_bottom_up): Don't include template instantiations
11493 in type unit DIE.
11494 (generate_skeleton): Likewise.
11495 (break_out_comdat_types): Move recursive call to break out nested
11496 types earlier.
11497 (prune_unused_types_mark_generic_parms_dies): Call
11498 is_template_parameter.
11499
040d8a1c
IB		 115002013-07-26 Ian Bolton <ian.bolton@arm.com>
11501
11502 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11503 uses vector registers.
89fdc743 11504 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11505
85bd4ac6 115062013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11507 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 11508
11509 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11510 where appropriate.
85bd4ac6
KT		 11511 * config/arm/arm.md (movcond_addsi): New splitter.
11512
b6051207
SA		 115132013-07-25 Sterling Augustine <saugustine@google.com>
11514
11515 * dwarf2out.c (size_of_pubnames): Move code to...
11516 (include_pubname_in_output): ...here. New.
11517 (want_pubnames): Rearrange.
11518 (output_pubnames): Call include_pubname_in_output. Move assertion.
11519
41a7c215
CM		 115202013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11521
11522 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11523
115242013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11525
11526 PR target/38836
11527 * doc/extend.texi: Remove obsolete builtins. Fix
11528 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11529
49bde175
JH		 115302013-07-25 Jan Hubicka <jh@suse.cz>
11531
11532 * cgraph.c (release_function_body): Break out from ...
11533 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11534 and DECL_ARGUMENTS.
11535 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11536 old_tree in the map.
11537 (create_specialized_node): Update.
11538 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11539 into index.
41a7c215
CM		 11540 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11541 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 11542 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11543 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11544 DECL_RESULT.
11545
4870352d
KT		 115462013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11547
11548 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11549 addsi3_carryin_alt2_<optab>): Correct output template.
11550
1d6aee1c
KT		 115512013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11552
11553 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11554 Adjust for arm_restrict_it.
11555 Remove trailing whitespace.
11556
c7cafd75
MK		 115572013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11558
017d38f5
MK		 11559 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11560 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11561
c7cafd75
MK		 11562 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11563
00b7527b
VM		 115642013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11565
11566 PR rtl-optimization/57960
11567 * lra-constraints.c (process_alt_operands): Use the right mode
11568 when checking strict_low.
11569
4f63dfc6
JH		 115702013-07-25 Jan Hubicka <jh@suse.cz>
11571
11572 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11573 * cgraph.c (cgraph_remove_node): Do not release function body
11574 when in cgraph streaming.
11575 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 11576 in other partitions are not considered reachable; fix handling of
11577 clones.
4f63dfc6 11578
e6dcfa18
RR		 115792013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11580
11581 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11582
0b93d3b6
RR		 115832013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11584
11585 PR target/19599
11586 PR target/57731
bb80c2eb 11587 PR target/57837
e6dcfa18 11588 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11589 Ss with US. Adjust output for v5 and v4t.
41a7c215 11590 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11591
0b93d3b6
RR		 11592 * config/arm/constraints.md ("Ss"): Rename to US.
11593
690688b3
TG		 115942013-07-25 Terry Guo <terry.guo@arm.com>
11595
11596 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11597 shift_add/shift_sub0/shift_sub1 RTXs.
11598
a4ad093b 115992013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11600 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 11601
11602 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11603 (altivec_vpks<VI_char>ss): Likewise.
11604 (altivec_vpks<VI_char>us): Likewise.
11605 (altivec_vpku<VI_char>us): Likewise.
11606 (altivec_vpku<VI_char>um): Likewise.
11607
8c5005ce
DM		 116082013-07-24 David Malcolm <dmalcolm@redhat.com>
11609
11610 Introduce context class.
11611
11612 * Makefile.in (CONTEXT_H): New.
11613 (OBJS): Add context.o.
11614 (toplev.o): Add CONTEXT_H to dependencies.
11615 (context.o): New.
11616
41a7c215 11617 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 11618
11619 * context.c: New.
11620
11621 * context.h: New.
11622
ce4a9422
JR		 116232013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11624
11625 PR rtl-optimization/57968
11626 * mode-switching.c (create_pre_exit): Allow instructions that
11627 don't set a return register to need a non-exit mode.
11628
493f4c9e 116292013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11630 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11631
b5860fd3
WS		 11632 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11633 operands to vperm for little endian.
11634 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11635 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11636
73792b92 116372013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11638 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 11639
11640 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11641 two operands for little-endian.
11642
14c21302
SE		 116432013-07-23 Steve Ellcey <sellcey@mips.com>
11644
11645 * config/mips/mips.c (mips_case_values_threshold): New.
11646 (TARGET_CASE_VALUES_THRESHOLD): Define.
11647
a6056198 116482013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11649 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 11650
11651 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11652 selection of field for vector splat in little endian mode.
11653
dd7a40e1
MM		 116542013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11655
11656 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11657 expanders to rs6000.md.
11658 (ior<mode>3): Likewise.
11659 (and<mode>3): Likewise.
11660 (one_cmpl<mode>2): Likewise.
11661 (nor<mode>3): Likewise.
11662 (andc<mode>3): Likewise.
11663 (eqv<mode>3): Likewise.
11664 (nand<mode>3): Likewise.
11665 (orc<mode>3): Likewise.
11666
11667 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11668 declaration.
11669
11670 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11671 to split multi-word logical operations.
11672 (rs6000_split_logical_di): Likewise.
11673 (rs6000_split_logical): Likewise.
11674
11675 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11676 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11677 and allow TImode operations in 32-bit.
11678 (vsx_and<mode>3_64bit): Likewise.
11679 (vsx_ior<mode>3_32bit): Likewise.
11680 (vsx_ior<mode>3_64bit): Likewise.
11681 (vsx_xor<mode>3_32bit): Likewise.
11682 (vsx_xor<mode>3_64bit): Likewise.
11683 (vsx_one_cmpl<mode>2_32bit): Likewise.
11684 (vsx_one_cmpl<mode>2_64bit): Likewise.
11685 (vsx_nor<mode>3_32bit): Likewise.
11686 (vsx_nor<mode>3_64bit): Likewise.
11687 (vsx_andc<mode>3_32bit): Likewise.
11688 (vsx_andc<mode>3_64bit): Likewise.
11689 (vsx_eqv<mode>3_32bit): Likewise.
11690 (vsx_eqv<mode>3_64bit): Likewise.
11691 (vsx_nand<mode>3_32bit): Likewise.
11692 (vsx_nand<mode>3_64bit): Likewise.
11693 (vsx_orc<mode>3_32bit): Likewise.
11694 (vsx_orc<mode>3_64bit): Likewise.
11695
11696 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11697 logical types in GPRs.
11698
11699 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11700 logical insns to rs6000.md, and allow TImode operations in
11701 32-bit.
11702 (altivec_ior<mode>3): Likewise.
11703 (altivec_xor<mode>3): Likewise.
11704 (altivec_one_cmpl<mode>2): Likewise.
11705 (altivec_nor<mode>3): Likewise.
11706 (altivec_andc<mode>3): Likewise.
11707
11708 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11709 attributes for moving the 128-bit logical operations into
11710 rs6000.md.
11711 (BOOL_REGS_OUTPUT): Likewise.
11712 (BOOL_REGS_OP1): Likewise.
11713 (BOOL_REGS_OP2): Likewise.
11714 (BOOL_REGS_UNARY): Likewise.
11715 (BOOL_REGS_AND_CR0): Likewise.
11716 (one_cmpl<mode>2): Add support for DI logical operations on
11717 32-bit, splitting the operations to 32-bit.
11718 (anddi3): Likewise.
11719 (iordi3): Likewise.
11720 (xordi3): Likewise.
11721 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11722 changes to combine the 32/64-bit code, allow logical operations on
11723 TI mode in 32-bit, and to use similar match_operator patterns like
11724 scalar mode uses. Combine the Altivec and VSX code for logical
11725 operations, and move it here.
11726 (ior<mode>3, 128-bit types): Likewise.
11727 (xor<mode>3, 128-bit types): Likewise.
11728 (one_cmpl<mode>3, 128-bit types): Likewise.
11729 (nor<mode>3, 128-bit types): Likewise.
11730 (andc<mode>3, 128-bit types): Likewise.
11731 (eqv<mode>3, 128-bit types): Likewise.
11732 (nand<mode>3, 128-bit types): Likewise.
11733 (orc<mode>3, 128-bit types): Likewise.
11734 (and<mode>3_internal): Likewise.
11735 (bool<mode>3_internal): Likewise.
11736 (boolc<mode>3_internal1): Likewise.
11737 (boolc<mode>3_internal2): Likewise.
11738 (boolcc<mode>3_internal1): Likewise.
11739 (boolcc<mode>3_internal2): Likewise.
11740 (eqv<mode>3_internal1): Likewise.
11741 (eqv<mode>3_internal2): Likewise.
11742 (one_cmpl1<mode>3_internal): Likewise.
11743
12211b99 117442013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 11745
11746 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11747 Rename flag_stack_usage to flag_stack_usage_info.
11748
12211b99 117492013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 11750
a5f257fd
ME		 11751 * config/microblaze/sync.md: New file.
11752 * config/microblaze/microblaze.md: Include sync.md
11753 * config/microblaze/microblaze.c: Add print_operand 'y'.
11754 * config/microblaze/constraints.md: Add memory_contraint
11755 'Q' which is a single register.
450b0ebb 11756
49fba14f
EB		 117572013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11758
11759 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11760
c7b3b99f
PCC		 117612013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11762
11763 * reload.c (find_reloads): Exit loop once we find this operand
11764 cannot be reloaded somehow for this alternative.
11765
11766 * reload.c (find_reloads): Exit loop once we find a hard register.
11767
11768 * rtlanal.c (computed_jump_p): Exit loop once we find label
11769 reference is used.
11770
11771 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11772
11773 * cfgloopmanip.c (remove_path): Exit loop after setting
11774 irred_invalidated.
11775
11776 * gensupport.c (subst_dup): Avoid loop if code is not
11777 MATCH_DUP nor MATCH_OP_DUP.
11778
c8fbf1fa
NBJ		 117792013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11780
11781 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11782
9259db42
YZ		 117832013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11784
11785 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11786 true for SP_REGNUM if mode == ptr_mode.
11787 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11788 with value R0_REGNUM + 31.
11789
00edcfbe
YZ		 117902013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11791
11792 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11793 pad pointer-typed argument downward.
11794
43be9a95
YZ		 117952013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11796
11797 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11798 and __ILP32__ when the ILP32 model is in use.
11799
28514dda
YZ		 118002013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11801
11802 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11803 (aarch64_load_symref_appropriately): In the case of
11804 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11805 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11806 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11807 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11808 if 'mode' doesn't equal to 'ptr_mode'.
11809 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11810 'vcall_offset'; change to call aarch64_emit_move differently depending
11811 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11812 to calculate the upper bound of 'vcall_offset'.
11813 (aarch64_cannot_force_const_mem): Change to also return true if
11814 mode != ptr_mode.
11815 (aarch64_legitimize_reload_address): In the case of large
11816 displacements, add new local variable 'xmode' and an assertion
11817 based on it; change to use 'xmode' to generate the new rtx and
11818 reload.
11819 (aarch64_asm_trampoline_template): Change to generate the template
11820 differently depending on TARGET_ILP32 or not; change to use
11821 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11822 (aarch64_trampoline_size): Removed.
11823 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11824 and replace immediate literals with it. Change to use 'ptr_mode'
11825 instead of 'DImode' and call convert_memory_address if the mode
11826 of 'fnaddr' doesn't equal to 'ptr_mode'.
11827 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11828 to output symbol.
11829 (aarch64_elf_asm_destructor): Likewise.
11830 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11831 on TARGET_ILP32 instead of aarch64_trampoline_size.
11832 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11833 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11834 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11835 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11836 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11837 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11838 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11839 depending on the value of 'mode'.
11840 (add_losym_<mode>): New.
11841 (ldr_got_small_<mode>): New, based on ldr_got_small.
11842 (ldr_got_small): Remove.
11843 (ldr_got_small_sidi): New.
11844 * config/aarch64/iterators.md (P): New.
11845 (PTR): Change to 'ptr_mode' in the condition.
11846
17a819cb
YZ		 118472013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11848
11849 * config.gcc (aarch64*-*-*): Support --with-abi.
11850 (aarch64*-*-elf): Support --with-multilib-list.
11851 (aarch64*-*-linux*): Likewise.
11852 (supported_defaults): Add abi to aarch64*-*-*.
11853 * configure.ac: Mention AArch64 for --with-multilib-list.
11854 * configure: Re-generated.
11855 * config/aarch64/biarchilp32.h: New file.
11856 * config/aarch64/biarchlp64.h: New file.
11857 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11858 (ABI_SPEC): Ditto.
11859 (MULTILIB_DEFAULTS): Ditto.
11860 (DRIVER_SELF_SPECS): Ditto.
11861 (ASM_SPEC): Update to also substitute -mabi.
11862 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11863 file whose name depends on -mabi= and -mbig-endian.
11864 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11865 TARGET_ILP32.
11866 (POINTER_SIZE): New define.
11867 (POINTERS_EXTEND_UNSIGNED): Ditto.
11868 (enum aarch64_abi_type): New enumeration tag.
11869 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11870 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11871 (TARGET_ILP32): New define.
11872 * config/aarch64/aarch64.opt (mabi): New.
11873 (aarch64_abi): New.
11874 (ilp32, lp64): New values for -mabi.
11875 * config/aarch64/t-aarch64 (comma): New define.
11876 (MULTILIB_OPTIONS): Ditto.
11877 (MULTILIB_DIRNAMES): Ditto.
11878 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11879 * doc/invoke.texi: Document -mabi for AArch64.
11880
9f9cbdce
GJL		 118812013-07-23 Georg-Johann Lay <avr@gjlay.de>
11882
11883 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11884
7061977b 118852013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11886 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 11887
11888 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11889 endianness when selecting field to splat.
11890
75a41878
EC		 118912013-07-22 Eric Christopher <echristo@gmail.com>
11892
11893 * dwarf2out.c (die_odr_checksum): New function to use
11894 CHECKSUM_ macros and ULEB128 for DIE tag.
11895 (generate_type_signature): Use.
41a7c215 11896
38ae58ca
EB		 118972013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11898
11899 * config.gcc (sparc*-*-*): Accept leon3 processor.
11900 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11901 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11902 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11903 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11904 (mfix-ut699): Adjust comment.
11905 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11906 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11907 (CPP_CPU_SPEC): Likewise.
11908 (ASM_CPU_SPEC): Likewise.
11909 * config/sparc/sparc.c (leon3_cost): New constant.
11910 (sparc_option_override): Add leon3 support.
11911 (mem_ref): New function.
11912 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11913 (sparc_do_work_around_errata): Look into the instruction in the delay
11914 slot and adjust accordingly. Add fix for the data cache nullify issues
11915 of the UT699. Change insertion position for the NOP.
11916 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11917 (leon3_load): New reservation.
11918 (leon_store): Bump latency to 2.
11919 (grfpu): New automaton.
11920 (grfpu_alu): New unit.
11921 (grfpu_ds): Likewise.
11922 (leon_fp_alu): Adjust.
11923 (leon_fp_mult): Delete.
11924 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11925 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11926 * config/sparc/sparc.md (cpu): Add leon3.
11927 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11928 (swapsi): Likewise.
11929 (atomic_test_and_set): Likewise.
11930 (ldstub): Likewise.
11931
04dfc6df
JU		 119322013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11933
11934 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11935 default for R5900 targets.
11936 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11937 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11938 * config/mips/mips.c (mips_option_override): Report an error for
11939 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11940 for -march=r5900 -mhard-float.
11941
c7b3b99f 119422013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 11943
11944 * df-problems.c (can_move_insns_across): Exit loop once we
11945 find a non-fixed, non-global register.
11946
11947 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11948 setting can_throw.
11949
11950 * omega.c (omega_eliminate_red): Break after setting red_found.
11951 (omega_problem_has_red_equations): Similarly after setting found.
11952 (omega_query_variable): Similarly after setting coupled.
11953
9b6e6981
MP		 119542013-07-22 Marek Polacek <polacek@redhat.com>
11955
11956 * gimplify.c: Don't include gimple.h twice.
11957
5e5f7673
KT		 119582013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11959
11960 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11961 instead of TARGET_THUMB1.
11962 (Pz): New constraint.
11963 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11964 encodings.
11965 (compare_negsi_si): Likewise.
11966 (compare_addsi2_op0): Likewise.
11967 (compare_addsi2_op1): Likewise.
11968 (addsi3_carryin_<optab>): Likewise.
11969 (addsi3_carryin_alt2_<optab>): Likewise.
11970 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11971 for arm_restrict_it.
11972 (subsi3_carryin): Likewise.
11973 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11974 (minmax_arithsi): Disable for arm_restrict_it.
11975 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11976 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11977 (satsi_<SAT:code>_shift): Likewise.
11978 (arm_shiftsi3): Add alternative for 16-bit encoding.
11979 (arm32_movhf): Disable for arm_restrict_it.
11980 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11981 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11982
859abddd
SN		 119832013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11984
11985 * config/arm/arm.md (attribute "insn"): Delete.
11986 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11987 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11988 (not_shiftsi): Update for attribute change.
11989 (not_shiftsi_compare0): Likewise.
11990 (not_shiftsi_compare0_scratch): Likewise.
11991 (arm_one_cmplsi2): Likewise.
11992 (thumb1_one_cmplsi2): Likewise.
11993 (notsi_compare0): Likewise.
11994 (notsi_compare0_scratch): Likewise.
11995 (thumb1_movdi_insn): Likewise.
11996 (arm_movsi_insn): Likewise.
11997 (movhi_insn_arch4): Likewise.
11998 (movhi_bytes): Likewise.
11999 (arm_movqi_insn): Likewise.
12000 (thumb1_movqi_insn): Likewise.
12001 (arm32_movhf): Likewise.
12002 (thumb1_movhf): Likewise.
12003 (arm_movsf_soft_insn): Likewise.
12004 (thumb1_movsf_insn): Likewise.
12005 (thumb_movdf_insn): Likewise.
12006 (movsicc_insn): Likewise.
12007 (movsfcc_soft_insn): Likewise.
12008 (and_scc): Likewise.
12009 (cond_move): Likewise.
12010 (if_move_not): Likewise.
12011 (if_not_move): Likewise.
12012 (if_shift_move): Likewise.
12013 (if_move_shift): Likewise.
12014 (if_shift_shift): Likewise.
12015 (if_not_arith): Likewise.
12016 (if_arith_not): Likewise.
12017 (cond_move_not): Likewise.
12018 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12019 (neon_mov<mode>): Likewise.
12020 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12021 (thumb2_movsi_vfp): Likewise.
12022 (movsf_vfp): Likewise.
12023 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 12024 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12025 change.
859abddd
SN		 12026 (cortexa7_older_only): Likewise.
12027 (cortexa7_younger): Likewise.
12028 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12029 (1020alu_shift_op): Likewise.
12030 (1020alu_shift_reg_op): Likewise.
12031 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12032 (alu_shift_op): Likewise.
12033 (alu_shift_reg_op): Likewise.
12034 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12035 (11_alu_shift_op): Likewise.
12036 (11_alu_shift_reg_op): Likewise.
12037 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12038 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12039 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12040 change.
859abddd
SN		 12041 (cortex_a15_alu_shift): Likewise.
12042 (cortex_a15_alu_shift_reg): Likewise.
12043 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12044 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 12045 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12046 change.
859abddd 12047 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 12048 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12049 change.
859abddd
SN		 12050 (cortex_a7_alu_reg): Likewise.
12051 (cortex_a7_alu_shift): Likewise.
12052 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12053 (cortex_a8_alu_shift): Likewise.
12054 (cortex_a8_alu_shift_reg): Likewise.
12055 (cortex_a8_mov): Likewise.
12056 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12057 (cortex_a9_dp_shift): Likewise.
12058 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12059 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12060 (cortex_r4_mov): Likewise.
12061 (cortex_r4_alu_shift): Likewise.
12062 (cortex_r4_alu_shift_reg): Likewise.
12063 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12064 (526_alu_shift_op): Likewise.
12065 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12066 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12067 (626te_alu_shift_op): Likewise.
12068 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12069 (726te_alu_op): Likewise.
12070 (726te_alu_shift_op): Likewise.
12071 (726te_alu_shift_reg_op): Likewise.
12072 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12073 (mp626_alu_shift_op): Likewise.
12074 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12075 (pj4_alu_e1_conds): Likewise.
12076 (pj4_alu): Likewise.
12077 (pj4_alu_conds): Likewise.
12078 (pj4_shift): Likewise.
12079 (pj4_shift_conds): Likewise.
12080 (pj4_alu_shift): Likewise.
12081 (pj4_alu_shift_conds): Likewise.
12082
95b97fac
KT		 120832013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12084
12085 * config/arm/predicates.md (shiftable_operator_strict_it):
12086 New predicate.
12087 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12088 Disable cond_exec version for arm_restrict_it.
12089 (thumb2_smaxsi3): Convert to generate cond_exec.
12090 (thumb2_sminsi3): Likewise.
12091 (thumb32_umaxsi3): Likewise.
12092 (thumb2_uminsi3): Likewise.
12093 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12094 (thumb2_neg_abssi2): Likewise.
12095 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12096 (thumb2_movsicc_insn): Adjust alternatives.
12097 (thumb2_mov_negscc): Disable for arm_restrict_it.
12098 (thumb2_mov_negscc_strict_it): New pattern.
12099 (thumb2_mov_notscc_strict_it): New pattern.
12100 (thumb2_mov_notscc): Disable for arm_restrict_it.
12101 (thumb2_ior_scc): Likewise.
12102 (thumb2_ior_scc_strict_it): New pattern.
12103 (thumb2_cond_move): Adjust for arm_restrict_it.
12104 (thumb2_cond_arith): Disable for arm_restrict_it.
12105 (thumb2_cond_arith_strict_it): New pattern.
12106 (thumb2_cond_sub): Adjust for arm_restrict_it.
12107 (thumb2_movcond): Likewise.
12108 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12109 (thumb2_zero_extendhisi2_v6): Likewise.
12110 (thumb2_zero_extendqisi2_v6): Likewise.
12111 (orsi_notsi_si): Likewise.
12112 (orsi_not_shiftsi_si): Likewise.
12113
d5a6ef82
GJL		 121142013-07-22 Georg-Johann Lay <avr@gjlay.de>
12115
12116 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12117 instruction sequence is 1 byte shorter.
12118
d9e71154
UB		 121192013-07-22 Uros Bizjak <ubizjak@gmail.com>
12120
12121 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12122 it is not needed after split.
12123
8c6381ec
IS		 121242013-07-20 Iain Sandoe <iain@codesourcery.com>
12125
12126 PR target/51784
12127 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12128 second label for nonlocal goto receivers. Don't output pic base labels
12129 unless we're producing PIC; mark that action unreachable().
12130 (ix86_save_reg): If the function contains a nonlocal label, save the
12131 PIC base reg.
12132 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 12133 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 12134 (update_pic_label_number_if_needed): New.
12135 (machopic_output_function_base_name): Adjust for nonlocal receiver
12136 case.
12137 (machopic_should_output_picbase_label): New.
12138 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12139 (nonlocal_goto_receiver): New insn and split.
12140
d05d0709
JG		 121412013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12142
12143 * config/aarch64/aarch64-builtins.c
12144 (aarch64_fold_builtin): Fold abs in all modes.
12145 * config/aarch64/aarch64-simd-builtins.def
12146 (abs): Enable for all modes.
12147 * config/aarch64/arm_neon.h
12148 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12149 (vabs_f64): Add missing intrinsic.
12150
ba590e63
IB		 121512013-07-19 Ian Bolton <ian.bolton@arm.com>
12152
12153 * config/aarch64/arm_neon.h (vabs_s64): New function
12154
02371798
GJL		 121552013-07-19 Georg-Johann Lay <avr@gjlay.de>
12156
12157 PR target/57516
12158 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12159 * config/avr/avr.md (adjust_len): Add `round'.
12160 * config/avr/avr-protos.h (avr_out_round): New prototype.
12161 (avr_out_plus): Add `out_label' argument.
12162 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12163 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12164 Handle the case where `insn' is just a pattern.
12165 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12166 (avr_out_round): New function.
12167 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12168
12211b99 121692013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 12170
12171 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12172 Add check for flag_stack_usage to handle -fstack-usage support
12173
12211b99 121742013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 12175
12176 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12177 interaction for new Power8 flags and VSX.
12178
df9dda2d
ST		 121792013-07-18 Sriraman Tallam <tmsriram@google.com>
12180
12181 PR middle-end/57698
12182 * tree-inline.c (expand_call_inline): Emit errors during
12183 early_inlining only if optimization is not turned on.
12184
0f9cdef4
DM		 121852013-07-18 David Malcolm <dmalcolm@redhat.com>
12186
12187 * passes.def: New.
12188
12189 * passes.c (init_optimization_passes): Move the construction of
12190 the pass hierarchy into a new passes.def file.
12191
12192 * Makefile.in (passes.o): Add dependency on passes.def.
12193
2efa4087
DM		 121942013-07-18 David Malcolm <dmalcolm@redhat.com>
12195
12196 * passes.c (init_optimization_passes): Introduce macros for
12197 constructing the tree of passes (INSERT_PASSES_AFTER,
12198 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12199 TERMINATE_PASS_LIST).
12200
bc404e1b
VM		 122012013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12202 Wei Mi <wmi@google.com>
12203
12204 PR rtl-optimization/57878
12205 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12206 top.
12207 (reload_pseudo_compare_func): Check nregs first for reload
12208 pseudos.
12209
d675843f
DM		 122102013-07-18 David Malcolm <dmalcolm@redhat.com>
12211
12212 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12213
c7b3b99f 122142013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 12215
12216 * read-rtl.c (validate_const_int): Once an invalid character is
12217 seen, quit the loop.
12218
12219 * gengtype.c (write_roots): Similarly once we find the "deletable"
12220 or "if_marked" option.
12221
52fceb44
SN		 122222013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12223
12224 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12225 "xtab" and "sat". Move value "clz" from here to ...
12226 (attriubte "type"): ... here.
12227 (satsi_<SAT:code>): Delete "insn" attribute.
12228 (satsi_<SAT:code>_shift): Likewise.
12229 (arm_zero_extendqisi2addsi): Likewise.
12230 (arm_extendqisi2addsi): Likewise.
12231 (clzsi2): Update for attribute changes.
12232 (rbitsi2): Likewise.
371e77e3
UB		 12233 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12234 attribute.
52fceb44
SN		 12235 (arm_usatsihi): Likewise.
12236 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12237
006bd006
SN		 122382013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12239
12240 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 12241 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12242 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 12243 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12244 in alphabetical order.
12245 (attribute "core_cycles"): Update for attribute changes.
12246 (arm_addsi3): Likewise.
12247 (addsi3_compare0): Likewise.
12248 (addsi3_compare0_scratch): Likewise.
12249 (addsi3_compare_op1): Likewise.
12250 (addsi3_compare_op2): Likewise.
12251 (compare_addsi2_op0): Likewise.
12252 (compare_addsi2_op1): Likewise.
12253 (addsi3_carryin_shift_<optab>): Likewise.
12254 (subsi3_carryin_shift): Likewise.
12255 (rsbsi3_carryin_shift): Likewise.
12256 (arm_subsi3_insn): Likewise.
12257 (subsi3_compare0): Likewise.
12258 (subsi3_compare): Likewise.
12259 (arm_andsi3_insn): Likewise.
12260 (thumb1_andsi3_insn): Likewise.
12261 (andsi3_compare0): Likewise.
12262 (andsi3_compare0_scratch): Likewise.
12263 (zeroextractsi_compare0_scratch
12264 (andsi_not_shiftsi_si): Likewise.
12265 (iorsi3_insn): Likewise.
12266 (iorsi3_compare0): Likewise.
12267 (iorsi3_compare0_scratch): Likewise.
12268 (arm_xorsi3): Likewise.
12269 (thumb1_xorsi3_insn): Likewise.
12270 (xorsi3_compare0): Likewise.
12271 (xorsi3_compare0_scratch): Likewise.
12272 (satsi_<SAT:code>_shift): Likewise.
12273 (rrx): Likewise.
12274 (arm_shiftsi3): Likewise.
12275 (shiftsi3_compare0): Likewise.
12276 (not_shiftsi): Likewise.
12277 (not_shiftsi_compare0): Likewise.
12278 (not_shiftsi_compare0_scratch): Likewise.
12279 (arm_one_cmplsi2): Likewise.
12280 (thumb_one_complsi2): Likewise.
12281 (notsi_compare0): Likewise.
12282 (notsi_compare0_scratch): Likewise.
12283 (thumb1_zero_extendhisi2): Likewise.
12284 (arm_zero_extendhisi2): Likewise.
12285 (arm_zero_extendhisi2_v6): Likewise.
12286 (arm_zero_extendhisi2addsi): Likewise.
12287 (thumb1_zero_extendqisi2): Likewise.
12288 (thumb1_zero_extendqisi2_v6): Likewise.
12289 (arm_zero_extendqisi2): Likewise.
12290 (arm_zero_extendqisi2_v6): Likewise.
12291 (arm_zero_extendqisi2addsi): Likewise.
12292 (thumb1_extendhisi2): Likewise.
12293 (arm_extendhisi2): Likewise.
12294 (arm_extendhisi2_v6): Likewise.
12295 (arm_extendqisi): Likewise.
12296 (arm_extendqisi_v6): Likewise.
12297 (arm_extendqisi2addsi): Likewise.
12298 (thumb1_extendqisi2): Likewise.
12299 (thumb1_movdi_insn): Likewise.
12300 (arm_movsi_insn): Likewise.
12301 (movsi_compare0): Likewise.
12302 (movhi_insn_arch4): Likewise.
12303 (movhi_bytes): Likewise.
12304 (arm_movqi_insn): Likewise.
12305 (thumb1_movqi_insn): Likewise.
12306 (arm32_movhf): Likewise.
12307 (thumb1_movhf): Likewise.
12308 (arm_movsf_soft_insn): Likewise.
12309 (thumb1_movsf_insn): Likewise.
12310 (movdf_soft_insn): Likewise.
12311 (thumb_movdf_insn): Likewise.
12312 (arm_cmpsi_insn): Likewise.
12313 (cmpsi_shiftsi): Likewise.
12314 (cmpsi_shiftsi_swp): Likewise.
12315 (arm_cmpsi_negshiftsi_si): Likewise.
12316 (movsicc_insn): Likewise.
12317 (movsfcc_soft_insn): Likewise.
12318 (arith_shiftsi): Likewise.
12319 (arith_shiftsi_compare0
12320 (arith_shiftsi_compare0_scratch
12321 (sub_shiftsi): Likewise.
12322 (sub_shiftsi_compare0
12323 (sub_shiftsi_compare0_scratch
12324 (and_scc): Likewise.
12325 (cond_move): Likewise.
12326 (if_plus_move): Likewise.
12327 (if_move_plus): Likewise.
12328 (if_move_not): Likewise.
12329 (if_not_move): Likewise.
12330 (if_shift_move): Likewise.
12331 (if_move_shift): Likewise.
12332 (if_shift_shift): Likewise.
12333 (if_not_arith): Likewise.
12334 (if_arith_not): Likewise.
12335 (cond_move_not): Likewise.
12336 (thumb1_ashlsi3): Set type attribute.
12337 (thumb1_ashrsi3): Likewise.
12338 (thumb1_lshrsi3): Likewise.
12339 (thumb1_rotrsi3): Likewise.
12340 (shiftsi3_compare0_scratch): Likewise.
12341 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12342 (neon_mov<mode>): Likewise.
371e77e3
UB		 12343 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12344 attribute changes.
006bd006
SN		 12345 (thumb2_movsi_insn): Likewise.
12346 (thumb2_cmpsi_neg_shiftsi): Likewise.
12347 (thumb2_extendqisi_v6): Likewise.
12348 (thumb2_zero_extendhisi2_v6): Likewise.
12349 (thumb2_zero_extendqisi2_v6): Likewise.
12350 (thumb2_shiftsi3_short): Likewise.
12351 (thumb2_addsi3_compare0_scratch): Likewise.
12352 (orsi_not_shiftsi_si): Likewise.
12353 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12354 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12355 changes.
12356 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12357 (1020alu_shift_op): Likewise.
12358 (1020alu_shift_reg_op): Likewise.
12359 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12360 (alu_shift_op): Likewise.
12361 (alu_shift_reg_op): Likewise.
12362 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12363 (11_alu_shift_op): Likewise.
12364 (11_alu_shift_reg_op): Likewise.
12365 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12366 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12367 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12368 changes.
006bd006
SN		 12369 (cortex_a15_alu_shift): Likewise.
12370 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 12371 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12372 changes.
006bd006
SN		 12373 (cortex_a5_alu_shift): Likewise.
12374 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12375 changes.
12376 (cortex_a53_alu_shift): Likewise.
12377 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12378 changes.
12379 (cortex_a7_alu_reg): Likewise.
12380 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 12381 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12382 changes.
006bd006
SN		 12383 (cortex_a8_alu_shift): Likewise.
12384 (cortex_a8_alu_shift_reg): Likewise.
12385 (cortex_a8_mov): Likewise.
12386 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12387 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 12388 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12389 changes.
12390 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12391 changes.
006bd006
SN		 12392 (cortex_r4_mov): Likewise.
12393 (cortex_r4_alu_shift): Likewise.
12394 (cortex_r4_alu_shift_reg): Likewise.
12395 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12396 (526_alu_shift_op): Likewise.
12397 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12398 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12399 (626te_alu_shift_op): Likewise.
12400 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12401 (726te_alu_op): Likewise.
12402 (726te_alu_shift_op): Likewise.
12403 (726te_alu_shift_reg_op): Likewise.
12404 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12405 (mp626_alu_shift_op): Likewise.
12406 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12407 (pj4_alu_e1_conds): Likewise.
12408 (pj4_alu): Likewise.
12409 (pj4_alu_conds): Likewise.
12410 (pj4_shift): Likewise.
12411 (pj4_shift_conds): Likewise.
12412 (pj4_alu_shift): Likewise.
12413 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 12414 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12415 changes.
006bd006
SN		 12416 (cortexa7_older_only): Likewise.
12417 (cortexa7_younger): Likewise.
12418
651df1b2
DM		 124192013-07-18 David Malcolm <dmalcolm@redhat.com>
12420
12421 * ipa-pure-const.c (generate_summary): Rename to...
12422 (pure_const_generate_summary): ... this.
12423
ea717bd7
IS		 124242013-07-17 Iain Sandoe <iain@codesourcery.com>
12425
12426 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12427
a1259a13
YR		 124282013-07-17 Yvan Roux <yvan.roux@linaro.org>
12429
12430 PR target/57909
12431 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12432 usage in HI mode.
12433
167f68ed
AK		 124342013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12435
12436 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12437 enabled without -march=zEC12.
12438 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12439 flags to be set.
12440
f900a982
MR		 124412013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12442
12443 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12444 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12445 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12446 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12447 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12448 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12449 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12450 ISA_HAS_FP4.
12451 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12452 and ISA_HAS_NMADD3_NMSUB3.
12453 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12454 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12455 (nmsub4<mode>, nmsub3<mode>): Likewise.
12456 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12457
d4ed27eb
MR		 124582013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12459
12460 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12461 TARGET_MIPS5400 checking.
12462
7a5add18
PB		 124632013-07-16 Jakub Jelinek <jakub@redhat.com>
12464 Peter Bergner <bergner@vnet.ibm.com>
12465
12466 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12467 registers in the comment.
12468 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12469 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12470 rather than FIRST_PSEUDO_REGISTERS.
12471
124722013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 12473
12474 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12475 enable extra ISA flags with TARGET_HTM.
12476
d2ab0929
MR		 124772013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12478
12479 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12480 Fix comment typos.
12481
839c74bc
CH		 124822013-07-15 Cong Hou <congh@google.com>
12483
12484 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12485 in compare function for sorting.
12486
7a5add18 124872013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 12488
12489 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12490 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12491 * config/rs6000/rs6000.opt: Add -mhtm option.
12492 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12493 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12494 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12495 __HTM__ if the HTM instructions are available.
12496 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12497 htm_spr_reg_operand): New define_predicates.
12498 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12499 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12500 Include htm.md.
12501 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12502 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12503 HTM builtin functions.
12504 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12505 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12506 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12507 (rs6000_builtin_mask_calculate): Likewise.
12508 (rs6000_option_override_internal): Likewise.
12509 (bdesc_htm): Add new HTM builtin support.
12510 (htm_spr_num): New function.
12511 (htm_spr_regno): Likewise.
12512 (rs6000_htm_spr_icode): Likewise.
12513 (htm_expand_builtin): Likewise.
12514 (htm_init_builtins): Likewise.
12515 (rs6000_expand_builtin): Add support for HTM builtin functions.
12516 (rs6000_init_builtins): Likewise.
371e77e3
UB		 12517 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12518 option.
0258b6e4
PB		 12519 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12520 (TARGET_HTM, MASK_HTM): Define macros.
12521 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12522 (FIXED_REGISTERS): Likewise.
12523 (CALL_USED_REGISTERS): Likewise.
12524 (CALL_REALLY_USED_REGISTERS): Likewise.
12525 (REG_ALLOC_ORDER): Likewise.
12526 (enum reg_class): Likewise.
12527 (REG_CLASS_NAMES): Likewise.
12528 (REG_CLASS_CONTENTS): Likewise.
12529 (REGISTER_NAMES): Likewise.
12530 (ADDITIONAL_REGISTER_NAMES): Likewise.
12531 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12532 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12533 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12534 * config/rs6000/htm.md: New file.
12535 * config/rs6000/htmintrin.h: New file.
12536 * config/rs6000/htmxlintrin.h: New file.
12537
87dd8ab0
MS		 125382013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12539
12540 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12541 Define SYMBOL_TINY_GOT, update comment.
12542 * config/aarch64/aarch64.c
12543 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12544 (aarch64_expand_mov_immediate): Likewise.
12545 (aarch64_print_operand): Likewise.
12546 (aarch64_classify_symbol): Likewise.
12547 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12548 (ldr_got_tiny): Define.
12549
52d676b6
TG		 125502013-07-13 Tobias Grosser <tobias@grosser.es>
12551
12552 PR tree-optimization/54094
12553 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12554 scheduling dimension for the parallelism check from the polyhedral
12555 information in the AST.
12556 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12557 in 2D + 1 form.
12558
286e8fc1
JM		 125592013-07-13 Jason Merrill <jason@redhat.com>
12560
12561 * print-tree.c (debug_vec_tree): Use debug_raw.
12562 (debug_raw (vec<tree, va_gc> &)): New.
12563 (debug_raw (vec<tree, va_gc> *)): New.
12564 * tree.h: Declare them.
12565
e43257e8
BC		 125662013-07-13 Bin Cheng <bin.cheng@arm.com>
12567
12568 * ifcvt.c (ifcvt_after_combine): New static variable.
12569 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12570 for size.
12571 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12572 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12573 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12574
da582d46
MR		 125752013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12576
12577 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12578
18b0ea8f
MM		 125792013-07-12 Michael Matz <matz@suse.de>
12580
12581 PR middle-end/55771
12582 * convert.c (convert_to_real): Reject non-float inner types.
12583
e4f0f84d
TB		 125842013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12585
12586 * config/aarch64/aarch64-protos.h
12587 (aarch64_simd_immediate_valid_for_move): Remove.
12588 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12589 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12590 cases.
12591 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12592
76500022
SE		 125932013-07-11 Steve Ellcey <sellcey@mips.com>
12594
12595 * config/mips/mips.c (mips_conditional_register_usage): Do not
12596 use t[0-7] registers in MIPS16 mode when optimizing for size.
12597
78039734
ST		 125982013-07-11 Sriraman Tallam <tmsriram@google.com>
12599
12600 * config/i386/i386.c (dispatch_function_versions): Fix array
12601 indexing of function_version_info to match actual_versions.
12602
7b24b675
TJ		 126032013-07-11 Teresa Johnson <tejohnson@google.com>
12604
12605 * vec.h (struct va_gc): Move release out-of-line.
12606 (va_gc::release): Call ggc_free on released vec.
12607
85d4cbb8
UW		 126082013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12609
12610 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12611 Require GOT register as additional operand in UNSPEC.
12612 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12613 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12614 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12615 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12616 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12617 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12618 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12619
52b3f9e2
GJL		 126202013-07-11 Georg-Johann Lay <avr@gjlay.de>
12621
12622 PR target/57631
12623 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12624 name seen by assembler/linker rather if available.
12625
3f3e266e
AS		 126262013-07-11 Andreas Schwab <schwab@suse.de>
12627
12628 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12629
21231ca6
VM		 126302013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12631
371e77e3 12632 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12633
d331d374
JM		 126342013-07-10 Joseph Myers <joseph@codesourcery.com>
12635
673c2f63
JM		 12636 * doc/tm.texi.in: Move hook documentation to ....
12637 * target.def: ... here.
12638
d331d374
JM		 12639 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12640 text on @hook line.
12641 * doc/tm.texi: Regenerate.
12642
dfeadaa0
PC		 126432013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12644
12645 PR c++/57869
12646 * doc/invoke.texi: Document -Wconditionally-supported.
12647
bb6f2bac
GJL		 126482013-07-10 Georg-Johann Lay <avr@gjlay.de>
12649
12650 PR target/57844
12651 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12652 of my_fp.
12653
5219b232
GJL		 126542013-07-10 Georg-Johann Lay <avr@gjlay.de>
12655
12656 PR target/57506
12657 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12658 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12659 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12660 Remove duplicate devices.
12661 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12662 * config/avr/t-multilib: Regenerate.
12663 * config/avr/avr-tables.opt: Regenerate.
12664 * doc/avr-mmcu.texi: Regenerate.
12665
283045c0
GJL		 126662013-07-10 Georg-Johann Lay <avr@gjlay.de>
12667
12668 PR target/56987
12669 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12670
ceb2d59f 126712013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 12672
ceb2d59f
GS		 12673 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12674 the cost of MULT when optimizing for size.
12675
055e0a99
JBG		 126762013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12677
12678 * config/cr16/cr16-protos.h: Don't include target.h.
12679
ef43e868
JM		 126802013-07-09 Joseph Myers <joseph@codesourcery.com>
12681
12682 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12683 adjust register size for TDmode and TFmode for VSX registers.
12684
4a283fdf
KT		 126852013-07-08 Kai Tietz <ktietz@redhat.com>
12686
12687 PR target/56892
12688 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12689 hook_bool_const_tree_true.
12690
2cf4c39e
AK		 126912013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12692
12693 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12694 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12695 * config/s390/s390.md: Define FPR*_REGNUM constants.
12696 Fix FPR2_REGNUM constant (18 -> 17).
12697 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12698 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12699 ("trunc<BFP:mode><DFP_ALL:mode>2")
12700 ("trunc<DFP_ALL:mode><BFP:mode>2")
12701 ("*extend<BFP:mode><DFP_ALL:mode>2")
12702 ("*extend<DFP_ALL:mode><BFP:mode>2")
12703 ("extend<BFP:mode><DFP_ALL:mode>2")
12704 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12705 FPR4_REGNUM.
12706
e6ac0270 127072013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 12708
6a2fa4b2 12709 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 12710
b89b22fc
AK		 127112013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12712
12713 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12714 and cfun_fpr_bit_p to cfun_fpr_save_p.
12715 (s390_frame_area, s390_register_info, s390_frame_info)
12716 (s390_emit_prologue, s390_emit_epilogue)
12717 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12718 register numbers.
12719 * config/s390/s390.h: Define *_REGNUM macros for floating point
12720 register numbers.
12721
07f398aa
EB		 127222013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12723
12724 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12725
75fefa91
PCC		 127262013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12727
12728 PR rtl-optimization/57786
12729 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12730 and break out of the loop when it is set to false.
12731
046f1eee
JJ		 127322013-07-08 Jakub Jelinek <jakub@redhat.com>
12733
8140c065
JJ		 12734 PR target/57819
12735 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12736 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12737 (const_int 63)) 0)).
12738 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12739 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12740 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12741
046f1eee
JJ		 12742 PR rtl-optimization/57829
12743 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12744 mask bits outside of mode are just sign-extension from mode to HWI.
12745
930b700b
MZ		 127462013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12747
12748 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12749 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12750 adjust_address instead of change_address to keep info about alignment.
12751 (emit_strmov): Remove.
12752 (emit_memmov): New function.
12753 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12754 (expand_movmem_epilogue): Likewise and return updated rtx for
12755 destination.
12756 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12757 destination and source.
12758 (decide_alignment): Refactor, handle vector_loop.
12759 (ix86_expand_movmem): Likewise.
12760 (ix86_expand_setmem): Likewise.
12761 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12762
7b9d1bd8
UB		 127632013-07-07 Uros Bizjak <ubizjak@gmail.com>
12764
12765 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12766 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12767
eabb5f48
UB		 127682013-07-06 Uros Bizjak <ubizjak@gmail.com>
12769
12770 * config/i386/sse.md (sse_movlhps): Change alternative 3
12771 of operand 2 to "m".
12772
127732013-07-06 Uros Bizjak <ubizjak@gmail.com>
12774
12775 PR target/57807
12776 * config/i386/sse.md (iptr): New mode attribute.
12777 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12778 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12779 (<sse>_vmmul<mode>3): Ditto.
12780 (<sse>_vmdiv<mode>3): Ditto.
12781 (sse_vmrcpv4sf2): Ditto.
12782 (<sse>_vmsqrt<mode>2): Ditto.
12783 (sse_vmrsqrtv4sf2): Ditto.
12784 (<sse>_vm<code><mode>3): Ditto.
12785 (avx_vmcmp<mode>3): Ditto.
12786 (<sse>_vmmaskcmp<mode>3): Ditto.
12787 (<sse>_comi): Ditto.
12788 (<sse>_ucomi): Ditto.
12789 (*xop_vmfrcz_<mode>): Ditto.
12790 (*fmai_fmadd_<mode>): Ditto.
12791 (*fmai_fmsub_<mode>): Ditto.
12792 (*fmai_fnmadd_<mode>): Ditto.
12793 (*fmai_fnmsub_<mode>): Ditto.
12794 (*fma4i_vmfmadd_<mode>): Ditto.
12795 (*fma4i_vmfmsub_<mode>): Ditto.
12796 (*fma4i_vmfnmadd_<mode>): Ditto.
12797 (*fma4i_vmfnmsub_<mode>): Ditto.
12798 (*xop_vmfrcz_<mode>): Ditto.
12799 (sse_cvtps2pi): Ditto.
12800 (sse_cvttps2pi): Ditto.
12801 (sse_cvtss2si): Ditto.
12802 (sse_cvtss2si_2): Ditto.
12803 (sse_cvtss2siq_2): Ditto.
12804 (sse_cvttss2si): Ditto.
12805 (sse_cvttss2siq): Ditto.
12806 (sse_cvtsd2si): Ditto.
12807 (sse_cvtsd2si_2): Ditto.
12808 (sse_cvtsd2siq_2): Ditto.
12809 (sse_cvttsd2si): Ditto.
12810 (sse_cvttsd2siq): Ditto.
12811 (sse_cvtsd2ss): Ditto.
12812 (sse_cvtss2sd): Ditto.
12813 (avx2_pbroadcast<mode>): Ditto.
12814 (avx2_pbroadcast<mode>_1): Ditto.
12815 (*avx_vperm_broadcast_v4sf): Ditto.
12816
12817 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12818 (sse_movlhps): Ditto.
12819 (sse_storehps): Ditto.
12820 (sse_loadhps): Ditto.
12821 (sse_storelps): Ditto.
12822 (sse_loadlps): Ditto.
12823 (*vec_concatv4sf): Ditto.
12824 (*vec_interleave_highv2df): Ditto.
12825 (*vec_interleave_lowv2df): Ditto.
12826 (*vec_extractv2df_1_sse): Ditto.
12827 (*vec_extractv2df_0_sse): Ditto.
12828 (sse2_storelpd): Ditto.
12829 (sse2_loadlpd): Ditto.
12830 (sse2_movsd): Ditto.
12831 (*vec_concatv4si): Ditto.
12832 (vec_concatv2di): Ditto.
12833
12834 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12835 for Intel asm dialect.
12836 (mmx_punpcklwd): Ditto.
12837 (mmx_punpckldq): Ditto.
12838
12839 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12840 for intel assembler dialect.
12841
1f6eac90
JJ		 128422013-07-06 Jakub Jelinek <jakub@redhat.com>
12843
12844 PR target/29776
12845 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12846 for BUILT_IN_C{LZ,LRSB}*.
12847 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12848 * tree-vrp.c (extract_range_basic): Handle
12849 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12850 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12851 fall thru to code calling set_value*.
12852 * builtins.c (expand_builtin): Remove *IMAX cases.
12853 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12854 if width is bigger than 2*HWI.
12855
2b778c9d
VM		 128562013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12857
12858 PR rtl-optimization/55342
12859 * lra-int.h (lra_subreg_reload_pseudos): New.
12860 * lra.c: Add undoing optional reloads to the block diagram.
12861 (lra_subreg_reload_pseudos): New.
12862 (lra_optional_reload_pseudos): Change comments.
12863 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12864 lra_optional_reload_pseudos after undo transformations.
12865 * lra-assigns.c (pseudo_prefix_title): New.
12866 (lra_setup_reg_renumber): Use it.
12867 (spill_for): Ditto. Check subreg reload pseudos too.
12868 (assign_by_spills): Consider subreg reload pseudos too.
12869 * lra-constraints.c (simplify_operand_subreg): Use
12870 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12871 (curr_insn_transform): Recognize and do optional reloads.
12872 (undo_optional_reloads): New.
12873 (lra_undo_inheritance): Call undo_optional_reloads.
12874
207156e4
TQ		 128752013-07-05 Thomas Quinot <quinot@adacore.com>
12876
12877 * tree-complex.c (expand_complex_operations_1): Fix typo.
12878
8990e73a
TB		 128792013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12880
12881 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12882 (tune_params): New member 'const vec_costs'.
12883 * config/aarch64/aarch64.c (generic_vector_cost): New.
12884 (generic_tunings): New member 'generic_vector_cost'.
12885 (aarch64_builtin_vectorization_cost): New.
12886 (aarch64_add_stmt_cost): New.
12887 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12888 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12889
f7bc421d
JJ		 128902013-07-03 Jakub Jelinek <jakub@redhat.com>
12891
12892 PR target/57777
12893 * config/i386/predicates.md (vsib_address_operand): Disallow
12894 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12895
e90d1568
HPN		 128962013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12897
12898 PR middle-end/55030
12899 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12900 expand_builtin_setjmp_receiver.
12901 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12902 with NULL for the label parameter.
12903 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12904 the frame-pointer. Adjust comments.
12905 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12906 only if LABEL is non-NULL.
12907
cbd7413b
YZ		 129082013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12909
12910 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12911 (ARM_ABI_AAPCS64): Ditto.
12912 (arm_abi): Ditto.
12913 (ARM_DEFAULT_ABI): Ditto.
12914
dec11868
JG		 129152013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12916
12917 * config/aarch64/aarch64-builtins.c
12918 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12919 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12920 (st1): Likewise.
12921 * config/aarch64/aarch64-simd.md
12922 (aarch64_ld1<VALL:mode>): New.
12923 (aarch64_st1<VALL:mode>): Likewise.
12924 * config/aarch64/arm_neon.h
12925 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12926
786fb9b6
ST		 129272013-07-02 Sriraman Tallam <tmsriram@google.com>
12928
0ba6aff3
UB		 12929 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12930 target ISA is AVX.
786fb9b6
ST		 12931 (ix86_option_override_internal):Turn on all -mavx target flags by
12932 default as they are dependent on AVX anyway.
12933
4b043553
CC		 129342013-07-02 Cary Coutant <ccoutant@google.com>
12935
12936 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12937 deterministic hash.
12938 (loc_checksum_ordered): Likewise.
12939 (hash_loc_operands): Remove inline keyword.
12940
c4e87a13
JJ		 129412013-07-02 Jakub Jelinek <jakub@redhat.com>
12942
12943 PR tree-optimization/57741
12944 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12945 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12946 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12947 Allow REAL_CST step_exprs if flag_associative_math.
12948 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12949
26366d28
IB		 129502013-07-02 Ian Bolton <ian.bolton@arm.com>
12951
0ba6aff3 12952 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 12953
2879bb2b
IB		 129542013-07-02 Ian Bolton <ian.bolton@arm.com>
12955
12956 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12957
afc5e8a6
KT		 129582013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12959
12960 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12961 encoding.
12962 (iorsi3_insn): Likewise.
12963 (arm_xorsi3): Likewise.
12964
31f8442b
SN		 129652013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12966
12967 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12968 to ...
12969 (attribute "type"): ... here, and prefix with "wmmx_".
12970 (attribute "core_cycles"): Update for attribute changes.
12971 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12972 (tbcstv4hi): Likewise.
12973 (tbcstv2si): Likewise.
12974 (iwmmxt_iordi3): Likewise.
12975 (iwmmxt_xordi3): Likewise.
12976 (iwmmxt_anddi3): Likewise.
12977 (iwmmxt_nanddi3): Likewise.
12978 (iwmmxt_arm_movdi): Likewise.
12979 (iwmmxt_movsi_insn): Likewise.
12980 (mov<mode>_internal): Likewise.
12981 (and<mode>3_iwmmxt): Likewise.
12982 (ior<mode>3_iwmmxt): Likewise.
12983 (xor<mode>3_iwmmxt): Likewise.
12984 (add<mode>3_iwmmxt): Likewise.
12985 (ssaddv8qi3): Likewise.
12986 (ssaddv4hi3): Likewise.
12987 (ssaddv2si3): Likewise.
12988 (usaddv8qi3): Likewise.
12989 (usaddv4hi3): Likewise.
12990 (usaddv2si3): Likewise.
12991 (sub<mode>3_iwmmxt): Likewise.
12992 (sssubv8qi3): Likewise.
12993 (sssubv4hi3): Likewise.
12994 (sssubv2si3): Likewise.
12995 (ussubv8qi3): Likewise.
12996 (ussubv4hi3): Likewise.
12997 (ussubv2si3): Likewise.
12998 (mulv4hi3_iwmmxt): Likewise.
12999 (smulv4hi3_highpart): Likewise.
13000 (umulv4hi3_highpart): Likewise.
13001 (iwmmxt_wmacs): Likewise.
13002 (iwmmxt_wmacsz): Likewise.
13003 (iwmmxt_wmacu): Likewise.
13004 (iwmmxt_wmacuz): Likewise.
13005 (iwmmxt_clrdi): Likewise.
13006 (iwmmxt_clrv8qi): Likewise.
13007 (iwmmxt_clr4hi): Likewise.
13008 (iwmmxt_clr2si): Likewise.
13009 (iwmmxt_uavgrndv8qi3): Likewise.
13010 (iwmmxt_uavgrndv4hi3): Likewise.
13011 (iwmmxt_uavgv8qi3): Likewise.
13012 (iwmmxt_uavgv4hi3): Likewise.
13013 (iwmmxt_tinsrb): Likewise.
13014 (iwmmxt_tinsrh): Likewise.
13015 (iwmmxt_tinsrw): Likewise.
13016 (iwmmxt_textrmub): Likewise.
13017 (iwmmxt_textrmsb): Likewise.
13018 (iwmmxt_textrmuh): Likewise.
13019 (iwmmxt_textrmsh): Likewise.
13020 (iwmmxt_textrmw): Likewise.
13021 (iwmxxt_wshufh): Likewise.
13022 (eqv8qi3): Likewise.
13023 (eqv4hi3): Likewise.
13024 (eqv2si3): Likewise.
13025 (gtuv8qi3): Likewise.
13026 (gtuv4hi3): Likewise.
13027 (gtuv2si3): Likewise.
13028 (gtv8qi3): Likewise.
13029 (gtv4hi3): Likewise.
13030 (gtv2si3): Likewise.
13031 (smax<mode>3_iwmmxt): Likewise.
13032 (umax<mode>3_iwmmxt): Likewise.
13033 (smin<mode>3_iwmmxt): Likewise.
13034 (umin<mode>3_iwmmxt): Likewise.
13035 (iwmmxt_wpackhss): Likewise.
13036 (iwmmxt_wpackwss): Likewise.
13037 (iwmmxt_wpackdss): Likewise.
13038 (iwmmxt_wpackhus): Likewise.
13039 (iwmmxt_wpackwus): Likewise.
13040 (iwmmxt_wpackdus): Likewise.
13041 (iwmmxt_wunpckihb): Likewise.
13042 (iwmmxt_wunpckihh): Likewise.
13043 (iwmmxt_wunpckihw): Likewise.
13044 (iwmmxt_wunpckilb): Likewise.
13045 (iwmmxt_wunpckilh): Likewise.
13046 (iwmmxt_wunpckilw): Likewise.
13047 (iwmmxt_wunpckehub): Likewise.
13048 (iwmmxt_wunpckehuh): Likewise.
13049 (iwmmxt_wunpckehuw): Likewise.
13050 (iwmmxt_wunpckehsb): Likewise.
13051 (iwmmxt_wunpckehsh): Likewise.
13052 (iwmmxt_wunpckehsw): Likewise.
13053 (iwmmxt_wunpckelub): Likewise.
13054 (iwmmxt_wunpckeluh): Likewise.
13055 (iwmmxt_wunpckeluw): Likewise.
13056 (iwmmxt_wunpckelsb): Likewise.
13057 (iwmmxt_wunpckelsh): Likewise.
13058 (iwmmxt_wunpckelsw): Likewise.
13059 (ror<mode>3): Likewise.
13060 (ashr<mode>3_iwmmxt): Likewise.
13061 (lshr<mode>3_iwmmxt): Likewise.
13062 (ashl<mode>3_iwmmxt): Likewise.
13063 (ror<mode>3_di): Likewise.
13064 (ashr<mode>3_di): Likewise.
13065 (lshr<mode>3_di): Likewise.
13066 (ashl<mode>3_di): Likewise.
13067 (iwmmxt_wmadds): Likewise.
13068 (iwmmxt_wmaddu): Likewise.
13069 (iwmmxt_tmia): Likewise.
13070 (iwmmxt_tmiaph): Likewise.
13071 (iwmmxt_tmiabb): Likewise.
13072 (iwmmxt_tmiatb): Likewise.
13073 (iwmmxt_tmiabt): Likewise.
13074 (iwmmxt_tmiatt): Likewise.
13075 (iwmmxt_tmovmskb): Likewise.
13076 (iwmmxt_tmovmskh): Likewise.
13077 (iwmmxt_tmovmskw): Likewise.
13078 (iwmmxt_waccb): Likewise.
13079 (iwmmxt_wacch): Likewise.
13080 (iwmmxt_waccw): Likewise.
13081 (iwmmxt_waligni): Likewise.
13082 (iwmmxt_walignr): Likewise.
13083 (iwmmxt_walignr0): Likewise.
13084 (iwmmxt_walignr1): Likewise.
13085 (iwmmxt_walignr2): Likewise.
13086 (iwmmxt_walignr3): Likewise.
13087 (iwmmxt_wsadb): Likewise.
13088 (iwmmxt_wsadh): Likewise.
13089 (iwmmxt_wsadbz): Likewise.
13090 (iwmmxt_wsadhz): Likewise.
13091 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13092 (iwmmxt_wabsdiffb): Likewise.
13093 (iwmmxt_wabsdiffh): Likewise.
13094 (iwmmxt_wabsdiffw): Likewise.
13095 (iwmmxt_waddsubhx): Likewise
13096 (iwmmxt_wsubaddhx): Likewise.
13097 (addc<mode>3): Likewise.
13098 (iwmmxt_avg4): Likewise.
13099 (iwmmxt_avg4r): Likewise.
13100 (iwmmxt_wmaddsx): Likewise.
13101 (iwmmxt_wmaddux): Likewise.
13102 (iwmmxt_wmaddsn): Likewise.
13103 (iwmmxt_wmaddun): Likewise.
13104 (iwmmxt_wmulwsm): Likewise.
13105 (iwmmxt_wmulwum): Likewise.
13106 (iwmmxt_wmulsmr): Likewise.
13107 (iwmmxt_wmulumr): Likewise.
13108 (iwmmxt_wmulwsmr): Likewise.
13109 (iwmmxt_wmulwumr): Likewise.
13110 (iwmmxt_wmulwl): Likewise.
13111 (iwmmxt_wqmulm): Likewise.
13112 (iwmmxt_wqmulwm): Likewise.
13113 (iwmmxt_wqmulmr): Likewise.
13114 (iwmmxt_wqmulwmr): Likewise.
13115 (iwmmxt_waddbhusm): Likewise.
13116 (iwmmxt_waddbhusl): Likewise.
13117 (iwmmxt_wqmiabb): Likewise.
13118 (iwmmxt_wqmiabt): Likewise.
13119 (iwmmxt_wqmiatb): Likewise.
13120 (iwmmxt_wqmiatt): Likewise.
13121 (iwmmxt_wqmiabbn): Likewise.
13122 (iwmmxt_wqmiabtn): Likewise.
13123 (iwmmxt_wqmiatbn): Likewise.
13124 (iwmmxt_wqmiattn): Likewise.
13125 (iwmmxt_wmiabb): Likewise.
13126 (iwmmxt_wmiabt): Likewise.
13127 (iwmmxt_wmiatb): Likewise.
13128 (iwmmxt_wmiatt): Likewise.
13129 (iwmmxt_wmiabbn): Likewise.
13130 (iwmmxt_wmiabtn): Likewise.
13131 (iwmmxt_wmiatbn): Likewise.
13132 (iwmmxt_wmiattn): Likewise.
13133 (iwmmxt_wmiawbb): Likewise.
13134 (iwmmxt_wmiawbt): Likewise.
13135 (iwmmxt_wmiawtb): Likewise.
13136 (iwmmxt_wmiawtt): Likewise.
13137 (iwmmxt_wmiawbbn): Likewise.
13138 (iwmmxt_wmiawbtn): Likewise.
13139 (iwmmxt_wmiawtbn): Likewise.
13140 (iwmmxt_wmiawttn): Likewise.
13141 (iwmmxt_wmerge): Likewise.
13142 (iwmmxt_tandc<mode>3): Likewise.
13143 (iwmmxt_torc<mode>3): Likewise.
13144 (iwmmxt_torvsc<mode>3): Likewise.
13145 (iwmmxt_textrc<mode>3): Likewise.
13146 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13147 (wmmxt_pack): Likewise.
13148 (wmmxt_mult_c1): Likewise.
13149 (wmmxt_mult_c2): Likewise.
13150 (wmmxt_alu_c1): Likewise.
13151 (wmmxt_alu_c2): Likewise.
13152 (wmmxt_alu_c3): Likewise.
13153 (wmmxt_transfer_c1): Likewise.
13154 (wmmxt_transfer_c2): Likewise.
13155 (wmmxt_transfer_c3): Likewise.
13156 (marvell_f_iwmmxt_wstr): Likewise.
13157 (marvell_f_iwmmxt_wldr): Likewise.
13158
bdb7bf8a
YZ		 131592013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13160
0ba6aff3 13161 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 13162
e6631d38
VM		 131632013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13164
13165 Revert:
13166 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13167 * lra-constraints.c (need_for_split_p): Check call used hard regs
13168 living through calls.
13169
13170 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13171 call used regs for call insn.
13172
53c1275b
JJ		 131732013-06-28 Jakub Jelinek <jakub@redhat.com>
13174
13175 PR target/57736
0ba6aff3
UB		 13176 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13177 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 13178
1141ed3f
BI		 131792013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13180
13181 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 13182
e32e4c4a
VM		 131832013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13184
13185 * lra-constraints.c (need_for_split_p): Check call used hard regs
13186 living through calls.
13187
f161bfd3
MM		 131882013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13189
13190 PR target/57744
13191 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13192 to tie with any other modes. Eliminate Altivec vector mode tests,
13193 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13194 code, to return 0 if testing MODE2 for a condition, if we've
13195 already tested MODE1 for the same condition.
13196
7eda14e1
MS		 131972013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13198
13199 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13200 layout.
13201
c822f852
MS		 132022013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13203
13204 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13205 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13206
da4f13a4
MS		 132072013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13208
0ba6aff3
UB		 13209 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13210 Define.
da4f13a4
MS		 13211 (aarch64_symbolic_constant_p): Remove.
13212 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13213 static. Fix line length and white space.
13214 (aarch64_symbolic_constant_p): Remove.
13215 * config/aarch64/predicates.md (aarch64_valid_symref):
13216 Use aarch64_classify_symbol_expression.
13217
956a95a5
KT		 132182013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13219
13220 * config/arm/constraints.md (Ts): New constraint.
13221 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13222 16-bit encodings.
13223 (compare_scc): Use "Ts" constraint for operand 0.
13224 (ior_scc_scc): Likewise.
13225 (and_scc_scc): Likewise.
13226 (and_scc_scc_nodom): Likewise.
13227 (ior_scc_scc_cmp): Likewise for operand 7.
13228 (and_scc_scc_cmp): Likewise.
13229 * config/arm/thumb2.md (thumb2_movsi_insn):
13230 Add alternatives for 16-bit encodings.
13231 (thumb2_movhi_insn): Likewise.
13232 (thumb2_movsicc_insn): Likewise.
13233 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13234 (thumb2_negscc): Use "Ts" constraint.
13235 Move mvn instruction outside cond_exec block.
13236 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13237 for 16-bit encodings.
13238
31dfce10
KT		 132392013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13240
13241 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13242 encoding.
13243 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13244 (mulsi3subsi): Likewise.
13245 (mulsidi3adddi): Likewise.
13246 (mulsidi3_v6): Likewise.
13247 (umulsidi3_v6): Likewise.
13248 (umulsidi3adddi_v6): Likewise.
13249 (smulsi3_highpart_v6): Likewise.
13250 (umulsi3_highpart_v6): Likewise.
13251 (mulhisi3tb): Likewise.
13252 (mulhisi3bt): Likewise.
13253 (mulhisi3tt): Likewise.
13254 (maddhisi4): Likewise.
13255 (maddhisi4tb): Likewise.
13256 (maddhisi4tt): Likewise.
13257 (maddhidi4): Likewise.
13258 (maddhidi4tb): Likewise.
13259 (maddhidi4tt): Likewise.
13260 (zeroextractsi_compare0_scratch): Likewise.
13261 (insv_zero): Likewise.
13262 (insv_t2): Likewise.
13263 (anddi_notzesidi_di): Likewise.
13264 (anddi_notsesidi_di): Likewise.
13265 (andsi_notsi_si): Likewise.
13266 (iordi_zesidi_di): Likewise.
13267 (xordi_zesidi_di): Likewise.
13268 (andsi_iorsi3_notsi): Likewise.
13269 (smax_0): Likewise.
13270 (smax_m1): Likewise.
13271 (smin_0): Likewise.
13272 (not_shiftsi): Likewise.
13273 (unaligned_loadsi): Likewise.
13274 (unaligned_loadhis): Likewise.
13275 (unaligned_loadhiu): Likewise.
13276 (unaligned_storesi): Likewise.
13277 (unaligned_storehi): Likewise.
13278 (extv_reg): Likewise.
13279 (extzv_t2): Likewise.
13280 (divsi3): Likewise.
13281 (udivsi3): Likewise.
13282 (arm_zero_extendhisi2addsi): Likewise.
13283 (arm_zero_extendqisi2addsi): Likewise.
13284 (compareqi_eq0): Likewise.
13285 (arm_extendhisi2_v6): Likewise.
13286 (arm_extendqisi2addsi): Likewise.
13287 (arm_movt): Likewise.
13288 (thumb2_ldrd): Likewise.
13289 (thumb2_ldrd_base): Likewise.
13290 (thumb2_ldrd_base_neg): Likewise.
13291 (thumb2_strd): Likewise.
13292 (thumb2_strd_base): Likewise.
13293 (thumb2_strd_base_neg): Likewise.
13294 (arm_negsi2): Add alternative for 16-bit encoding.
13295 (arm_one_cmplsi2): Likewise.
13296
1572e697
KT		 132972013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13298
13299 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13300 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13301 (movdfcc): Likewise.
13302 * config/arm/vfp.md (*thumb2_movsf_vfp):
13303 Disable predication for arm_restrict_it.
13304 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13305 (*thumb2_movdfcc_vfp): Likewise.
13306 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13307 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13308 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13309 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13310 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13311 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13312 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13313 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13314 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13315 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13316 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13317 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13318 Disable predication for arm_restrict_it.
13319
ebf8f0ea
KY		 133202013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13321
13322 * config/i386/bmiintrin.h (_bextr_u32): New.
13323 (_bextr_u64): Ditto.
13324
a290fcda
RS		 133252013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13326
13327 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13328 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13329 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13330 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13331 * config/mips/n32-elf.h: ...this new file.
13332
ea169996
MG		 133332013-06-27 Marc Glisse <marc.glisse@inria.fr>
13334
13335 PR target/57224
13336 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13337 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13338
12211b99 133392013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 13340
13341 * config/mips/mips-tables.opt: Regenerate.
13342 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 13343 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 13344 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 13345
2444244e
JJ		 133462013-06-27 Jakub Jelinek <jakub@redhat.com>
13347
cd06d2a2
JJ		 13348 PR target/57623
13349 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13350 constraints of operand 1 and 2.
13351
2444244e
JJ		 13352 PR target/57623
13353 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13354 to match RTL canonicalization. Swap predicates and
13355 constraints of operand 1 and 2.
13356
2f259720
VM		 133572013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13358
0ba6aff3
UB		 13359 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13360 Process OP_INOUT regs for splitting too.
2f259720 13361
fef4d2b3
JJ		 133622013-06-27 Jakub Jelinek <jakub@redhat.com>
13363
13364 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13365 decl before the loop, initialize to NULL.
13366 (vectorizable_load): Initialize ptr_incr to NULL.
13367
79ee9826
MJ		 133682013-06-27 Martin Jambor <mjambor@suse.cz>
13369
13370 PR lto/57208
13371 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13372 * ipa-ref.c (ipa_maybe_record_reference): New function.
13373 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13374 * ipa-cp.c (create_specialized_node): Record potential references from
13375 aggvals.
13376 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13377
e18b4a81
YZ		 133782013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13379
13380 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13381 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13382 'mode' to force_reg.
13383 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13384 (aarch64_expand_mov_immediate): Likewise.
13385
9c023bf0
YZ		 133862013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13387
13388 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13389 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13390
5a3fe9b6
AK		 133912013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13392
13393 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13394 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13395 (struct machine_function): Add tbegin_p.
13396 (s390_canonicalize_comparison): Fold CC mode compares to
13397 conditional jump if possible.
13398 (s390_emit_jump): Return the emitted jump.
13399 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13400 Handle CCRAWmode compares.
13401 (s390_option_override): Default to -mhtm if available.
13402 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13403 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13404 FPRs instead of df_regs_ever_live_p.
13405 (s390_optimize_nonescaping_tx): New function.
13406 (s390_init_frame_layout): Extend clobbered_regs array to cover
13407 FPRs as well.
13408 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13409 (s390_expand_tbegin): New function.
13410 (enum s390_builtin): New enum definition.
13411 (code_for_builtin): New array definition.
13412 (s390_init_builtins): New function.
13413 (s390_expand_builtin): New function.
13414 (TARGET_INIT_BUILTINS): Define.
13415 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 13416 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 13417 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13418 (s390_alc_comparison): Likewise.
13419 * config/s390/s390-modes.def: Add CCRAWmode.
13420 * config/s390/s390.h (processor_flags): Add PF_TX.
13421 (TARGET_CPU_HTM): Define macro.
13422 (TARGET_HTM): Define macro.
13423 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13424 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13425 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13426 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 13427 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 13428 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13429 ("*cc_to_int"): Move up.
13430 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13431 constants other than 0.
13432 ("*ccraw_to_int"): New insn and splitter definition.
13433 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13434 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13435 ("tx_assist"): New expander.
13436 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13437 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13438 * config/s390/s390.opt: Add -mhtm option.
13439 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13440 * config/s390/htmxlintrin.h: New file.
13441 * config/s390/htmintrin.h: New file.
13442 * config/s390/s390intrin.h: New file.
13443 * doc/extend.texi: Document htm builtins.
13444 * config.gcc: Add the new header files to extra_headers.
13445
9e216629
TS		 134462013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13447
13448 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13449 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13450
fd8c65e7 134512013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 13452 Pat Haugen <pthaugen@us.ibm.com>
13453 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 13454
13455 * config/rs6000/power8.md: New.
13456 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13457 setting for power8 entry.
13458 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13459 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13460 test for Power4/Power5 only.
13461 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13462 support.
13463 (force_new_group): Adjust comment.
13464 * config/rs6000/rs6000.md: Include power8.md.
13465
b24a2ce5
GY		 134662013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13467
13468 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13469 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13470 declaration.
13471 (tune_params): Update comment.
13472 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13473 (arm_max_conditional_execute): New function.
13474 (thumb2_final_prescan_insn): Use max_insn_skipped and
13475 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13476
f6fdeb62
JJ		 134772013-06-25 Jakub Jelinek <jakub@redhat.com>
13478
13479 PR tree-optimization/57705
13480 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13481 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13482 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 13483 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13484
c13bc3d9
MJ		 134852013-06-25 Martin Jambor <mjambor@suse.cz>
13486
13487 PR middle-end/57670
13488 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13489 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13490 calls in the dump.
13491 (ipa_note_param_call): Initialize member_ptr flag.
13492 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13493 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13494 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13495 (ipa_read_indirect_edge_info): Likewise.
13496
8a845901
RB		 134972013-06-25 Richard Biener <rguenther@suse.de>
13498
13499 PR middle-end/56977
13500 * passes.c (init_optimization_passes): Move pass_fold_builtins
13501 and pass_dce earlier with -Og.
13502
33be0bec
EB		 135032013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13504
13505 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13506 <BIT_FIELD_REF>: Remove trailing TAB.
13507 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13508 remove blank line.
13509
5fe8e757
MJ		 135102013-06-24 Martin Jambor <mjambor@suse.cz>
13511
13512 PR tree-optimization/57358
13513 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13514 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13515 (ipa_analyze_params_uses): Generate pessimistic info when true.
13516
44a60244
MJ		 135172013-06-24 Martin Jambor <mjambor@suse.cz>
13518
13519 PR tree-optimization/57539
13520 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13521 global.inlined_to of the new node to it. All callers changed.
13522 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13523 inlining_into, pass it to cgraph_clone_node.
13524 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13525 ipa_free_edge_args_substructures.
13526 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13527 rdesc linked list. Do not assert rdesc edges have inlined caller.
13528 Assert we have found an rdesc in the rdesc list.
13529
7c5848b8
RB		 135302013-06-24 Richard Biener <rguenther@suse.de>
13531
0ba6aff3 13532 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 13533 (pointer_set_lookup): Declare.
13534 (class pointer_map): New template class implementing a
13535 generic pointer to T map.
13536 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13537 pointer_map<T>::contains, pointer_map<T>::insert,
13538 pointer_map<T>::traverse): New functions.
13539 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13540 (pointer_set_lookup): New function.
13541 (pointer_set_contains): Use pointer_set_lookup.
13542 (pointer_set_insert): Likewise.
13543 (insert_aux): Remove.
13544 (struct pointer_map_t): Embed a pointer_set_t.
13545 (pointer_map_create): Adjust.
13546 (pointer_map_destroy): Likewise.
13547 (pointer_map_contains): Likewise.
13548 (pointer_map_insert): Likewise.
13549 (pointer_map_traverse): Likewise.
13550 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13551 pointer_map<unsigned> instead of a pointer_map_t.
13552 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13553 (streamer_tree_cache_lookup): Likewise.
13554 (streamer_tree_cache_create): Likewise.
13555 (streamer_tree_cache_delete): Likewise.
13556 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13557 pointer_map<unsigned> instead of a pointer_map_t.
13558 (lto_init_tree_ref_encoder): Adjust.
13559 (lto_destroy_tree_ref_encoder): Likewise.
13560 * lto-section-out.c (lto_output_decl_index): Likewise.
13561 (lto_record_function_out_decl_state): Likewise.
13562 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13563
3dbc97a9
RB		 135642013-06-24 Richard Biener <rguenther@suse.de>
13565
13566 PR tree-optimization/57488
13567 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13568
8345965a
AM		 135692013-06-24 Alan Modra <amodra@gmail.com>
13570
13571 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13572 (gen_easy_altivec_constant): Likewise.
13573 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13574 easy_vector_constant_msb): Likewise.
13575
b7df379f
JJ		 135762013-06-23 Jakub Jelinek <jakub@redhat.com>
13577
13578 PR target/57688
13579 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13580 add missing return true.
13581
12211b99 135822013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 13583
13584 PR target/52483
13585 * config/sh/predicates.md (general_extend_operand): Invoke
13586 general_movsrc_operand for memory operands.
13587 (general_movsrc_operand): Allow reg+reg addressing, do not use
13588 general_operand for memory operands.
13589
97db2bf7
ST		 135902013-06-23 Sriraman Tallam <tmsriram@google.com>
13591
13592 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13593 when current target options does not apply.
13594 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13595 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 13596 * config/i386/bmiintrin.h: Pass appropriate target
13597 attributes to header.
97db2bf7
ST		 13598 * config/i386/mmintrin.h: Ditto.
13599 * config/i386/nmmintrin.h: Ditto.
13600 * config/i386/avx2intrin.h: Ditto.
13601 * config/i386/fxsrintrin.h: Ditto.
13602 * config/i386/tbmintrin.h: Ditto.
13603 * config/i386/xsaveintrin.h: Ditto.
13604 * config/i386/f16cintrin.h: Ditto.
13605 * config/i386/xtestintrin.h: Ditto.
13606 * config/i386/xsaveoptintrin.h: Ditto.
13607 * config/i386/bmi2intrin.h: Ditto.
13608 * config/i386/lzcntintrin.h: Ditto.
13609 * config/i386/smmintrin.h: Ditto.
13610 * config/i386/wmmintrin.h: Ditto.
13611 * config/i386/x86intrin.h: Remove all header include guards.
13612 * config/i386/prfchwintrin.h: Ditto.
13613 * config/i386/pmmintrin.h: Ditto.
13614 * config/i386/tmmintrin.h: Ditto.
13615 * config/i386/xmmintrin.h: Ditto.
13616 * config/i386/popcntintrin.h: Ditto.
13617 * config/i386/rdseedintrin.h: Ditto.
13618 * config/i386/ammintrin.h: Ditto.
13619 * config/i386/emmintrin.h: Ditto.
13620 * config/i386/immintrin.h: Remove all header include guards.
13621 * config/i386/fma4intrin.h: Ditto.
13622 * config/i386/lwpintrin.h: Ditto.
13623 * config/i386/xopintrin.h: Ditto.
13624 * config/i386/ia32intrin.h: Ditto.
13625 * config/i386/avxintrin.h: Ditto.
13626 * config/i386/rtmintrin.h: Ditto.
13627 * config/i386/fmaintrin.h: Ditto.
13628 * config/i386/mm3dnow.h: Ditto.
13629
495e6879
ST		 136302013-06-22 Sriraman Tallam <tmsriram@google.com>
13631
13632 * common/config/i386/i386-common.c: Handle LZCNT.
13633
11e8f687
AK		 136342013-06-22 Andi Kleen <ak@linux.intel.com>
13635
13636 * doc/extend.texi: Use __atomic_store_n instead of
13637 __atomic_store in HLE example.
13638
12211b99 136392013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 13640
13641 * config/sh/sh.c: Remove <cstdlib> workaround.
13642
927f908b
AK		 136432013-06-21 Andi Kleen <ak@linux.intel.com>
13644
0ba6aff3 13645 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13646
cca410b6
AK		 136472013-06-21 Andi Kleen <ak@linux.intel.com>
13648
13649 * doc/extend.texi: Document that __atomic_clear and
13650 __atomic_test_and_set should only be used with bool.
13651
d84db1ed
JH		 136522013-06-20 Jan Hubicka <jh@suse.cz>
13653
13654 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13655 types_same_for_odr.
13656 * tree.c (decls_same_for_odr): New function.
13657 (same_for_edr): New function.
13658 (types_same_for_odr): New function.
13659 (get_binfo_at_offset): Use it.
13660 * tree.h (types_same_for_odr): Declare.
13661
12211b99 136622013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 13663 Jason Merrill <jason@redhat.com>
13664
13665 * system.h: Include <cstdlib> as well as <stdlib.h>.
13666
49b4ff63
UB		 136672013-06-20 Uros Bizjak <ubizjak@gmail.com>
13668
13669 PR target/57655
13670 * config/i386/i386.c (construct_container): Report error if
13671 long double is used with disabled x87 float returns.
13672
7a89ad00
JH		 136732013-06-20 Jan Hubicka <jh@suse.cz>
13674
13675 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13676
12211b99 136772013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 13678
13679 PR rtl-optimization/57425
13680 PR rtl-optimization/57569
13681 * alias.c (write_dependence_p): Remove parameters mem_mode and
13682 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13683 Changed all callers.
13684 (canon_anti_dependence): Get comments and semantics in sync.
13685 Add parameter mem_canonicalized. Changed all callers.
13686 * rtl.h (canon_anti_dependence): Update prototype.
13687
d16e9a99
RB		 136882013-06-20 Richard Biener <rguenther@suse.de>
13689
13690 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13691 byte case, inline streamer_read_uchar and defer section
13692 overrun check.
13693
c3a9b91b
RB		 136942013-06-20 Richard Biener <rguenther@suse.de>
13695
13696 PR tree-optimization/57584
13697 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13698 SSA names into the expanded expression that take part in
13699 abnormal coalescing.
13700
371e77e3 137012013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 13702
13703 * gcov.c (print_usage): Handle new option.
13704 (process_args): Ditto.
13705 (get_gcov_intermediate_filename): New function.
13706 (output_intermediate_file): New function.
13707 (output_gcov_file): New function
13708 (generate_results): Handle new option.
13709 (release_function): Relase demangled name.
13710 (read_graph_file): Handle demangled name.
13711 (output_lines): Ditto.
13712 * doc/gcov.texi: Document gcov intermediate format.
13713
11f5b71a
VM		 137142013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13715
13716 PR bootstrap/57604
13717 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13718 (lra_emit_add): Use the functions. Add comment about Y as an
13719 address segment.
13720
e1f9340b
DE		 137212013-06-19 David Edelsohn <dje.gcc@gmail.com>
13722
13723 PR driver/57652
13724 * collect2.c (collect_atexit): New.
13725 (collect_exit): Delete.
13726 (main): Register collect_atexit with atexit.
13727 (collect_wait): Change collect_exit to exit.
13728 (do_wait): Same.
13729 * collect2.h (collect_exit): Delete.
13730 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13731
e3f9e0ac
WM		 137322013-06-19 Wei Mi <wmi@google.com>
13733
13734 PR rtl-optimization/57518
13735 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13736 if regno is used in paradoxical subreg.
13737 (update_equiv_regs): Check pdx_subregs[regno] before
13738 set a reg to be equivalent with a mem.
13739
a09f1a76
MK		 137402013-06-19 Matthias Klose <doko@ubuntu.com>
13741
13742 PR driver/57651
13743 * file-find.h (find_a_file): Add a mode parameter.
13744 * file-find.c (find_a_file): Likewise.
13745 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13746 with X_OK for the executables.
13747 * collect2.c (main): Call find_a_file with X_OK.
13748
7d18b0ad
SE		 137492013-06-19 Steve Ellcey <sellcey@mips.com>
13750
13751 PR target/56942
13752 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13753 Use NEXT_INSN instead of next_real_insn.
13754
6a6dac52
JH		 137552013-06-19 Jan Hubicka <jh@suse.cz>
13756
13757 * cgraph.h (const_value_known_p): Replace by ...
13758 (ctor_for_folding): .. this one.
13759 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13760 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13761 * expr.c (expand_expr_real_1): Likewise.
13762 (string_constant): Likewise.
13763 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13764 * ipa.c (process_references): Likewise.
13765 (symtab_remove_unreachable_nodes): Likewise.
13766 * ipa-inline-analysis.c (param_change_prob): Likewise.
13767 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13768 (get_base_constructor): Likwise.
13769 * varpool.c (varpool_remove_node): Likewise.
13770 (varpool_remove_initializer): LIkewise.
13771 (dump_varpool_node): LIkwise.
13772 (const_value_known_p): Rewrite to ...
13773 (ctor_for_folding): ... this one.
13774
216c12ab
JJ		 137752013-06-19 Jakub Jelinek <jakub@redhat.com>
13776
13777 PR driver/57651
13778 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13779 PERSONALITY in $PATH derived prefixes.
13780
d53e2f99
JL		 137812013-06-19 Jeff Law <law@redhat.com>
13782
ecdbf306
JL		 13783 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13784 in comment.
13785
d53e2f99
JL		 13786 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13787 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13788 booleans.
13789
89e7fc04
SN		 137902013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13791
13792 * config/arm/vfp.md: Move VFP instruction classification documentation
13793 to ...
13794 * config/arm/arm.md: ... here. Update instruction classification
13795 documentation.
13796
5e7f6aaa
RE		 137972013-06-19 Richard Earnshaw <rearnsha@arm.com>
13798
13799 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13800 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13801 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13802
5fdc1e5d
SB		 138032013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13804
13805 PR target/57609
13806 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13807 with NEXT_INSN. Use tablejump_p to check for jump table data
13808 insns.
13809
e952afba
PC		 138102013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13811
13812 PR c++/56544
13813 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13814 that now in C++ the value is correct per the C++ standards.
13815
3fd0303a
RB		 138162013-06-19 Richard Biener <rguenther@suse.de>
13817
13818 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13819 for global context.
13820
b04e4b08
AK		 138212013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13822
13823 Revert:
13824 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13825
13826 PR target/57609
13827 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13828 with next_active_insn.
13829
bef8491a
ST		 138302013-06-18 Sriraman Tallam <tmsriram@google.com>
13831
13832 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13833 functions are inlined during failures to flag an error.
13834 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13835 in early inline pass.
22fc7d3d 13836
c21b257d
L		 138372013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13838
13839 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13840 in comments.
13841
33255ae3
JB		 138422013-06-18 Julian Brown <julian@codesourcery.com>
13843
13844 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13845 Permit virtual register pre-reload if !strict.
13846 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13847 change.
13848 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13849 prototype.
13850 * config/arm/neon.md (movmisalign<mode>): Use
13851 neon_perm_struct_or_reg_operand instead of
13852 neon_struct_or_register_operand.
13853 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13854 neon_permissive_struct_operand instead of neon_struct_operand.
13855 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13856 neon_vector_mem_operand.
13857 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13858 neon_vector_mem_operand.
13859 (neon_permissive_struct_operand): New.
13860 (neon_struct_or_register_operand): Rename to...
13861 (neon_perm_struct_or_reg_operand): This. Adjust call to
13862 neon_vector_mem_operand.
13863
d579fcda
RB		 138642013-06-18 Richard Biener <rguenther@suse.de>
13865
13866 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13867 * lto-streamer.h: Include pointer-set.h.
13868 (struct lto_decl_slot): Remove.
13869 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13870 Remove next_index entry.
13871 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13872 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13873 (lto_init_tree_ref_encoder): Adjust.
13874 (lto_destroy_tree_ref_encoder): Likewise.
13875 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13876 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13877 (lto_output_decl_index): Adjust.
13878 (lto_new_out_decl_state): Likewise.
13879 (lto_record_function_out_decl_state): Likewise.
13880 * lto-streamer-out.c (copy_function): Likewise.
13881
31ee20ba
RB		 138822013-06-18 Richard Biener <rguenther@suse.de>
13883
13884 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13885 * cgraphunit.c: Include cfgloop.h.
13886 (init_lowered_empty_function): Initialize the loop tree.
13887 (assemble_thunk): Insert new BBs into loops.
13888
ac0511f2
RB		 138892013-06-18 Richard Biener <rguenther@suse.de>
13890
13891 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13892 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13893 the map from cache entry to cache index optional.
13894 (streamer_tree_cache_replace_tree): Adjust accordingly.
13895 (streamer_tree_cache_append): Likewise.
13896 (streamer_tree_cache_delete): Likewise.
13897 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13898 streamer cache map from cache entry to cache index.
13899 * lto-streamer-out.c (create_output_block): Adjust.
13900
09485a08
SN		 139012013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13902
22fc7d3d
UB		 13903 * config/arm/arm.md (attribute "insn"): Move multiplication and
13904 division attributes to...
09485a08
SN		 13905 (attribute "type"): ... here. Remove mult.
13906 (attribute "mul32"): New attribute.
13907 (attribute "mul64"): Add umaal.
13908 (*arm_mulsi3): Update attributes.
13909 (*arm_mulsi3_v6): Likewise.
13910 (*thumb_mulsi3): Likewise.
13911 (*thumb_mulsi3_v6): Likewise.
13912 (*mulsi3_compare0): Likewise.
13913 (*mulsi3_compare0_v6): Likewise.
13914 (*mulsi_compare0_scratch): Likewise.
13915 (*mulsi_compare0_scratch_v6): Likewise.
13916 (*mulsi3addsi): Likewise.
13917 (*mulsi3addsi_v6): Likewise.
13918 (*mulsi3addsi_compare0): Likewise.
13919 (*mulsi3addsi_compare0_v6): Likewise.
13920 (*mulsi3addsi_compare0_scratch): Likewise.
13921 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13922 (*mulsi3subsi): Likewise.
13923 (*mulsidi3adddi): Likewise.
13924 (*mulsi3addsi_v6): Likewise.
13925 (*mulsidi3adddi_v6): Likewise.
13926 (*mulsidi3_nov6): Likewise.
13927 (*mulsidi3_v6): Likewise.
13928 (*umulsidi3_nov6): Likewise.
13929 (*umulsidi3_v6): Likewise.
13930 (*umulsidi3adddi): Likewise.
13931 (*umulsidi3adddi_v6): Likewise.
13932 (*smulsi3_highpart_nov6): Likewise.
13933 (*smulsi3_highpart_v6): Likewise.
13934 (*umulsi3_highpart_nov6): Likewise.
13935 (*umulsi3_highpart_v6): Likewise.
13936 (mulhisi3): Likewise.
13937 (*mulhisi3tb): Likewise.
13938 (*mulhisi3bt): Likewise.
13939 (*mulhisi3tt): Likewise.
13940 (maddhisi4): Likewise.
13941 (*maddhisi4tb): Likewise.
13942 (*maddhisi4tt): Likewise.
13943 (maddhidi4): Likewise.
13944 (*maddhidi4tb): Likewise.
13945 (*maddhidi4tt): Likewise.
13946 (divsi3): Likewise.
13947 (udivsi3): Likewise.
13948 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13949 (thumb2_mulsi_short_compare0): Likewise.
13950 (thumb2_mulsi_short_compare0_scratch): Likewise.
13951 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13952 (1020mult2): Likewise.
13953 (1020mult3): Likewise.
13954 (1020mult4): Likewise.
13955 (1020mult5): Likewise.
13956 (1020mult6): Likewise.
22fc7d3d
UB		 13957 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13958 change.
09485a08
SN		 13959 (cortex_a15_mult64): Likewise.
13960 (cortex_a15_sdiv): Likewise.
13961 (cortex_a15_udiv): Likewise.
13962 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13963 (mult2): Likewise.
13964 (mult3): Likewise.
13965 (mult4): Likewise.
13966 (mult5): Likewise.
13967 (mult6): Likewise.
13968 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13969 (pj4_ir_div): Likewise.
13970 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13971 (11_mult2): Likewise.
13972 (11_mult3): Likewise.
13973 (11_mult4): Likewise.
13974 (11_mult5): Likewise.
13975 (11_mult6): Likewise.
13976 (11_mult7): Likewise.
13977 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13978 (cortex_a8_mla): Likewise.
13979 (cortex_a8_mull): Likewise.
13980 (cortex_a8_smulwy): Likewise.
13981 (cortex_a8_smlald): Likewise.
13982 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13983 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13984 (cortex_r4_mul_3): Likewise.
13985 (cortex_r4_mla_4): Likewise.
13986 (cortex_r4_mla_3): Likewise.
13987 (cortex_r4_smlald): Likewise.
13988 (cortex_r4_mull): Likewise.
13989 (cortex_r4_sdiv): Likewise.
13990 (cortex_r4_udiv): Likewise.
13991 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13992 (cortex_a7_idiv): Likewise.
13993 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13994 (9_mult2): Likewise.
13995 (9_mult3): Likewise.
13996 (9_mult4): Likewise.
13997 (9_mult5): Likewise.
13998 (9_mult6): Likewise.
13999 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14000 (cortex_a53_sdiv): Likewise.
14001 (cortex_a53_udiv): Likewise.
14002 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14003 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14004 (mp626_mult2): Likewise.
14005 (mp626_mult3): Likewise.
14006 (mp626_mult4): Likewise.
14007 * config/arm/fa526.md (526_mult1): Update attribute change.
14008 (526_mult2): Likewise.
14009 * config/arm/arm-generic.md (mult): Update attribute change.
14010 (mult_ldsched_strongarm): Likewise.
14011 (mult_ldsched): Likewise.
14012 (multi_cycle): Likewise.
14013 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14014 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14015 (606te_mult2): Likewise.
14016 (606te_mult3): Likewise.
14017 (606te_mult4): Likewise.
14018 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14019 (cortex_a9_mac16): Likewise.
14020 (cortex_a9_multiply): Likewise.
14021 (cortex_a9_mac): Likewise.
14022 (cortex_a9_multiply_long): Likewise.
14023 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14024 (626te_mult2): Likewise.
14025 (626te_mult3): Likewise.
14026 (626te_mult4): Likewise.
14027
8951f345
RB		 140282013-06-18 Richard Biener <rguenther@suse.de>
14029
14030 PR lto/57334
14031 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14032
d8101d05
AK		 140332013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14034
14035 PR target/57609
14036 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14037 with next_active_insn.
14038
69eff9da
AM		 140392013-06-18 Alan Modra <amodra@gmail.com>
14040
14041 * config/rs6000/rs6000.h (enum data_align): New.
14042 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14043 (DATA_ABI_ALIGNMENT): Define.
14044 (CONSTANT_ALIGNMENT): Correct comment.
14045 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14046 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14047
8ed00d76
DM		 140482013-06-17 David Malcolm <dmalcolm@redhat.com>
14049
14050 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14051 ATTRIBUTE_UNUSED marking.
14052
3d33d06b
SN		 140532013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14054
14055 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14056 alternative and update.
14057 (aarch64_dup_lanedi): Delete.
14058 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14059 * config/aarch64/aarch64-simd-builtins.def: Update.
14060
ee03e71d
RB		 140612013-06-17 Richard Biener <rguenther@suse.de>
14062
14063 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14064 (lto_input_scc): Declare.
14065 (lto_input_tree_1): Likewise.
14066 (struct lto_stats_d): Add num_tree_bodies_output and
14067 num_pickle_refs_output.
14068 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14069 (lto_read_tree_1): Split out from ...
14070 (lto_read_tree): ... this.
14071 (lto_input_scc): New function.
14072 (lto_input_tree_1): Split out from ...
14073 (lto_input_tree): ... this. Handle LTO_tree_scc.
14074 (lto_data_in_create): Create the streamer cache without hashes.
14075 * lto-streamer-out.c (create_output_block): Create the streamer
14076 cache with hashes when not doing WPA.
14077 (lto_write_tree_1): Split out from ...
14078 (lto_write_tree): ... this.
14079 (get_symbol_initial_value): New function.
14080 (lto_output_tree_1): Split out from ...
14081 (lto_output_tree): ... this. Write trees as series of SCCs
14082 using a DFS walk via DFS_write_tree.
14083 (struct sccs, struct scc_entry): New types.
14084 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14085 (DFS_write_tree_body): New function.
14086 (DFS_write_tree): Likewise.
14087 (hash_tree): Likewise.
14088 (scc_entry_compare): Likewise.
14089 (hash_scc): Likewise.
14090 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14091 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14092 TREE_CHAIN as regular reference.
14093 (streamer_read_integer_cst): Remove.
14094 (streamer_get_pickled_tree): Adjust.
14095 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14096 of DECL_EXTERNALs in BLOCK_VARS for now.
14097 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14098 reference.
14099 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14100 Add hash value argument and record that if hashes are recorded
14101 in the cache.
14102 (streamer_tree_cache_insert_1): Adjust.
14103 (streamer_tree_cache_insert): Likewise.
14104 (streamer_tree_cache_insert_at): Rename to ...
14105 (streamer_tree_cache_replace_tree): ... this and adjust.
14106 (streamer_tree_cache_append): Adjust.
14107 (record_common_node): Likewise.
14108 (streamer_tree_cache_create): Add argument whether to
14109 record hash values together with trees.
14110 (streamer_tree_cache_delete): Adjust.
14111 * tree-streamer.h (struct streamer_tree_cache_d): Add
14112 vector of hashes.
14113 (streamer_read_integer_cst): Remove.
14114 (streamer_tree_cache_insert): Adjust.
14115 (streamer_tree_cache_append): Likewise.
14116 (streamer_tree_cache_insert_at): Rename to ...
14117 (streamer_tree_cache_replace_tree): ... this and adjust.
14118 (streamer_tree_cache_create): Add argument whether to record hashes.
14119 (streamer_tree_cache_get): Rename to ...
14120 (streamer_tree_cache_get_tree): ... this.
14121 (streamer_tree_cache_get_hash): New function.
14122 * tree.c (cache_integer_cst): New function.
14123 * tree.h (cache_integer_cst): Declare.
14124 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14125 * lto-symtab.c (lto_varpool_replace_node): Only release
14126 DECL_INITIAL of non-prevailing decls.
14127 * varpool.c (varpool_remove_initializer): Do not release
14128 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14129
0de86a92
JU		 141302013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14131
14132 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14133 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14134 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14135 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14136 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14137 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14138 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14139 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14140 instead of TARGET_64BIT.
14141 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14142 Require ISA_HAS_<D>DIV.
14143
74240413
RS		 141442013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14145
14146 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14147 (mips*-*-linux*): Move default with_llsc setting to where other
14148 defaults are set.
14149 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14150 with_arch block.
14151 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14152 Likewise. Remove default with_tune setting. Move default float
14153 setting to its own block. Handle with_llsc in the same block as above.
14154
12211b99 141552013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 14156
14157 PR rtl-optimization/57425
14158 PR rtl-optimization/57569
299a5f6a 14159 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 14160 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14161 Changed all callers.
14162 (canon_anti_dependence): New function.
14163 * cse.c (check_dependence): Use canon_anti_dependence.
14164 * cselib.c (cselib_invalidate_mem): Likewise.
14165 * rtl.h (canon_anti_dependence): Declare.
14166
b259d352
JU		 141672013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14168
14169 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14170 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14171 ".set mips3" for 64-bit targets.
14172
4de09b85
DC		 141732013-06-15 Dehao Chen <dehao@google.com>
14174
14175 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14176 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14177 (gimple_check_call_args): Likewise.
14178 * value-prof.c (check_ic_target): Likewise.
14179 * ipa-inline.c (early_inliner): Likewise.
14180 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14181 * cgraph.c (cgraph_create_edge_1): Likewise.
14182 (cgraph_make_edge_direct): Likewise.
14183
d1aa4a76
MM		 141842013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14185
14186 PR target/57615
14187 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14188 rs6000_output_move_128bit to handle emitting quad memory
14189 operations. Set attribute length to 8 bytes.
14190
12211b99 141912013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 14192
14193 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14194 New pattern.
14195 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14196 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14197 (aarch64_<su>mlsl<mode>): Likewise.
14198
f5273574
MS		 141992013-06-14 Mike Stump <mikestump@comcast.net>
14200
14201 * Makefile.in (TARGET_H): Add insn-codes.h.
14202
ef7e7ab2
AM		 142032013-06-14 Alan Modra <amodra@gmail.com>
14204
14205 PR middle-end/57134
14206 PR middle-end/57586
14207 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14208 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14209 bitfield expansion when EXPAND_MEMORY.
14210 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14211
135faab6
MM		 142122013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14213
14214 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14215 test for clearing quad memory on 32-bit later.
14216
948a1fd9
MG		 142172013-06-13 Marc Glisse <marc.glisse@inria.fr>
14218
14219 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14220 (fold_negate_expr): Likewise.
14221 (fold_real_zero_addition_p): Handle vectors.
14222 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14223
0597fb9c
AM		 142242013-06-14 Alan Modra <amodra@gmail.com>
14225
14226 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14227
e5b962d0
JH		 142282013-06-13 Jan Hubicka <jh@suse.cz>
14229
14230 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14231 Local comdats are not externally visible.
14232 * symtab.c (dump_symtab_base): Dump externally visible.
14233 (verify_symtab_base): Verify back links in the symtab hash.
14234
a8355e51
BC		 142352013-06-13 Bin Cheng <bin.cheng@arm.com>
14236
14237 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14238 CONVERT_EXPR as equal nodes.
14239
8f7e6e33
BC		 142402013-06-13 Bin Cheng <bin.cheng@arm.com>
14241
22fc7d3d 14242 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 14243
5d418483
MG		 142442013-06-13 Marc Glisse <marc.glisse@inria.fr>
14245
14246 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14247 Generalize to complex and vector.
14248 * tree.c (build_all_ones_cst): New function.
14249 * tree.h (build_all_ones_cst): Declare it.
14250
50751417
AM		 142512013-06-13 Alan Modra <amodra@gmail.com>
14252
14253 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14254 * config/rs6000/rs6000.md (signbittf2): New insn.
14255 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14256 (abstf2_internal, cmptf_internal2): Likewise.
14257 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14258
4b02c962 142592013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14260 Pat Haugen <pthaugen@us.ibm.com>
14261 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 14262
14263 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14264 power8 byte, half-word, and quad-word atomic instructions.
14265 (emit_store_conditional): Likewise.
14266 (rs6000_expand_atomic_compare_and_swap): Likewise.
14267 (rs6000_expand_atomic_op): Likewise.
14268
14269 * config/rs6000/sync.md (larx): Add new modes for power8.
14270 (stcx): Likewise.
14271 (AINT): New mode iterator to include TImode as well as normal
14272 integer modes on power8.
14273 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14274 that VSX registers are not considered. Use AINT mode iterator
14275 instead of INT1 to allow inclusion of quad word atomic operations
14276 on power8.
14277 (load_locked<mode>): Likewise.
14278 (store_conditional<mode>): Likewise.
14279 (atomic_compare_and_swap<mode>): Likewise.
14280 (atomic_exchange<mode>): Likewise.
14281 (atomic_nand<mode>): Likewise.
14282 (atomic_fetch_<fetchop_name><mode>): Likewise.
14283 (atomic_nand_fetch<mode>): Likewise.
14284 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14285 each type.
14286 (ATOMIC): On power8, add QImode, HImode modes.
14287 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14288 modes that promote to SImode.
14289 (load_lockedti): Convert TImode arguments to PTImode, so that we
14290 get a guaranteed even/odd register pair.
14291 (load_lockedpti): Likewise.
14292 (store_conditionalti): Likewise.
14293 (store_conditionalpti): Likewise.
14294
14295 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14296 atomic load/store instructions.
14297 (HSI): Likewise.
14298
710b6cc1
RS		 142992013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14300
14301 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14302 loads.
14303 (insn_count): New attribute, with most cases extracted from...
14304 (length): ...here. Redefine most cases in terms of insn_count.
14305 (single_insn): Delete.
14306 (can_delay): Use insn_count to check for single instructions.
14307 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14308 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14309 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14310 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14311 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14312 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14313 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14314 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14315 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14316 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14317 rather than "length".
14318 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14319 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14320 Use "insn_count" rather than "length".
14321 * config/mips/mips-dsp.md
14322 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14323 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14324 length attributes.
14325
1951f101
MG		 143262013-06-12 Marc Glisse <marc.glisse@inria.fr>
14327
14328 PR tree-optimization/57361
14329 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14330
8b033a8a
SN		 143312013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14332
22fc7d3d
UB		 14333 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14334 to split.
8b033a8a
SN		 14335 (aarch64_simd_combine<mode>): New instruction expansion.
14336 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14337 function prototype.
14338 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14339 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14340
ca0f62a8
JH		 143412013-06-12 Jan Hubicka <jh@suse.cz>
14342
14343 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14344 decl has when in streaming stage.
14345 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14346 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14347
12211b99 143482013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 14349
14350 PR target/57578
14351 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14352
bfdeda2c
JJ		 143532013-06-12 Jakub Jelinek <jakub@redhat.com>
14354
14355 PR tree-optimization/57537
14356 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14357 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14358
a4fa02d1
RB		 143592013-06-12 Richard Biener <rguenther@suse.de>
14360
14361 * data-streamer.h (streamer_write_char_stream): CSE
14362 obs->current_pointer.
14363 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14364 streamer_write_char_stream manually and optimize the resulting loop.
14365 (streamer_write_hwi_stream): Likewise.
14366
bbf9ad07
JH		 143672013-06-12 Jan Hubicka <jh@suse.cz>
14368
14369 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14370 * cgraph.h (varpool_create_empty_node): Declare.
14371 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14372 duplicated nodes.
14373 * symtab.c (symtab_unregister_node): Be lax about missin entries
14374 in node hash.
14375 (symtab_get_node): Update comment.
14376 * varpool.c (varpool_create_empty_node): Break out from ...
14377 (varpool_node_for_decl): ... here.
14378 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14379
a53f90ad
EB		 143802013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14381
14382 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14383 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14384 part. Use straight-line flow at the end.
14385 <COMPONENT_REF>: Remove superfluous else.
14386 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14387
946f9306
JJ		 143882013-06-12 Jakub Jelinek <jakub@redhat.com>
14389
14390 PR target/56564
14391 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14392 target hook even for !TREE_PUBLIC decls. If no resolution info
14393 is available, return false for common and external decls.
14394
a7fca6f0
KP		 143952013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14396
14397 * config/rl78/constraints.md (U): New constraint.
14398 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14399 valloc attribute.
14400
c7d8f446
MM		 144012013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14402
14403 PR target/57589
14404 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14405 to allow returning address to AT_PLATFORM name.
14406
08346abd
JH		 144072013-06-11 Jan Hubicka <jh@suse.cz>
14408
14409 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14410 * cgraph.h (symtab_node_base): Add weakref flag.
14411 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14412 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14413 (output_weakrefs): Use weakref flag.
14414 * fold-const.c (simple_operand_p): Handle WEAK.
14415 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14416 * ipa.c (varpool_externally_visible_p): Drop weakref.
14417 (function_and_variable_visibility): Update comment; fix weakref
14418 sanity checks; do not clear DECL_WEAK on them.
14419 * lto-cgraph.c (lto_output_node): update.
14420 (lto_output_varpool_node): Update.
14421 (input_overwrite_node): Update.
14422 (input_node): Update.
14423 (input_varpool_node): Update.
14424 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14425 (lto_symtab_merge_symbols): Add sanity check.
14426 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14427 * passes.c (rest_of_decl_compilation): Set static flag, too.
14428 * symtab.c (dump_symtab_base): Dump weakref.
14429 (verify_symtab_base): Sanity check weakrefs.
14430 (symtab_make_decl_local): Remove duplicated code.
14431 (symtab_alias_ultimate_target): Simplify.
14432 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 14433
2e6a2f88
TV		 144342013-06-11 Tom de Vries <tom@codesourcery.com>
14435
14436 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14437 sequence_vect == NULL.
14438
d9af3652
DD		 144392013-06-11 DJ Delorie <dj@redhat.com>
14440
14441 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14442 (rl78_unwind_word_mode): New.
14443
ffa66012
DM		 144442013-06-11 David Malcolm <dmalcolm@redhat.com>
14445
14446 * final.c (debug_prefix_maps): Make static.
14447
783f2d35
DM		 144482013-06-11 David Malcolm <dmalcolm@redhat.com>
14449
14450 * function.c (initial_trampoline): Remove stray copy.
14451
7af26def
SN		 144522013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14453
14454 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14455
97756c0e
MJ		 144562013-06-11 Martin Jambor <mjambor@suse.cz>
14457
14458 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14459 within bounds at the beginning of the function.
14460
2c7b8bf6
AM		 144612013-06-11 Alan Modra <amodra@gmail.com>
14462
14463 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14464 reporting.
14465 (get_named_section): Don't NULL !DECL_P decl.
14466
77cea46e
IZ		 144672013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14468
14469 * doc/invoke.texi (core-avx2): Document.
14470 (slm): Likewise.
14471 (atom): Updated with MOVBE.
14472
888f0920
RB		 144732013-06-11 Richard Biener <rguenther@suse.de>
14474
22fc7d3d 14475 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14476
6b438d58
AB		 144772013-06-11 Anton Blanchard <anton@samba.org>
14478
14479 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14480 correct shift value in little-endian mode.
14481
3bc66938
JJ		 144822013-06-11 Jakub Jelinek <jakub@redhat.com>
14483
14484 PR target/56564
14485 * varasm.c (get_variable_align): Move #endif to the right place.
14486
a5dc7f88
CC		 144872013-06-10 Cary Coutant <ccoutant@google.com>
14488
14489 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14490 for hash so that hash table traversal order is deterministic.
14491
19be72ab 144922013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14493 Pat Haugen <pthaugen@us.ibm.com>
14494 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 14495
14496 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14497 of vectors in GPRS if they are direct moves or quad word load or
14498 store moves.
14499
14500 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14501 declaration.
14502 (direct_move_p): Likewise.
14503 (quad_load_store_p): Likewise.
14504
14505 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14506 classes into bins based on the physical register type.
14507 (reg_class_to_reg_type): Likewise.
14508 (IS_STD_REG_TYPE): Likewise.
14509 (IS_FP_VECT_REG_TYPE): Likewise.
14510 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14511 use direct move instructions.
14512 (reload_gpr_vsx): Likewise.
14513 (reload_vsx_gpr): Likewise.
14514 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14515 information that is a simplification of register classes. Also
14516 precalculate direct move reload helpers.
14517 (direct_move_p): New function to return true if the operation can
14518 be done as a direct move instruciton.
14519 (quad_load_store_p): New function to return true if the operation
14520 is a quad memory operation.
14521 (rs6000_legitimize_address): If quad memory, only allow register
14522 indirect for TImode addresses.
14523 (rs6000_legitimate_address_p): Likewise.
14524 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14525 (rs6000_reload_register_type): Likewise.
14526 (register_to_reg_type): Return register type.
14527 (rs6000_secondary_reload_simple_move): New helper function for
14528 secondary reload and secondary memory needed to identify anything
14529 that is a simple move, and does not need reloading.
14530 (rs6000_secondary_reload_direct_move): New helper function for
14531 secondary reload to identify cases that can be done with several
14532 instructions via the direct move instructions.
14533 (rs6000_secondary_reload_move): New helper function for secondary
14534 reload to identify moves between register types that can be done.
14535 (rs6000_secondary_reload): Add support for quad memory operations
14536 and for direct move.
14537 (rs6000_secondary_memory_needed): Likewise.
14538 (rs6000_debug_secondary_memory_needed): Change argument names.
14539 (rs6000_output_move_128bit): New function to return the move to
14540 use for 128-bit moves, including knowing about the various
14541 limitations of quad memory operations.
14542
14543 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14544 memory operations. call rs6000_output_move_128bit for the actual
14545 instruciton(s) to generate.
14546 (vsx_movti_64bit): Likewise.
14547
14548 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14549 (UNSPEC_P8V_MTVSRWZ): Likewise.
14550 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14551 (UNSPEC_P8V_MTVSRD): Likewise.
14552 (UNSPEC_P8V_XXPERMDI): Likewise.
14553 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14554 (UNSPEC_FUSION_GPR): Likewise.
14555 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14556 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 14557 (f32_sv): Likewise.
14558 (f32_dm): Likewise.
14559 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14560 loads and direct move instructions.
14561 (zero_extendsidi2_lfiwzx): Likewise.
14562 (extendsidi2_lfiwax): Likewise.
14563 (extendsidi2_nocell): Likewise.
14564 (floatsi<mode>2_lfiwax): Likewise.
14565 (lfiwax): Likewise.
14566 (floatunssi<mode>2_lfiwzx): Likewise.
14567 (lfiwzx): Likewise.
14568 (fix_trunc<mode>_stfiwx): Likewise.
14569 (fixuns_trunc<mode>_stfiwx): Likewise.
14570 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14571 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14572 (parity<mode>2_cmpb): Set length/type attr.
14573 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14574 for 'mr.' to fast_compare.
14575 (bpermd_<mode>): Change type attr to popcnt.
14576 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14577 (p8_mtvsrwz_1): Likewise.
14578 (p8_mtvsrwz_2): Likewise.
14579 (reload_fpr_from_gpr<mode>): Likewise.
14580 (p8_mtvsrd_1): Likewise.
14581 (p8_mtvsrd_2): Likewise.
14582 (p8_xxpermdi_<mode>): Likewise.
14583 (reload_vsx_from_gpr<mode>): Likewise.
14584 (reload_vsx_from_gprsf): Likewise.
14585 (p8_mfvsrd_3_<mode>): LIkewise.
14586 (reload_gpr_from_vsx<mode>): Likewise.
14587 (reload_gpr_from_vsxsf): Likewise.
14588 (p8_mfvsrd_4_disf): Likewise.
14589 (multi-word GPR splits): Do not split direct moves or quad memory
14590 operations.
14591
16876bdc
DM		 145922013-06-10 David Malcolm <dmalcolm@redhat.com>
14593
14594 * tree-into-ssa.c (interesting_blocks): Make static.
14595
df8a1d28
JJ		 145962013-06-10 Jakub Jelinek <jakub@redhat.com>
14597
14598 PR target/56564
14599 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14600 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14601 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14602 (get_variable_align): New function.
14603 (get_variable_section, emit_bss, emit_common,
14604 assemble_variable_contents, place_block_symbol): Use
14605 get_variable_align instead of DECL_ALIGN.
14606 (assemble_noswitch_variable): Add align argument, use it
14607 instead of DECL_ALIGN.
14608 (assemble_variable): Adjust caller. Use get_variable_align
14609 instead of DECL_ALIGN.
14610 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14611 caller.
14612 (DATA_ABI_ALIGNMENT): Define.
14613 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14614 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14615 opt is false, only return the psABI mandated alignment increase.
14616 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14617 (DATA_ABI_ALIGNMENT): ... this.
14618 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14619 (DATA_ABI_ALIGNMENT): ... this.
14620 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14621 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14622 (DATA_ABI_ALIGNMENT): ... this.
14623 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14624 * doc/tm.texi: Regenerated.
14625
57b29ca6
UB		 146262013-06-10 Uros Bizjak <ubizjak@gmail.com>
14627
14628 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14629 cmp_code to construct REG_EQUAL note.
14630
3e56ed50
JJ		 146312013-06-09 Jakub Jelinek <jakub@redhat.com>
14632
14633 PR target/57568
14634 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14635 that operands[2] doesn't overlap with operands[0].
14636
74fe2a1d
DE		 146372013-06-09 David Edelsohn <dje.gcc@gmail.com>
14638 Jan Hubicka <jh@suse.cz>
14639
14640 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14641 hack to mark symbols as used.
14642
ed52a84e
VM		 146432013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14644
14645 PR rtl-optimization/57559
14646 * lra-constraints.c (process_alt_operands): Don't discourage
14647 memory with known offset for offsetable memory constraint.
14648 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14649
299316ed
EB		 146502013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14651
14652 * varasm.c (struct oc_local_state): Reorder fields.
14653 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14654 and adjust accordingly.
14655 (output_constructor): Reorder initialization code and adjust call to
14656 output_constructor_bitfield.
14657
88ac60d3
JH		 146582013-06-07 Jan Hubicka <jh@suse.cz>
14659
14660 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14661
4cdce1a8
DM		 146622013-06-07 David Malcolm <dmalcolm@redhat.com>
14663
14664 * tree-object-size.c (unknown): Make const.
14665
b5e0425c
AK		 146662013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14667
14668 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14669 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14670 for last alternative in the cpu_facility attribute.
14671
79678d04
KT		 146722013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14673
14674 PR target/56315
14675 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14676 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14677 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14678 * config/arm/constraints.md (Dg): New constraint.
14679 * config/arm/neon.md (xordi3_neon): Remove.
14680 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14681 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14682
b31ddbdb
KT		 146832013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14684
14685 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14686 Clean up alternatives.
14687
8f90b7d4
AM		 146882013-06-07 Alan Modra <amodra@gmail.com>
14689
14690 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14691 va_list_gpr_size.
14692
961ce119
AM		 146932013-06-07 Alan Modra <amodra@gmail.com>
14694
14695 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14696
33e49835
KT		 146972013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14698
14699 * config/arm/constraints.md (Df): New constraint.
14700 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14701 Correct length attribute for last two alternatives.
14702
acd17ae6
AM		 147032013-06-07 Alan Modra <amodra@gmail.com>
14704
14705 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14706 override user -mfp-in-toc.
14707 (offsettable_ok_by_alignment): Consider just the current access
14708 rather than the whole object, unless BLKmode. Handle
14709 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14710 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14711 for -mcmodel=medium.
14712 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14713 override user -mfp-in-toc or -msum-in-toc. Default to
14714 -mno-fp-in-toc for -mcmodel=medium.
14715
73310b0e
DD		 147162013-06-06 DJ Delorie <dj@redhat.com>
14717
14718 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14719 TARGET_VALID_POINTER_MODE.
14720
0bd62dca 147212013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14722 Pat Haugen <pthaugen@us.ibm.com>
14723 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 14724
14725 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14726 Document new power8 builtins.
14727
14728 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14729 condition code register, to allow 128-bit logical operations to be
14730 done in the VSX or GPR registers.
14731 (nor<mode>3): Use the canonical form for nor.
14732 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14733 vclz*, and vpopcnt* vector instructions.
14734 (nand<mode>3): Likewise.
14735 (orc<mode>3): Likewise.
14736 (clz<mode>2): LIkewise.
14737 (popcount<mode>2): Likewise.
14738
14739 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14740 that only the GPRs are recognized.
14741
14742 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14743 support for new power8 builtins.
14744
14745 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14746 builtin functions.
14747 (xscvdpspn): Likewise.
14748 (vclz): Likewise.
14749 (vclzb): Likewise.
14750 (vclzh): Likewise.
14751 (vclzw): Likewise.
14752 (vclzd): Likewise.
14753 (vpopcnt): Likewise.
14754 (vpopcntb): Likewise.
14755 (vpopcnth): Likewise.
14756 (vpopcntw): Likewise.
14757 (vpopcntd): Likewise.
14758 (vgbbd): Likewise.
14759 (vmrgew): Likewise.
14760 (vmrgow): Likewise.
14761 (eqv): Likewise.
14762 (eqv_v16qi3): Likewise.
14763 (eqv_v8hi3): Likewise.
14764 (eqv_v4si3): Likewise.
14765 (eqv_v2di3): Likewise.
14766 (eqv_v4sf3): Likewise.
14767 (eqv_v2df3): Likewise.
14768 (nand): Likewise.
14769 (nand_v16qi3): Likewise.
14770 (nand_v8hi3): Likewise.
14771 (nand_v4si3): Likewise.
14772 (nand_v2di3): Likewise.
14773 (nand_v4sf3): Likewise.
14774 (nand_v2df3): Likewise.
14775 (orc): Likewise.
14776 (orc_v16qi3): Likewise.
14777 (orc_v8hi3): Likewise.
14778 (orc_v4si3): Likewise.
14779 (orc_v2di3): Likewise.
14780 (orc_v4sf3): Likewise.
14781 (orc_v2df3): Likewise.
14782
14783 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14784 allow power8 quad mode in 64-bit.
14785 (rs6000_builtin_vectorized_function): Add support to vectorize
14786 ISA 2.07 count leading zeros, population count builtins.
14787 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14788 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14789 (builtin_function_type): Add vgbbd builtin function which takes an
14790 unsigned argument.
14791 (altivec_expand_vec_perm_const): Add support for new power8 merge
14792 instructions.
14793
14794 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14795 that does not include TImdoe for use with 32-bit.
14796 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14797 instructions.
14798 (UNSPEC_VSX_CVDPSPN): Likewise.
14799 (vsx_xscvdpspn): Likewise.
14800 (vsx_xscvspdpn): Likewise.
14801 (vsx_xscvdpspn_scalar): Likewise.
14802 (vsx_xscvspdpn_directmove): Likewise.
14803 (vsx_and<mode>3): Split logical operations into 32-bit and
14804 64-bit. Add support to do logical operations on TImode as well as
14805 VSX vector types. Allow logical operations to be done in either
14806 VSX registers or in general purpose registers in 64-bit mode. Add
14807 splitters if GPRs were used. For AND, add clobber of CCmode to
14808 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14809 encoding.
14810 (vsx_and<mode>3_32bit): Likewise.
14811 (vsx_and<mode>3_64bit): Likewise.
14812 (vsx_ior<mode>3): Likewise.
14813 (vsx_ior<mode>3_32bit): Likewise.
14814 (vsx_ior<mode>3_64bit): Likewise.
14815 (vsx_xor<mode>3): Likewise.
14816 (vsx_xor<mode>3_32bit): Likewise.
14817 (vsx_xor<mode>3_64bit): Likewise.
14818 (vsx_one_cmpl<mode>2): Likewise.
14819 (vsx_one_cmpl<mode>2_32bit): Likewise.
14820 (vsx_one_cmpl<mode>2_64bit): Likewise.
14821 (vsx_nor<mode>3): Likewise.
14822 (vsx_nor<mode>3_32bit): Likewise.
14823 (vsx_nor<mode>3_64bit): Likewise.
14824 (vsx_andc<mode>3): Likewise.
14825 (vsx_andc<mode>3_32bit): Likewise.
14826 (vsx_andc<mode>3_64bit): Likewise.
14827 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14828 and xxlorc instructions.
14829 (vsx_eqv<mode>3_64bit): Likewise.
14830 (vsx_nand<mode>3_32bit): Likewise.
14831 (vsx_nand<mode>3_64bit): Likewise.
14832 (vsx_orc<mode>3_32bit): Likewise.
14833 (vsx_orc<mode>3_64bit): Likewise.
14834
14835 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14836
14837 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14838 instruction.
14839 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14840 (p8_vmrgow): Likewise.
14841 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14842 GPRs to be split under VSX.
14843 (p8v_clz<mode>2): Add power8 count leading zero support.
14844 (p8v_popcount<mode>2): Add power8 population count support.
14845 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14846 support.
14847
14848 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14849 instruction.
14850
14851 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14852 builtin functions.
14853 (vec_nand): Likewise.
14854 (vec_vclz): Likewise.
14855 (vec_vclzb): Likewise.
14856 (vec_vclzd): Likewise.
14857 (vec_vclzh): Likewise.
14858 (vec_vclzw): Likewise.
14859 (vec_vgbbd): Likewise.
14860 (vec_vmrgew): Likewise.
14861 (vec_vmrgow): Likewise.
14862 (vec_vpopcnt): Likewise.
14863 (vec_vpopcntb): Likewise.
14864 (vec_vpopcntd): Likewise.
14865 (vec_vpopcnth): Likewise.
14866 (vec_vpopcntw): Likewise.
14867
37684c46
VM		 148682013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14869
14870 PR rtl-optimization/57468
14871 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14872 spilled pseudos.
14873
d89ae27c
VM		 148742013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14875
14876 PR rtl-optimization/57459
14877 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14878 type when setting live regs.
14879
3597e113
VM		 148802013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14881
14882 * config/s390/s390.opt (mlra): New option.
14883 * config/s390/s390.c (s390_decompose_address): Check displacement
14884 for all registers for LRA.
14885 (s390_secondary_reload): Don't used secondary reloads for LRA.
14886 (s390_lra_p): New function.
14887 (TARGET_LRA_P): Define.
14888 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14889 of attribute cpu_facility to zarch for the last alternative.
14890 (*cmpmem_short): Ditto.
14891
01037aeb
EB		 148922013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14893
14894 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14895 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14896 (arm_expand_prologue): Likewise.
14897
3371a64f
TJ		 148982013-06-06 Teresa Johnson <tejohnson@google.com>
14899
14900 PR c++/53743
14901 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14902 as this is now done by redirect_edge_and_branch_force.
14903 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14904 barriers, and fix interaction with splitting.
14905 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14906 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14907 reflect changes made in the routine.
14908 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14909 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14910 since this is called in cfglayout mode, and replace partition fixup
14911 with assert as that is now done by force_nonfallthru_and_redirect.
14912 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14913 already be marked with region crossing note.
14914 (insert_section_boundary_note): Make non-static, gate on flag
14915 has_bb_partition, rewrite to also check for multiple partitions.
14916 (rest_of_handle_reorder_blocks): Remove call to
14917 insert_section_boundary_note, now done later during free_cfg.
14918 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14919 * bb-reorder.h (insert_section_boundary_note): Declare.
14920 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14921 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14922 invoke insert_section_boundary_note.
14923 (try_redirect_by_replacing_jump): Remove unnecessary
14924 check for region crossing note.
14925 (fixup_partition_crossing): New function.
14926 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14927 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14928 in non-cfglayout mode.
14929 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14930 remove old code that tried to do this. Emit barrier correctly
14931 when we are in cfglayout mode.
14932 (last_bb_in_partition): New function.
14933 (rtl_split_edge): Correctly fixup partition boundaries.
14934 (commit_one_edge_insertion): Remove old code that tried to
14935 fixup region crossing edge since this is now handled in
14936 split_block, and set up insertion point correctly since
14937 block may now end in a jump.
14938 (verify_hot_cold_block_grouping): Guard against checking when not in
14939 linearized RTL mode.
14940 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14941 notes.
14942 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14943 rtl_verify_flow_info, so not called in cfglayout mode.
14944 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14945 (fixup_reorder_chain): Remove old code that attempted to fixup region
14946 crossing note as this is now handled in force_nonfallthru_and_redirect.
14947 (duplicate_insn_chain): Don't duplicate switch section notes.
14948 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14949 note.
14950 * basic-block.h (emit_barrier_after_bb): Declare.
14951
66071e10
KT		 149522013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14953
14954 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14955 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14956 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14957
0e26bf3d
KT		 149582013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14959
14960 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14961 where appropriate.
14962 * config/arm/ldmstm.md: Regenerate.
14963
12b4e7ef
KT		 149642013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14965
14966 * config/arm/sync.md (atomic_loaddi_1):
14967 Disable predication for arm_restrict_it.
14968 (arm_load_exclusive<mode>): Likewise.
14969 (arm_load_exclusivesi): Likewise.
14970 (arm_load_exclusivedi): Likewise.
14971 (arm_load_acquire_exclusive<mode>): Likewise.
14972 (arm_load_acquire_exclusivesi): Likewise.
14973 (arm_load_acquire_exclusivedi): Likewise.
14974 (arm_store_exclusive<mode>): Likewise.
14975 (arm_store_exclusive<mode>): Likewise.
14976 (arm_store_release_exclusivedi): Likewise.
14977 (arm_store_release_exclusive<mode>): Likewise.
14978
15d16c8a
RB		 149792013-06-06 Richard Biener <rguenther@suse.de>
14980
14981 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14982 after LTO_null.
14983 (lto_tag_is_tree_code_p): Adjust.
14984 (lto_tag_is_gimple_code_p): Likewise.
14985 (lto_gimple_code_to_tag): Likewise.
14986 (lto_tag_to_gimple_code): Likewise.
14987 (lto_tree_code_to_tag): Likewise.
14988 (lto_tag_to_tree_code): Likewise.
14989 * data-streamer.h (streamer_write_hwi_in_range): Use
14990 uhwi streaming to stream the normalized range.
14991 (streamer_read_hwi_in_range): Likewise.
14992
17a7fc37
KT		 149932013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14994
14995 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14996 (predicable_short_it): Likewise.
14997 (predicated): Likewise.
14998 (enabled): Handle above.
14999 (define_cond_exec): Set predicated attribute to yes.
15000
b57ca59b
MS		 150012013-06-05 Mike Stump <mikestump@comcast.net>
15002
15003 * gdbinit.in (__FUNCTION__): Add.
15004
c1e183a9
UB		 150052013-06-05 Uros Bizjak <ubizjak@gmail.com>
15006
15007 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15008 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15009
0378bddb
JH		 150102013-06-05 Jan Hubicka <jh@suse.cz>
15011
15012 * varasm.c (mark_decl_referenced): Revert the removal until targets
15013 are fixed.
15014
8f8a2057
DE		 150152013-06-05 David Edelsohn <dje.gcc@gmail.com>
15016
15017 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15018 instead of mark_decl_referenced.
15019
edb983b2
JH		 150202013-06-05 Jan Hubicka <jh@suse.cz>
15021
15022 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15023 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15024 and symtab_used_from_object_file_p.
15025 (cgraph_make_node_local_1): Clear forced_by_abi.
15026 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15027 * cgraph.h (symtab_node_base): Add forced_by_abi.
15028 (decide_is_variable_needed): Remove.
15029 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15030 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15031 (decide_is_symbol_needed): ... this one; handle symbols in general;
15032 always analyze virtuals; honnor forced_by_abi.
15033 (cgraph_finalize_function): Update.
15034 (varpool_finalize_decl): Update.
15035 (symbol_defined_and_needed): Remove.
15036 (analyze_functions): Update.
15037 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15038 output_refs, input_overwrite_node): Handle forced_by_abi.
15039 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15040 (address_taken_from_non_vtable_p): ... this one.
15041 (comdat_can_be_unshared_p_1): New function.
15042 (cgraph_comdat_can_be_unshared_p): Rename to ...
15043 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15044 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15045 (function_and_variable_visibility): Clear forced_by_abi as needed.
15046 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15047 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 15048 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15049 * varpool.c (decide_is_variable_needed): Remove.
15050
9912dbe5
KT		 150512013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15052
15053 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15054 (arm_option_override): Override arm_restrict_it where appropriate.
15055 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15056 * config/arm/arm.opt (mrestrict-it): New command-line option.
15057 * doc/invoke.texi: Document -mrestrict-it.
15058
34c136b6
DM		 150592013-06-05 David Malcolm <dmalcolm@redhat.com>
15060
15061 * tsan.c (tsan_atomic_table): Make const.
15062
a367df53
RB		 150632013-06-05 Richard Biener <rguenther@suse.de>
15064
15065 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15066 index associated with the tree we are supposed to replace.
15067 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15068 TREE_ASM_WRITTEN as zero for everything but SSA names.
15069
70d51a19
DM		 150702013-06-05 David Malcolm <dmalcolm@redhat.com>
15071
15072 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15073
b9bd2075
JH		 150742013-06-04 Jan Hubicka <jh@suse.cz>
15075
15076 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15077 (input_node, input_varpool_node): Handle correctly external same
15078 body aliases.
15079 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15080 nodes at ltrans stage.
15081
aaae719d
JH		 150822013-06-04 Jan Hubicka <jh@suse.cz>
15083
15084 * ipa-inline.c (update_caller_keys): Fix availability test.
15085 (update_callee_keys): Likewise.
15086 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15087 to follow ELF standard.
15088
107eea2c
JU		 150892013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15090
15091 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15092 (mips64r5900el-*-elf*): New configurations.
15093 * config/mips/mips-cpus.def (r5900): New processor.
15094 * config/mips/mips-tables.opt: Regenerate.
15095 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15096 (mips_issue_rate): Handle PROCESSOR_R5900.
15097 (mips_reorg_process_insns): Force reorder mode for the R5900.
15098 * config/mips/mips.h (TARGET_MIPS5900): Define.
15099 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15100 TARGET_MIPS5900.
15101 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15102 TARGET_MIPS5900.
15103 * config/mips/mips.md (processor): Add r5900.
15104 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15105
b7342d25
IB		 151062013-06-04 Ian Bolton <ian.bolton@arm.com>
15107
15108 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15109 into function to generate MOVI instruction.
c1e183a9 15110 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 15111 (aarch64_preferred_simd_mode): Turn into wrapper.
15112 (aarch64_output_scalar_simd_mov_immediate): New function.
15113 * config/aarch64/aarch64-protos.h: Add prototype for above.
15114
81c2dfb9
IB		 151152013-06-04 Ian Bolton <ian.bolton@arm.com>
15116
15117 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15118 element_char member.
15119 (sizetochar): Return signed char.
15120 (aarch64_simd_valid_immediate): Remove elchar and other
15121 unnecessary variables.
15122 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15123 Calculate element_char as required.
15124 * config/aarch64/aarch64-protos.h: Update and move prototype
15125 for aarch64_output_simd_mov_immediate.
15126 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15127 Update arguments.
15128
48063b9d
IB		 151292013-06-04 Ian Bolton <ian.bolton@arm.com>
15130
15131 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15132 information completed by aarch64_simd_valid_immediate.
15133 (aarch64_legitimate_constant_p): Update arguments.
15134 (aarch64_simd_valid_immediate): Work with struct rather than many
15135 pointers.
15136 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15137 (aarch64_simd_make_constant): Update arguments.
15138 (aarch64_output_simd_mov_immediate): Work with struct rather than
15139 many pointers. Output immediate directly rather than as operand.
15140 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15141 Update prototype.
15142 * config/aarch64/constraints.md (Dn): Update arguments.
15143
3ea63f60
IB		 151442013-06-04 Ian Bolton <ian.bolton@arm.com>
15145
15146 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15147 longer static.
15148 (aarch64_simd_immediate_valid_for_move): Remove.
15149 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15150 (aarch64_simd_make_constant): Update call.
15151 (aarch64_output_simd_mov_immediate): Update call.
15152 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15153 Add prototype.
15154 * config/aarch64/constraints.md (Dn): Update call.
15155
d8edd899
IB		 151562013-06-04 Ian Bolton <ian.bolton@arm.com>
15157
15158 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15159 return type to bool for prototype.
15160 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15161 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15162 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15163
44b20bb8
CM		 151642013-06-04 Catherine Moore <clm@codesourcery.com>
15165
15166 * config/mips/mips.opt (meva): New.
c1e183a9 15167 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 15168 (ASM_SPEC): Handle -meva.
15169 * doc/invoke.texi (meva): Document.
15170
52befbd8
AM		 151712013-06-04 Alan Modra <amodra@gmail.com>
15172
15173 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15174 constant output.
15175
aadaf24e
KT		 151762013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15177
15178 * rtl.def: Add extra fourth optional field to define_cond_exec.
15179 * gensupport.c (process_one_cond_exec): Process attributes from
15180 define_cond_exec.
15181 * doc/md.texi: Document fourth field in define_cond_exec.
15182
0cd9e9ee
EB		 151832013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15184
15185 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15186 out the processing order as in store_bit_field_1.
15187
cb0f665d
JH		 151882013-06-04 Jan Hubicka <jh@suse.cz>
15189
15190 PR middle-end/57500
15191 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15192 non-VAR_DECL node if it does not exist yet.
15193
53984b9b
RS		 151942013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15195
15196 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15197 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15198 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15199 target_cpu_default setting.
15200
5d318fd4
TJ		 152012013-06-03 Teresa Johnson <tejohnson@google.com>
15202
15203 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15204 default to -fopt-info=optimized instead of all.
15205 * doc/invoke.texi: Ditto.
15206 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15207 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15208 (execute_vect_slp): Emit BB vectorization success under
15209 MSG_OPTIMIZED_LOCATIONS.
15210 * tree-vect-slp.c (vect_slp_transform_bb): Change
15211 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15212 * tree-vect-loop.c (vect_transform_loop): Ditto.
15213
ec3af349
JM		 152142013-06-03 Jason Merrill <jason@redhat.com>
15215
15216 PR c++/57415
15217 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15218 Use TARGET_EXPR for C++.
15219
1baf9159
JJ		 152202013-06-03 Jakub Jelinek <jakub@redhat.com>
15221
15222 PR rtl-optimization/57268
0cd9e9ee 15223 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 15224 if DEBUG_INSN_P (insn).
15225
15226 Reapply
15227 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15228
15229 PR rtl-optimization/57268
15230 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15231 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15232
0f1d3965
YR		 152332013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15234
15235 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15236 (ix86_avoid_lea_for_addr): Likewise.
15237 (exact_dependency_1): Likewise.
15238 (ix86_adjust_cost): Likewise.
15239 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15240 removed.
15241 (do_reorder_for_imul): Fix typo, formatting and
15242 !reload_completed check removed.
15243 (ix86_sched_reorder): Fix typo and formatting.
15244 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15245 list.
15246
3bc30db4
SN		 152472013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15248
15249 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15250
19cc6fac
EB		 152512013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15252
15253 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15254 <STRING_CST>: Likewise.
15255 <VECTOR_CST>: Likewise.
15256
c68a6e08
JW		 152572013-06-01 Janus Weil <janus@gcc.gnu.org>
15258 Mikael Morin <mikael@gcc.gnu.org>
15259
15260 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15261 * config.in: Regenerated.
15262 * configure: Regenerated.
15263
38e55e5c
JH		 152642013-06-01 Jan Hubicka <jh@suse.cz>
15265
15266 PR middle-end/57366
15267 * cgraphunit.c (compile): When weakref is not supported,
15268 set up transparent aliases before final output pass.
15269 * varasm.c (assemble_alias): Do not try to do it here.
15270
eb51d2ff
JH		 152712013-06-01 Jan Hubicka <jh@suse.cz>
15272
15273 PR middle-end/57467
15274 * passes.c (for_per_function): Skip unanalyzed functions.
15275
40a7fe1e
JH		 152762013-06-01 Jan Hubicka <jh@suse.cz>
15277
15278 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15279 (lto_symtab_merge_symbols_1): ... this one.
15280 (lto_symtab_merge_cgraph_nodes): Rename to ...
15281 (lto_symtab_merge_symbols): ... this one; simplify.
15282 * cgraph.c (same_body_aliases_done): Rename to ...
15283 (cpp_implicit_aliases_done): ... this one.
15284 (cgraph_create_function_alias): Update.
15285 (cgraph_same_body_alias): Update.
c1e183a9 15286 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 15287 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 15288 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 15289 (cgraph_node): Remove same_body_alias.
15290 (varpool_node): Remove alias_of and extra_name_alias.
15291 (same_body_aliases_done): Rename to ..
15292 (cpp_implicit_aliases_done): ... this one.
15293 (symtab_alias_ultimate_target): Add default parameter.
15294 (symtab_resolve_alias): New function.
15295 (fixup_same_cpp_alias_visibility): Declare.
15296 (cgraph_function_node): Add default parameter.
15297 (cgraph_node_asm_name): Likewise.
15298 (cgraph_function_or_thunk_node): Add default parameter; do
15299 not ICE when it is NULL.
15300 (varpool_variable_node): Likewise.
15301 * tree-emutls.c (create_emultls_var): Update.
15302 (ipa_lower_emutls): Update.
15303 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15304 (cgraph_reset_node): Reset alias info.
15305 (cgraph_finalize_function): Update.
15306 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15307 (analyze_function): Simplify.
15308 (cgraph_process_same_body_aliases): Simplify.
15309 (analyze_functions): Fixup same body aliases.
15310 (handle_alias_pairs): Simplify.
15311 (assemble_thunk): Update.
15312 (assemble_thunks_and_aliases): Update.
15313 (output_weakrefs): Rewrite.
15314 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15315 (lto_output_varpool_node): Likewise.
15316 (compute_ltrans_boundary): Remve assert.
15317 (get_alias_symbol): New functoin.
15318 (input_node): Rewrite alias handling.
15319 (input_varpool_node): Likewise.
15320 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15321 * ipa.c (process_references): Handle weakrefs correctly.
15322 (symtab_remove_unreachable_nodes): Likewise.
15323 * trans-mem.c (get_cg_data): Update.
15324 (ipa_tm_create_version_alias): Update.
15325 (ipa_tm_execute): Update.
15326 * symtab.c (dump_symtab_base): Dump aliases.
15327 (verify_symtab_base): Verify aliases.
15328 (symtab_node_availability): New function.
15329 (symtab_alias_ultimate_target): Simplify.
15330 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15331 handle all the fixup cases.
15332 (symtab_resolve_alias): New function.
15333 * passes.c (ipa_write_summaries): Handle weakrefs.
15334 * varpool.c (varpool_analyze_node): Simplify.
15335 (assemble_aliases): Update.
15336 (varpool_create_variable_alias): Simplify.
15337 (varpool_extra_name_alias): Simplify.
15338 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15339 (lto_symtab_merge_symbols): ... this one.
15340
21759881
DT		 153412013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15342
15343 Revert
15344 PR rtl-optimization/57268
15345 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15346 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15347
daf4e940
TB		 153482013-06-01 Tobias Burnus <burnus@net-b.de>
15349
15350 Partially reverted:
15351 2013-05-31 Tobias Burnus <burnus@net-b.de>
15352
15353 PR middle-end/57073
15354 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15355 further up.
15356
c359d8d0
DT		 153572013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15358
15359 PR rtl-optimization/57268
15360 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15361 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15362
4172215d
EB		 153632013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15364
15365 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15366 unordered comparison operators when -fno-trapping-math is in effect
15367 on the e500.
15368 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15369 and implement unordered comparison operators properly on the e500.
15370
a8c50132
EB		 153712013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15372
15373 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15374 for constant scalar integers.
15375 (simplify_relational_operation_1): Likewise.
15376
44626634
SB		 153772013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15378
15379 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15380 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15381 Fix comment.
15382
55805e54
YR		 153832013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15384 Igor Zamyatin <igor.zamyatin@intel.com>
15385
15386 Silvermont (SLM) architecture performance tuning.
15387 * config/i386/i386.h (enum ix86_tune_indices): Add
15388 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15389 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15390
15391 * config/i386/i386.c (initial_ix86_tune_features)
15392 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15393 (ix86_lea_outperforms): Handle Silvermont tuning.
15394 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15395 call.
15396 (ix86_use_lea_for_mov): Likewise.
15397 (ix86_avoid_lea_for_addr): Likewise.
15398 (ix86_lea_for_add_ok): Likewise.
15399 (exact_dependency_1): New function.
15400 (exact_store_load_dependency): Likewise.
15401 (ix86_adjust_cost): Handle Silvermont tuning.
15402 (do_reoder_for_imul): Likewise.
15403 (swap_top_of_ready_list): New function.
15404 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15405
15406 * config/i386/i386.md (peepholes that split memory operand in fp
15407 converts): New.
15408
e19c9de2
MS		 154092013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15410
15411 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15412 Remove un-necessary braces.
15413
38e6c9a6
MS		 154142013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15415
15416 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15417 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15418
e3530904
TB		 154192013-05-31 Tobias Burnus <burnus@net-b.de>
15420
15421 PR middle-end/57073
c1e183a9 15422 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 15423
b6af05a9
KT		 154242013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15425
15426 PR target/56315
15427 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15428 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15429 * config/arm/neon.md (iordi3_neon): Remove.
15430 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15431 * config/arm/predicates.md (imm_for_neon_logic_operand):
15432 Move to earlier in the file.
15433 (neon_logic_op2): Likewise.
15434 (arm_iordi_operand_neon): New predicate.
15435
f800c166
RB		 154362013-05-31 Richard Biener <rguenther@suse.de>
15437
15438 PR tree-optimization/57478
15439 PR tree-optimization/57453
15440 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15441 are life as well.
15442
e571fa59
KP		 154432013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15444
15445 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15446 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15447
0fa6e0ef
TB		 154482013-05-30 Tobias Burnus <burnus@net-b.de>
15449 Thomas Koenig <tkoenig@gcc.gnu.org>
15450
15451 PR middle-end/57073
15452 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15453 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15454
4688ddf5
SB		 154552013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15456
0fa6e0ef 15457 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15458
3b9ceb4b
VM		 154592013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15460
15461 * target.def (register_usage_leveling_p): New hook.
15462 * targhooks.c (default_register_usage_leveling_p): New.
15463 * targhooks.h (default_register_usage_leveling_p): New prototype.
15464 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15465 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15466 * doc/tm.texi: Update.
15467 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15468
65f9e789
IB		 154692013-05-30 Ian Bolton <ian.bolton@arm.com>
15470
15471 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15472 (*insv_reg<mode>): New define_insn.
15473
12211b99 154742013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 15475
15476 PR rtl-optimization/57439
15477 * postreload.c (move2add_valid_value_p): Check that we have
15478 a zero subreg_regno_offset when accessing the register in
15479 the requested mode.
15480
0b871ccf
YR		 154812013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15482 Igor Zamyatin <igor.zamyatin@intel.com>
15483
15484 Silvermont (SLM) architecture pipeline model, tuning and
15485 insn selection.
15486 * config.gcc: Add slm config options and target.
15487
15488 * config/i386/slm.md: New.
15489
15490 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15491
f43245d1 15492 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR		 15493 PROCESSOR_SLM.
15494 (ix86_target_macros_internal): Likewise.
15495
f43245d1 15496 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR		 15497 (m_SLM): New macro flag.
15498 (initial_ix86_tune_features): Set m_SLM.
15499 (x86_accumulate_outgoing_args): Likewise.
15500 (x86_arch_always_fancy_math_387): Likewise.
15501 (processor_target_table): Add slm cost.
15502 (cpu_names): Add slm cpu name.
15503 (x86_option_override_internal): Set SLM ISA.
15504 (ix86_issue_rate): New case PROCESSOR_SLM.
15505 (ia32_multipass_dfa_lookahead): Likewise.
15506 (fold_builtin_cpu): Add slm.
15507
15508 * config/i386/i386.h (TARGET_SLM): New target macro.
15509 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15510 (processor_type): Add PROCESSOR_SLM.
15511
15512 * config/i386/i386.md (cpu): Add new value "slm".
15513 (slm.md): Include slm.md.
15514
24d5b097
XG		 155152013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15516 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15517
15518 * config/arm/arm-protos.h: Add and update function protos.
15519 * config/arm/arm.c (use_simple_return_p): New added.
15520 (thumb2_expand_return): Check simple_return flag.
15521 * config/arm/arm.md: Add simple_return and conditional simple_return.
15522 * config/arm/iterators.md: Add iterator for return and simple_return.
15523
c1cccc15
ZC		 155242013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15525
15526 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15527 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15528 (arm_emit_vfp_multi_reg_pop): Likewise.
15529 (thumb2_emit_ldrd_pop): Likewise.
15530 (arm_expand_epilogue): Add misc REG_CFA notes.
15531 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15532
3a4f280b
LC		 155332013-05-29 Lawrence Crowl <crowl@google.com>
15534
15535 * config/arm/t-arm: Update for below.
15536
15537 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15538 Change type to hash_table. Update dependent calls and types.
15539
15540 * config/i386/t-cygming: Update for below.
15541
15542 * config/i386/t-interix: Update for below.
15543
15544 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15545 Change type to hash_table. Update dependent calls and types.
15546 (i386_find_on_wrapper_list::wrappers): Likewise.
15547
15548 * config/ia64/t-ia64: Update for below.
15549
15550 * config/ia64/ia64.c (bundle_state_table):
15551 Change type to hash_table. Update dependent calls and types.
15552
15553 * config/mips/mips.c (mips_reorg_process_insns::htab):
15554 Change type to hash_table. Update dependent calls and types.
15555
15556 * config/sol2.c (solaris_comdat_htab):
15557 Change type to hash_table. Update dependent calls and types.
15558
15559 * config/t-sol2: Update for above.
15560
1388a0e3
TJ		 155612013-05-29 Teresa Johnson <tejohnson@google.com>
15562
15563 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15564 functions are not yet marked as defined.
15565
a5965b52 155662013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15567 Pat Haugen <pthaugen@us.ibm.com>
15568 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 15569
15570 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15571 instructions.
15572 (VEC_A): Likewise.
15573 (VEC_C): Likewise.
15574 (vrotl<mode>3): Likewise.
15575 (vashl<mode>3): Likewise.
15576 (vlshr<mode>3): Likewise.
15577 (vashr<mode>3): Likewise.
15578
15579 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15580 support for power8 V2DI builtins.
15581
15582 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15583 power8 V2DI builtins.
15584 (vupkhsw): Likewise.
15585 (vupklsw): Likewise.
15586 (vaddudm): Likewise.
15587 (vminsd): Likewise.
15588 (vmaxsd): Likewise.
15589 (vminud): Likewise.
15590 (vmaxud): Likewise.
15591 (vpkudum): Likewise.
15592 (vpksdss): Likewise.
15593 (vpkudus): Likewise.
15594 (vpksdus): Likewise.
15595 (vrld): Likewise.
15596 (vsld): Likewise.
15597 (vsrd): Likewise.
15598 (vsrad): Likewise.
15599 (vsubudm): Likewise.
15600 (vcmpequd): Likewise.
15601 (vcmpgtsd): Likewise.
15602 (vcmpgtud): Likewise.
15603 (vcmpequd_p): Likewise.
15604 (vcmpgtsd_p): Likewise.
15605 (vcmpgtud_p): Likewise.
15606 (vupkhsw): Likewise.
15607 (vupklsw): Likewise.
15608 (vaddudm): Likewise.
15609 (vmaxsd): Likewise.
15610 (vmaxud): Likewise.
15611 (vminsd): Likewise.
15612 (vminud): Likewise.
15613 (vpksdss): Likewise.
15614 (vpksdus): Likewise.
15615 (vpkudum): Likewise.
15616 (vpkudus): Likewise.
15617 (vrld): Likewise.
15618 (vsld): Likewise.
15619 (vsrad): Likewise.
15620 (vsrd): Likewise.
15621 (vsubudm): Likewise.
15622
15623 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15624 support for power8 V2DI instructions.
15625
15626 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15627 power8 V2DI instructions. Combine pack and unpack insns to use an
15628 iterator for each mode. Check whether a particular mode supports
15629 Altivec instructions instead of just checking TARGET_ALTIVEC.
15630 (UNSPEC_VPKUWUM): Likewise.
15631 (UNSPEC_VPKSHSS): Likewise.
15632 (UNSPEC_VPKSWSS): Likewise.
15633 (UNSPEC_VPKUHUS): Likewise.
15634 (UNSPEC_VPKSHUS): Likewise.
15635 (UNSPEC_VPKUWUS): Likewise.
15636 (UNSPEC_VPKSWUS): Likewise.
15637 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15638 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15639 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15640 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15641 (UNSPEC_VUPKHSB): Likewise.
15642 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15643 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15644 (UNSPEC_VUPKHSH): Likewise.
15645 (UNSPEC_VUPKLSB): Likewise.
15646 (UNSPEC_VUPKLSH): Likewise.
15647 (VI2): Likewise.
15648 (VI_char): Likewise.
15649 (VI_scalar): Likewise.
15650 (VI_unit): Likewise.
15651 (VP): Likewise.
15652 (VP_small): Likewise.
15653 (VP_small_lc): Likewise.
15654 (VU_char): Likewise.
15655 (add<mode>3): Likewise.
15656 (altivec_vaddcuw): Likewise.
15657 (altivec_vaddu<VI_char>s): Likewise.
15658 (altivec_vadds<VI_char>s): Likewise.
15659 (sub<mode>3): Likewise.
15660 (altivec_vsubcuw): Likewise.
15661 (altivec_vsubu<VI_char>s): Likewise.
15662 (altivec_vsubs<VI_char>s): Likewise.
15663 (altivec_vavgs<VI_char>): Likewise.
15664 (altivec_vcmpbfp): Likewise.
15665 (altivec_eq<mode>): Likewise.
15666 (altivec_gt<mode>): Likewise.
15667 (altivec_gtu<mode>): Likewise.
15668 (umax<mode>3): Likewise.
15669 (smax<mode>3): Likewise.
15670 (umin<mode>3): Likewise.
15671 (smin<mode>3): Likewise.
15672 (altivec_vpkuhum): Likewise.
15673 (altivec_vpkuwum): Likewise.
15674 (altivec_vpkshss): Likewise.
15675 (altivec_vpkswss): Likewise.
15676 (altivec_vpkuhus): Likewise.
15677 (altivec_vpkshus): Likewise.
15678 (altivec_vpkuwus): Likewise.
15679 (altivec_vpkswus): Likewise.
15680 (altivec_vpks<VI_char>ss): Likewise.
15681 (altivec_vpks<VI_char>us): Likewise.
15682 (altivec_vpku<VI_char>us): Likewise.
15683 (altivec_vpku<VI_char>um): Likewise.
15684 (altivec_vrl<VI_char>): Likewise.
15685 (altivec_vsl<VI_char>): Likewise.
15686 (altivec_vsr<VI_char>): Likewise.
15687 (altivec_vsra<VI_char>): Likewise.
15688 (altivec_vsldoi_<mode>): Likewise.
15689 (altivec_vupkhsb): Likewise.
15690 (altivec_vupkhs<VU_char>): Likewise.
15691 (altivec_vupkls<VU_char>): Likewise.
15692 (altivec_vupkhsh): Likewise.
15693 (altivec_vupklsb): Likewise.
15694 (altivec_vupklsh): Likewise.
15695 (altivec_vcmpequ<VI_char>_p): Likewise.
15696 (altivec_vcmpgts<VI_char>_p): Likewise.
15697 (altivec_vcmpgtu<VI_char>_p): Likewise.
15698 (abs<mode>2): Likewise.
15699 (vec_unpacks_hi_v16qi): Likewise.
15700 (vec_unpacks_hi_v8hi): Likewise.
15701 (vec_unpacks_lo_v16qi): Likewise.
15702 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15703 (vec_unpacks_lo_v8hi): Likewise.
15704 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15705 (vec_pack_trunc_v8h): Likewise.
15706 (vec_pack_trunc_v4si): Likewise.
15707 (vec_pack_trunc_<mode>): Likewise.
15708
15709 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15710 V2DI builtins.
15711 (vec_vmaxsd): Likewise.
15712 (vec_vmaxud): Likewise.
15713 (vec_vminsd): Likewise.
15714 (vec_vminud): Likewise.
15715 (vec_vpksdss): Likewise.
15716 (vec_vpksdus): Likewise.
15717 (vec_vpkudum): Likewise.
15718 (vec_vpkudus): Likewise.
15719 (vec_vrld): Likewise.
15720 (vec_vsld): Likewise.
15721 (vec_vsrad): Likewise.
15722 (vec_vsrd): Likewise.
15723 (vec_vsubudm): Likewise.
15724 (vec_vupkhsw): Likewise.
15725 (vec_vupklsw): Likewise.
15726
e70670cf
JH		 157272013-05-29 Jan Hubicka <jh@suse.cz>
15728
15729 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15730 flags; reorder rest of fields in more consistent way.
15731 (varpool_node): Remove analyzed, finalized and alias.
15732 (cgraph_ndoe): Likewise.
15733 (symtab_alias_ultimate_target): New function.
15734 (cgraph_function_node): Move offline.
15735 (cgraph_reset_node): Declare.
15736 (cgraph_comdat_can_be_unshared_p): Remove.
15737 (varpool_remove_initializer): Declare.
15738 (varpool_first_defined_variable, varpool_next_defined_variable
15739 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15740 (cgraph_function_with_gimple_body_p): Update.
15741 (varpool_all_refs_explicit_p): Update.
15742 (symtab_alias_target): New function.
15743 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15744 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 15745 (cgraph_function_or_thunk_node): Simplify using
15746 symtab_alias_ultimate_target.
e70670cf
JH		 15747 (varpool_variable_node): Likewise.
15748 * cgraph.c (cgraph_create_function_alias): Update.
15749 (cgraph_add_thunk): Update.
15750 (cgraph_remove_node): Update.
15751 (dump_cgraph_node): Do not dump removed flags.
15752 (cgraph_function_body_availability): Update.
15753 (cgraph_propagate_frequency): Update.
15754 (verify_cgraph_node): Check sanity of local flag.
15755 (cgraph_function_node): Move here from cgraph.h; revamp for
15756 cgraph_function_or_thunk_node.
15757 * lto-symtab.c (lto_varpool_replace_node): Update.
15758 (lto_symtab_resolve_can_prevail_p): Update.
15759 (lto_symtab_merge_cgraph_nodes): Update.
15760 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15761 propagate_constants_accross_call, devirtualization_time_bonus,
15762 ipcp_propagate_stage): Update.
15763 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 15764 * ipa-inline-transform.c (clone_inlined_nodes,
15765 preserve_function_body_p): Update.
e70670cf
JH		 15766 * ipa-reference.c (propagate): Update.
15767 (write_node_summary_p): Update.
15768 * toplev.c (wrapup_global_declaration_2): Update.
15769 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15770 (analyze_function) ... this one.
15771 (cgraph_process_new_functions): Update.
15772 (cgraph_reset_node): Export.
15773 (cgraph_finalize_function): Update.
15774 (cgraph_add_new_function): Update.
15775 (process_function_and_variable_attributes): Update.
15776 (varpool_finalize_decl): Update.
15777 (symbol_finalized): Remove.
15778 (symbol_finalized_and_needed): Rename to ...
15779 (symbol_defined_and_needed): ... update.
15780 (cgraph_analyze_functions): Update.
15781 (handle_alias_pairs): Update.
15782 (mark_functions_to_output): Update.
15783 (assemble_thunk): Update.
15784 (output_in_order): Update.
15785 (output_weakrefs): Update.
15786 (finalize_compilation_unit): Update.
15787 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15788 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15789 input_node, input_varpool_node): Update.
15790 * dbxout.c (dbxout_expand_expr): Update.
15791 * cgraphclones.c (cgraph_clone_node): Update.
15792 (cgraph_copy_node_for_versioning): Update.
15793 (cgraph_materialize_clone): Update.
15794 (cgraph_materialize_all_clones): Update.
15795 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15796 propagate_pure_const, propagate_nothrow): Update.
15797 * lto-streamer-out.c (lto_output, write_symbol): Update.
15798 * ipa-utils.c (ipa_reverse_postorder): Update.
15799 * ipa-inline.c (can_inline_edge_p): Update.
15800 (update_caller_keys, ipa_inline): Update.
15801 * dwarf2out.c (reference_to_unused,
15802 premark_types_used_by_global_vars_helper): Update.
15803 * tree-eh.c (tree_could_trap_p): Update.
15804 * ipa-split.c (consider_split, execute_split_functions): Update.
15805 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 15806 has_addr_references_p): Update; move ahead in file for better
15807 readability.
e70670cf
JH		 15808 (process_references): Simplify.
15809 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15810 bodies are removed.
15811 (cgraph_comdat_can_be_unshared_p): Make static.
15812 (cgraph_externally_visible_p): Update.
15813 (varpool_externally_visible_p): Update.
15814 (function_and_variable_visibility): Update.
15815 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15816 ipa_tm_mark_force_output_node): Update.
15817 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15818 estimate_edge_devirt_benefit, inline_generate_summary,
15819 inline_write_summary): Update.
15820 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15821 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 15822 (ipa_print_node_params, ipa_prop_read_section,
15823 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 15824 * varasm.c (mark_decl_referenced): Update.
15825 (assemble_alias, dump_tm_clone_pairs): Update.
15826 * tree-inline.c (copy_bb): Update.
15827 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15828 Update.
15829 * symtab.c (dump_symtab_base): Print new flags.
15830 (verify_symtab_base): Verify new flags.
15831 (symtab_alias_ultimate_target): New function.
15832 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15833 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15834 Update.
c1e183a9
UB		 15835 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15836 Update.
e70670cf
JH		 15837 * i386.c (ix86_get_function_versions_dispatcher,
15838 ix86_generate_version_dispatcher_body): Update.
15839 (fold_builtin_cpu): Use varpool_add_new_variable.
15840 * varpool.c (varpool_remove_initializer): Break out from ...
15841 (varpool_remove_node): ... this one.
15842 (dump_varpool_node, varpool_node_for_asm,
15843 cgraph_variable_initializer_availability, varpool_analyze_node,
15844 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 15845 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15846 Update.
e70670cf 15847
182802ad
JH		 158482013-05-29 Jan Hubicka <jh@suse.cz>
15849
15850 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15851
8190b609
ER		 158522013-05-29 Easwaran Raman <eraman@google.com>
15853
15854 PR tree-optimization/57442
15855 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15856 when control exits the main loop.
15857
69f5aa9b
SKS		 158582013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15859
15860 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15861 and RX600.
c1e183a9 15862 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 15863 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15864 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15865 option.
15866
4b847da9
BS		 158672013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15868
15869 PR tree-optimization/57441
15870 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15871 Don't limit size of incr_vec to number of candidates.
15872
4fc43c39
SE		 158732013-05-29 Steve Ellcey <sellcey@imgtec.com>
15874
15875 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15876 and mips16 directories.
c1e183a9 15877 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 15878 (MULTILIB_DIRNAMES): Ditto.
15879 (MULTILIB_EXCEPTIONS): Add new exceptions.
15880 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15881 (MULTILIB_DIRNAMES): Ditto.
15882 (MULTILIB_EXCEPTIONS): Add new exceptions.
15883
12211b99 158842012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 15885 Marcus Shawcroft <marcus.shawcroft@arm.com>
15886
15887 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15888 SYMBOL_TINY_ABSOLUTE.
15889 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15890 SYMBOL_TINY_ABSOLUTE.
15891 (aarch64_expand_mov_immediate): Likewise.
15892 (aarch64_classify_symbol): Likewise.
15893 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15894 Permit SYMBOL_TINY_ABSOLUTE.
15895 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15896
12211b99 158972013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 15898 Marcus Shawcroft <marcus.shawcroft@arm.com>
15899
15900 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15901 Refactor if/switch. Replace gcc_assert with if.
15902
c0186656
GG		 159032013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15904
15905 * config/i386/i386.c (initial_ix86_tune_features): Enable
15906 FP Reassociation for AMD bdver1 and bdver2.
15907
d20188f3
MJ		 159082013-05-29 Martin Jambor <mjambor@suse.cz>
15909
15910 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15911 and IMAGPART_EXPR do not occur within other handled_components.
15912
292cba13
RB		 159132013-05-29 Richard Biener <rguenther@suse.de>
15914
15915 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15916 access on whether the use is in the BB we currently try to
15917 vectorize.
15918 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15919 vectorize to vect_bb_slp_scalar_cost.
15920
6eddf228
RB		 159212013-05-29 Richard Biener <rguenther@suse.de>
15922
15923 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15924 computing scalar cost offsetted by stmts that are kept live
15925 by scalar uses.
15926 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15927 for computation of scalar cost.
15928
7df36117
SE		 159292013-05-28 Steve Ellcey <sellcey@mips.com>
15930
15931 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15932
36536d79
BI		 159332013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15934
15935 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15936 array notation built-in reduction functions.
15937 * doc/passes.texi (Passes): Added documentation about changes done
15938 for Cilk Plus.
15939 * doc/invoke.texi (C Dialect Options): Added documentation about
15940 the -fcilkplus flag.
15941 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15942 (BUILTINS_DEF): Depend on cilkplus.def.
15943 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15944 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15945 * cilkplus.def: New file.
15946
12211b99 159472013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 15948
15949 PR rtl-optimization/57439
15950 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15951
9daf6dbb
ER		 159522013-05-28 Easwaran Raman <eraman@google.com>
15953
15954 PR tree-optimization/57337
15955 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15956 (find_insert_point): Correctly identify the insertion point
15957 when two statements with the same UID is compared.
15958
fbd7e877
RB		 159592013-05-28 Richard Biener <rguenther@suse.de>
15960
15961 PR tree-optimization/56787
15962 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15963 from the list of data references.
15964 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15965 clobbers.
15966 (vect_analyze_loop_operations): Likewise.
15967 (vect_transform_loop): Remove clobbers.
15968
bbba1117
MJ		 159692013-05-28 Martin Jambor <mjambor@suse.cz>
15970
15971 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15972 and REALPART_EXPRs have scalar type.
15973
bd388c2a
RB		 159742013-05-28 Richard Biener <rguenther@suse.de>
15975
15976 PR tree-optimization/57411
15977 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15978 virtual operands.
15979 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15980 virtual operand propagation.
15981
2f56a311
EB		 159822013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15983
15984 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15985 destination register for bmasksi_vis.
15986 (vector_init_bshuffle): Likewise.
15987 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15988
5c3eacbb
EB		 159892013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15990
15991 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15992 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15993 mode if the instruction isn't available in the original mode.
15994 * config/sparc/sparc.opt (mfix-ut699): New option.
15995 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15996 (divdf3): Turn into expander.
15997 (divdf3_nofix): New insn.
15998 (divdf3_fix): Likewise.
15999 (divsf3): Disable if -mfix-ut699.
16000 (sqrtdf2): Turn into expander.
16001 (sqrtdf2_nofix): New insn.
16002 (sqrtdf2_fix): Likewise.
16003 (sqrtsf2): Disable if -mfix-ut699.
16004
a1756c0a
RB		 160052013-05-27 Richard Biener <rguenther@suse.de>
16006
16007 PR middle-end/57412
16008 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16009 block for the new loop.
16010
5a892248
RB		 160112013-05-27 Richard Biener <rguenther@suse.de>
16012
16013 PR tree-optimization/57343
16014 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16015 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16016 (number_of_iterations_cond): Do not build the folded tree.
16017
d1de852b
RB		 160182013-05-27 Richard Biener <rguenther@suse.de>
16019
16020 Revert
16021 PR middle-end/57381
16022 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16023 OEP_CONSTANT_ADDRESS_OF retained.
16024
16025 PR tree-optimization/57417
16026 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16027 for unchanged base.
16028 (set_ssa_val_to): Compare addresses using
16029 get_addr_base_and_unit_offset.
16030
12211b99 160312013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 16032
16033 PR rtl-optimization/56833
16034 * postreload.c (move2add_record_mode): New function.
16035 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16036 (move2add_use_add2_insn): Use move2add_record_sym_value.
16037 (move2add_use_add3_insn): Likewise.
16038 (reload_cse_move2add): Use move2add_valid_value_p and
16039 move2add_record_mode. Invalidate call-clobbered and REG_INC
16040 affected regs by setting reg_mode to VOIDmode.
16041 (move2add_note_store): Don't pretend the inside of a SUBREG is
16042 the actual destination. Invalidate single/leading registers by
16043 setting reg_mode to VOIDmode.
16044 Use move2add_record_sym_value, move2add_valid_value_p and
16045 move2add_record_mode.
16046
b03be25f
RB		 160472013-05-27 Richard Biener <rguenther@suse.de>
16048
16049 PR tree-optimization/57396
16050 * tree-affine.c (double_int_constant_multiple_p): Properly
16051 return false for val == 0 and div != 0.
16052
44e88db2
RS		 160532013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16054
16055 * config/mips/mips.h: Use #elif in preprocessor conditions.
16056
3b859704
RS		 160572013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16058
16059 PR target/53916
16060 * config/mips/constraints.md (kl): New constraint.
16061 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16062 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16063 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16064 from LO for MIPS16.
16065 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16066
c979d5f5
RS		 160672013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16068
16069 PR target/55777
16070 * config/mips/mips.c (mips_can_inline_p): New function.
16071 (TARGET_CAN_INLINE_P): Define.
16072
8e90de43
SB		 160732013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16074
16075 * sched-int.h (ds_t, dw_t): Make unsigned int.
16076 Fix documentation that describes how all the ds_t bits are used.
16077 Reserve the last bit for delayed-branch scheduling.
16078 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16079 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16080 (gen_dep_weak_1): Remove prototype.
16081 * sched-deps.c (get_dep_weak_1): Make static.
16082 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16083 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16084 * doc/tm.texi: Regenerate.
16085 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16086
cb5cb194
SB		 160872013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16088
16089 PR debug/56950
16090 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16091
5bd93ff6
NS		 160922013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16093 Sandra Loosemore <sandra@codesourcery.com>
16094
16095 * config.gcc (powerpc-*): Allow native for with-cpu.
16096
2343af65
JL		 160972013-05-24 Jeff Law <law@redhat.com>
16098
16099 PR tree-optimization/57124
16100 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16101 conversion feeding a condition if the range has an overflow
16102 if -fstrict-overflow. Add warnings for when we do make the
16103 transformation.
16104
3ad695b9
DC		 161052013-05-24 Dehao Chen <dehao@google.com>
16106
04960246 16107 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 16108 (locus_discrim_hasher): Likewise.
16109 (locus_discrim_hasher::hash): Likewise.
16110 (locus_discrim_hasher::equal): Likewise.
16111
82338059
MJ		 161122013-05-24 Martin Jambor <mjambor@suse.cz>
16113
16114 PR tree-optimization/57294
16115 * cgraph.h (ipa_record_stmt_references): Declare.
16116 * cgraphbuild.c (ipa_record_stmt_references): New function.
16117 (build_cgraph_edges): Use ipa_record_stmt_references.
16118 (rebuild_cgraph_edges): Likewise.
16119 (cgraph_rebuild_references): Likewise.
16120 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16121 associated with the old statement and build references from the
16122 newly built statements.
16123 * ipa-ref.c (ipa_remove_stmt_references): New function.
16124 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16125
1ccd4874
VM		 161262013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16127
55805e54 16128 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 16129 mem-mem moves.
55805e54 16130 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 16131 too.
55805e54 16132 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 16133 secondary memory moves.
16134 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16135 reg set up in the current insn.
16136
25e25c73
DC		 161372013-05-24 Dehao Chen <dehao@google.com>
16138
04960246 16139 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 16140 hash function.
16141 (locus_descrim_hasher::equal): Likewise.
04960246 16142 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 16143 (make_edges): Likewise.
16144 (next_discriminator_for_locus): Likewise.
16145 (same_line_p): Likewise.
16146 (assign_discriminators): Likewise.
16147 (make_cond_expr_edges): Likewise.
16148 (make_gimple_switch_edges): Likewise.
16149 (make_goto_expr_edges): Likewise.
16150 (make_gimple_asm_edges): Likewise.
16151
50d38551
IB		 161522013-05-24 Ian Bolton <ian.bolton@arm.com>
16153
16154 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16155 X format specifier to only display bottom 16 bits.
16156 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16157 immediate to match for operand 2, since it will be masked.
16158
aea0101d
RB		 161592013-05-24 Richard Biener <rguenther@suse.de>
16160
16161 PR tree-optimization/57287
16162 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16163 all SSA names that occur in abnormal PHIs.
16164
634e03d3
AI		 161652013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16166
16167 PR tree-ssa/57385
16168 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16169 that index is not negative.
16170
b17c024f
EB		 161712013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16172
16173 PR rtl-optimization/55177
16174 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16175 (simplify_byte_swapping_operation): New.
16176 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16177 (simplify_relational_operation_1): Deal with BSWAP.
16178
46aeac1b
RH		 161792013-05-23 Richard Henderson <rth@redhat.com>
16180
16181 PR target/56742
16182 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16183 (ix86_reorg): Call it.
16184
70cc1536
UB		 161852013-05-23 Uros Bizjak <ubizjak@gmail.com>
16186
16187 PR target/57379
16188 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16189 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16190 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16191
b3851501
CB		 161922013-05-23 Christian Bruel <christian.bruel@st.com>
16193
16194 PR debug/57351
16195 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16196
12211b99 161972013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 16198 Marcus Shawcroft <marcus.shawcroft@arm.com>
16199
16200 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16201 * config/aarch64/constraints.md (Usa): Remove.
16202 * doc/md.texi (AArch64 Usa): Remove.
16203
12211b99 162042013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 16205 Marcus Shawcroft <marcus.shawcroft@arm.com>
16206
16207 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16208 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16209 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16210 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16211
12211b99 162122013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 16213
16214 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16215 instruction (AdvSIMD).
16216 * config/aarch64/aarch64-builtins.c
16217 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16218 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16219
76c36cb1 162202013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 16221
16222 PR middle-end/57347
16223 * tree.h (contains_bitfld_component_ref_p): Declare.
16224 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 16225 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16226 caller.
7d2fb524
MJ		 16227 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16228 not access a bit-field. Assert all final offsets are byte-aligned.
16229
ce521ae6
RB		 162302013-05-23 Richard Biener <rguenther@suse.de>
16231
16232 PR tree-optimization/57380
16233 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16234 least one invariant or re-used load.
16235 * passes.c (init_optimization_passes): Move pass_phiprop before
16236 pass_forwprop.
16237
75c7257f
JG		 162382013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16239
16240 * config/aarch64/aarch64-simd.md
16241 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16242
af6d44b5
RB		 162432013-05-23 Richard Biener <rguenther@suse.de>
16244
16245 PR middle-end/57381
16246 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16247 OEP_CONSTANT_ADDRESS_OF retained.
16248
bd3647bf
JJ		 162492013-05-23 Jakub Jelinek <jakub@redhat.com>
16250
16251 PR middle-end/57344
70cc1536
UB		 16252 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16253 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 16254
a55757ea
RB		 162552013-05-23 Richard Biener <rguenther@suse.de>
16256
16257 PR rtl-optimization/57341
16258 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16259 instead of true_dependence.
16260
cfb00b41
DM		 162612013-05-22 David Malcolm <dmalcolm@redhat.com>
16262
dd1c676f
DM		 16263 * bb-reorder.c (branch_threshold): Make const.
16264 (exec_threshold): Ditto.
cfb00b41 16265
f62511da 162662013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16267 Pat Haugen <pthaugen@us.ibm.com>
16268 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16269
16270 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16271 documentation for the power8 crypto builtins.
16272
16273 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16274
16275 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16276 macros for defining power8 builtin functions.
16277 (BU_P8V_AV_2): Likewise.
16278 (BU_P8V_AV_P): Likewise.
16279 (BU_P8V_VSX_1): Likewise.
16280 (BU_P8V_OVERLOAD_1): Likewise.
16281 (BU_P8V_OVERLOAD_2): Likewise.
16282 (BU_CRYPTO_1): Likewise.
16283 (BU_CRYPTO_2): Likewise.
16284 (BU_CRYPTO_3): Likewise.
16285 (BU_CRYPTO_OVERLOAD_1): Likewise.
16286 (BU_CRYPTO_OVERLOAD_2): Likewise.
16287 (XSCVSPDP): Fix typo, point to the correct instruction.
16288 (VCIPHER): Add power8 crypto builtins.
16289 (VCIPHERLAST): Likewise.
16290 (VNCIPHER): Likewise.
16291 (VNCIPHERLAST): Likewise.
16292 (VPMSUMB): Likewise.
16293 (VPMSUMH): Likewise.
16294 (VPMSUMW): Likewise.
16295 (VPERMXOR_V2DI): Likewise.
16296 (VPERMXOR_V4SI: Likewise.
16297 (VPERMXOR_V8HI: Likewise.
16298 (VPERMXOR_V16QI: Likewise.
16299 (VSHASIGMAW): Likewise.
16300 (VSHASIGMAD): Likewise.
16301 (VPMSUM): Likewise.
16302 (VPERMXOR): Likewise.
16303 (VSHASIGMA): Likewise.
16304
16305 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16306 __CRYPTO__ if the crypto instructions are available.
16307 (altivec_overloaded_builtins): Add support for overloaded power8
16308 builtins.
16309
16310 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16311 support for power8 crypto builtins.
16312 (builtin_function_type): Likewise.
16313 (altivec_init_builtins): Add support for builtins that take vector
16314 long long (V2DI) arguments.
16315
16316 * config/rs6000/crypto.md: New file, define power8 crypto
16317 instructions.
16318
163192013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16320 Pat Haugen <pthaugen@us.ibm.com>
16321 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16322
16323 * doc/invoke.texi (Option Summary): Add power8 options.
16324 (RS/6000 and PowerPC Options): Likewise.
16325
16326 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16327 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16328 wm, wn, wr documentation.
16329
f43245d1 16330 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM		 16331 registers if direct move instructions are enabled.
16332 (wn): New constraint for no registers.
16333 (wq): New constraint for quad word even GPR registers.
16334 (wr): New constraint if 64-bit instructions are enabled.
16335 (wv): New constraint if power8 vector instructions are enabled.
16336 (wQ): New constraint for quad word memory locations.
16337
f43245d1 16338 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM		 16339 constraint for 0..15 for crypto instructions.
16340 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16341 as GPR and floating point registers.
16342 (int_reg_operand): New predicate to match only GPR registers.
16343 (base_reg_operand): New predicate to match base registers.
16344 (quad_int_reg_operand): New predicate to match even GPR registers
16345 for quad memory operations.
16346 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16347 operations in both GPR and VSX registers.
16348 (quad_memory_operand): New predicate for quad memory operations.
16349 (reg_or_indexed_operand): New predicate for direct move support.
16350
f43245d1 16351 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM		 16352 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16353 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16354 (POWERPC_MASKS): Add power8 options.
16355 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16356 various options.
16357
f43245d1 16358 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM		 16359 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16360
f43245d1 16361 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM		 16362 (-mpower8-fusion): New power8 options.
16363 (-mpower8-fusion-sign): Likewise.
16364 (-mpower8-vector): Likewise.
16365 (-mcrypto): Likewise.
16366 (-mdirect-move): Likewise.
16367 (-mquad-memory): Likewise.
16368
f43245d1 16369 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM		 16370 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16371 registers.
70cc1536 16372 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 16373 (rs6000_debug_vector_unit): Add p8_vector.
16374 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16375 definitions. Also print fusion state.
16376 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16377 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16378 (rs6000_option_override_internal): Add support for power8.
16379 (rs6000_common_init_builtins): Add debugging for skipped builtins
16380 if -mdebug=builtin.
16381 (rs6000_adjust_cost): Add power8 support.
16382 (rs6000_issue_rate): Likewise.
16383 (insn_must_be_first_in_group): Likewise.
16384 (insn_must_be_last_in_group): Likewise.
16385 (force_new_group): Likewise.
16386 (rs6000_register_move_cost): Likewise.
16387 (rs6000_opt_masks): Likewise.
16388
16389 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16390 power8 capable assembler, default to power7 options.
16391 (TARGET_DIRECT_MOVE): Likewise.
16392 (TARGET_CRYPTO): Likewise.
16393 (TARGET_P8_VECTOR): Likewise.
16394 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16395 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16396 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16397 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16398 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16399 (TARGET_XSCVDPSPN): Likewise.
16400 (TARGET_XSCVSPDPN): Likewsie.
16401 (TARGET_SYNC_HI_QI): Likewise.
16402 (TARGET_SYNC_TI): Likewise.
16403 (MASK_CRYPTO): Likewise.
16404 (MASK_DIRECT_MOVE): Likewise.
16405 (MASK_P8_FUSION): Likewise.
16406 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 16407 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16408 temporary used by some of the direct move instructions to get two FP
16409 temporary registers does not force creation of a stack frame.
f62511da
MM		 16410 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16411 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16412 that any VSX registers are tieable, even if they are also an
16413 Altivec vector mode.
16414 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16415 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16416 (RS6000_BTM_CRYPTO): Likewise.
16417 (RS6000_BTM_COMMON): Likewise.
16418
16419 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16420 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16421 (enum rs6000_vector): Add power8 vector support.
16422
73a1a707
RR		 164232013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16424
16425 PR target/19599
16426 PR target/57340
16427 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16428 (any_sibcall_could_use_r3): this and handle indirect calls.
16429 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16430
d28073d4
BS		 164312013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16432
16433 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16434
bee0b10c
RB		 164352013-05-22 Richard Biener <rguenther@suse.de>
16436
16437 PR middle-end/57349
16438 * profile.c (branch_prob): Do not split blocks that are
16439 abnormally receiving from ECF_RETURNS_TWICE functions.
16440
98409b51
RS		 164412013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16442
16443 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16444 address mode. Move pointer mode initialization to the same place.
16445
c0602ab8
MZ		 164462013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16447
16448 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16449 while it has any effect.
16450
4f7a634e
ER		 164512013-05-21 Easwaran Raman <eraman@google.com>
16452
16453 PR tree-optimization/57322
9daf6dbb
ER		 16454 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16455 UID of the statement added to the BB to be 1.
4f7a634e 16456
a4ce1258
JJ		 164572013-05-21 Jakub Jelinek <jakub@redhat.com>
16458
16459 PR tree-optimization/57331
70cc1536
UB		 16460 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16461 of conversion from pointer type to integral type with integer.
a4ce1258 16462
1b14621a
MJ		 164632013-05-21 Martin Jambor <mjambor@suse.cz>
16464
16465 PR lto/57289
16466 * ipa-prop.c (ipa_read_node_info): Process param_used and
16467 controlled_uses in the same order as when writing.
16468
e60661f0
MG		 164692013-05-21 Magnus Granberg <baldrick@free.fr>
16470
16471 PR plugins/56754
ee49aa34 16472 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16473
4a61be9a
RB		 164742013-05-21 Richard Biener <rguenther@suse.de>
16475
16476 PR tree-optimization/57318
16477 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16478 estimate stmts with side-effects as likely eliminated.
16479
c52da5f7
RB		 164802013-05-21 Richard Biener <rguenther@suse.de>
16481
16482 PR tree-optimization/57330
16483 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16484 preserve the call stmts fntype.
16485
7ec67e2a
RB		 164862013-05-21 Richard Biener <rguenther@suse.de>
16487
16488 PR tree-optimization/57303
16489 * tree-ssa-sink.c (statement_sink_location): Improve killing
16490 stmt detection and properly handle self-assignments.
16491
b112d513
CB		 164922013-05-21 Christian Bruel <christian.bruel@st.com>
16493
55805e54
YR		 16494 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16495 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 16496 registers. Set register size out of the PARALLEL loop.
16497
14c2ec26
OE		 164982013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16499
16500 PR target/56547
16501 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16502 (*fmasf4, *fmasf4_media): New insns.
16503
da734fa1
RS		 165042013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16505
16506 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16507 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16508 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16509 (mips_idiv_insns): Update the comments to say that the returned
16510 instruction counts are in units of BASE_INSN_LENGTH.
16511 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16512 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16513 using 2 rather than 4 as the length of indirect MIPS16 and
16514 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16515 length of a NOP. Don't divide MIPS16 lengths by 2.
16516 (mips16_split_long_branches): Assume a branch is long if the
16517 length is greater than 4 rather than 8.
16518 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16519 rather than multiplying them by 2. Multiply instruction counts
16520 by BASE_INSN_LENGTH rather than 4.
16521 (*jump_mips16, tls_get_tp_mips16_<mode>)
16522 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16523
13719e8b
RS		 165242013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16525
16526 * config/mips/mips.md (extended_mips16): Remove branch case.
16527 (length): Remove duplicated extended_mips16 test.
16528
c3850d14
RS		 165292013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16530
16531 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16532
8da2e059
RS		 165332013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16534
16535 * recog.h (Recog_data): Rename to...
16536 (recog_data_d): ...this.
16537 (recog_data): Update accordingly.
16538 * recog.c (recog_data): Likewise.
16539 * reload.c (save_recog_data): Likewise.
16540 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16541 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16542
2df013f3
JB		 165432013-05-17 Julian Brown <julian@codesourcery.com>
16544
16545 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16546 found in a REG_EQUAL note, invalidate it.
16547
371e77e3 165482013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 16549
16550 * tree-ssa-reassoc.c (find_insert_point): New function.
16551 (insert_stmt_after): Likewise.
16552 (get_def_stmt): Likewise.
16553 (ensure_ops_are_available): Likewise.
16554 (not_dominated_by): Likewise.
16555 (rewrite_expr_tree): Do not move statements beyond what is
16556 necessary. Remove call to swap_ops_for_binary_stmt...
16557 (reassociate_bb): ... and move it here.
16558 (build_and_add_sum): Assign UIDs for new statements.
16559 (linearize_expr): Likewise.
16560 (do_reassoc): Renumber gimple statement UIDs.
16561
e01c7cca
JH		 165622013-05-17 Jan Hubicka <jh@suse.cz>
16563
16564 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16565 weakrefs.
16566 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 16567 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16568 weakrefs.
e01c7cca
JH		 16569 (output_weakrefs): Update.
16570
c3272a92
PCC		 165712013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16572 Martin Jambor <mjambor@suse.cz>
16573
16574 PR middle-end/57276
16575 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16576 value that corresponds to the given aggval is found in values vector.
16577
11c2aa39
UB		 165782013-05-17 Uros Bizjak <ubizjak@gmail.com>
16579
16580 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16581 sse, sse2, sse3, ssse3 and sse4a flags to options.
16582
abecc8c6
DM		 165832013-05-17 David Malcolm <dmalcolm@redhat.com>
16584
16585 * gengtype-state.c: (s_expr_writer): New class, to handle
16586 prettifying of output layout of s-expressions.
16587 (state_writer): New class, to write out gtype.state.
16588 (state_written_type_count): Move this variable into member data of
16589 state_writer.
16590 (s_expr_writer::s_expr_writer): New code: constructor for new class
16591 (state_writer::state_writer(): ditto
16592 (s_expr_writer::write_new_line): New function
16593 (s_expr_writer::write_any_indent): ditto
16594 (s_expr_writer::begin_s_expr): ditto
16595 (s_expr_writer::end_s_expr): ditto
16596 (write_state_fileloc): convert to method of state_writer...
16597 (state_writer:: write_state_fileloc): ...and use methods of
16598 s_expr_writer to write indentation into the gtype.state output file
16599 to visually represent the hierarchical structure of the list
16600 structures
16601 (write_state_fields): ditto, renaming to...
16602 (state_writer::write_state_fields)
16603 (write_state_a_string): ditto, renaming to...
16604 (state_writer::write_state_a_string)
16605 (write_state_string_option): ditto, renaming to...
16606 (state_writer::write_state_string_option)
16607 (write_state_type_option): ditto, renaming to...
16608 (state_writer::write_state_type_option)
16609 (write_state_nested_option): ditto, renaming to...
16610 (state_writer::write_state_nested_option)
16611 (write_state_option): ditto, renaming to...
16612 (state_writer::write_state_option)
16613 (write_state_options): ditto, renaming to...
16614 (state_writer::write_state_options)
16615 (write_state_lang_bitmap): ditto, renaming to...
16616 (state_writer::write_state_lang_bitmap)
16617 (write_state_version): ditto, renaming to...
16618 (state_writer::write_state_version)
16619 (write_state_scalar_type): ditto, renaming to...
16620 (state_writer::write_state_scalar_type)
16621 (write_state_string_type): ditto, renaming to...
16622 (state_writer::write_state_string_type)
16623 (write_state_undefined_type): ditto, renaming to...
16624 (state_writer::write_state_undefined_type)
16625 (write_state_struct_union_type): ditto, renaming to...
16626 (state_writer::write_state_struct_union_type)
16627 (write_state_struct_type): ditto, renaming to...
16628 (state_writer::write_state_struct_type)
16629 (write_state_user_struct_type): ditto, renaming to...
16630 (state_writer::write_state_user_struct_type)
16631 (write_state_lang_struct_type): ditto, renaming to...
16632 (state_writer::write_state_lang_struct_type)
16633 (write_state_param_struct_type): ditto, renaming to...
16634 (state_writer::write_state_param_struct_type)
16635 (write_state_pointer_type): ditto, renaming to...
16636 (state_writer::write_state_pointer_type)
16637 (write_state_array_type): ditto, renaming to...
16638 (state_writer::write_state_array_type)
16639 (write_state_gc_used): ditto, renaming to...
16640 (state_writer::write_state_gc_used)
16641 (write_state_common_type_content): ditto, renaming to...
16642 (state_writer::write_state_common_type_content)
16643 (write_state_type): ditto, renaming to...
16644 (state_writer::write_state_type)
16645 (write_state_pair_list): ditto, renaming to...
16646 (state_writer::write_state_pair_list)
16647 (write_state_pair): ditto, renaming to...
16648 (state_writer::write_state_pair)
16649 (write_state_typedefs): ditto, renaming to...
16650 (state_writer::write_state_typedefs)
16651 (write_state_structures): ditto, renaming to...
16652 (state_writer::write_state_structures)
16653 (write_state_param_structs): ditto, renaming to...
16654 (state_writer::write_state_param_structs)
16655 (write_state_variables): ditto, renaming to...
16656 (state_writer::write_state_variables)
16657 (write_state_srcdir): ditto, renaming to...
16658 (state_writer::write_state_srcdir)
16659 (write_state_files_list): ditto, renaming to...
16660 (state_writer::write_state_files_list)
16661 (write_state_languages): ditto, renaming to...
16662 (state_writer::write_state_languages)
16663 (write_state): create a state_writer instance and use it when
16664 writing out the state file
16665
d6545f29
MS		 166662013-05-17 Mike Stump <mikestump@comcast.net>
16667
816a3f73 16668 PR rtl-optimization/57304
d6545f29
MS		 16669 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16670 accessing DF_REF_REAL_LOC.
16671
38c821cf
JJ		 166722013-05-17 Jakub Jelinek <jakub@redhat.com>
16673
16674 PR rtl-optimization/57281
16675 PR rtl-optimization/57300
16676 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16677 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16678 what the other splitter did if the registers are dead.
16679
2a293391
RB		 166802013-05-17 Richard Biener <rguenther@suse.de>
16681
16682 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16683 MEM_REF offsets.
16684
17042d2b
JJ		 166852013-05-17 Jakub Jelinek <jakub@redhat.com>
16686
16687 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16688 linking.
16689
5b115c1f
MP		 166902013-05-17 Marek Polacek <polacek@redhat.com>
16691
11c2aa39
UB		 16692 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16693 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 16694
52d84413
JJ		 166952013-05-17 Jakub Jelinek <jakub@redhat.com>
16696
68119618
JJ		 16697 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16698 vect_external_def oprnd1 with loop_vinfo, try to emit
16699 optional cast, negation and and stmts on the loop preheader
16700 edge instead of into the pattern def seq.
16701
52d84413
JJ		 16702 PR tree-optimization/57051
16703 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16704 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16705
17b962bd
NC		 167062013-05-16 Nick Clifton <nickc@redhat.com>
16707
16708 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16709 (rl78_is_naked_func): New function.
56aefbf7
UB		 16710 (rl78_expand_prologue): Skip prologue generation for naked functions.
16711 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 16712 * doc/extend.texi (naked): Add RL78 to the list of processors
16713 that supports this attribute.
16714
b1a0f84e
JL		 167152013-05-16 Jeff Law <law@redhat.com>
16716
16717 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16718
af0e415b
UB		 167192013-05-16 Uros Bizjak <ubizjak@gmail.com>
16720
16721 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16722 cache parameters using detect_caches_amd also for CYRIX,
16723 NSC and TM2 signatures.
16724
19db293a
UB		 167252013-05-16 Uros Bizjak <ubizjak@gmail.com>
16726 Dzianis Kahanovich <mahatma@eu.by>
16727
16728 PR target/45359
16729 PR target/46396
16730 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16731 VIA/Centaur processors and determine their cache parameters
16732 using detect_caches_amd.
16733
251a41b9
TJ		 167342013-05-16 Teresa Johnson <tejohnson@google.com>
16735
16736 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16737 (rtl_verify_edges): New function.
16738 (rtl_verify_bb_insns): Ditto.
16739 (rtl_verify_bb_pointers): Ditto.
16740 (rtl_verify_bb_insn_chain): Ditto.
16741 (rtl_verify_fallthru): Ditto.
16742 (rtl_verify_bb_layout): Ditto.
16743 (rtl_verify_flow_info_1): Outline checks into new functions.
16744 (rtl_verify_flow_info): Ditto.
16745
f14540b6
SE		 167462013-05-16 Steve Ellcey <sellcey@imgtec.com>
16747
16748 * cfghooks.c (copy_bbs): Add update_dominance argument.
16749 * cfghooks.h (copy_bbs): Update prototype.
16750 * tree-cfg.c (gimple_duplicate_sese_region):
16751 Add update_dominance argument.
16752 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16753 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16754 gimple_duplicate_sese_region call.
16755 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16756 Update copy_bbs call.
16757 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16758 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16759
7e9a3abb
JJ		 167602013-05-16 Jakub Jelinek <jakub@redhat.com>
16761
16762 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16763 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16764 vect_recog_rotate_pattern.
16765 (vect_recog_rotate_pattern): New function.
16766
427b248d
JM		 167672013-05-16 Jason Merrill <jason@redhat.com>
16768
16769 * Makefile.in (LLINKER): New variable.
16770 (mostlyclean): Remove link mutex.
16771 * configure.ac: Handle --enable-link-mutex.
16772 * lock-and-run.sh: New script.
16773
b871e3d2
RR		 167742013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16775
16776 PR target/19599
16777 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16778 for NULL decl.
16779
ce858126
RO		 167802013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16781
16782 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16783
798d3d04
GY		 167842013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16785
16786 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16787 * config/arm/arm.c (next_consecutive_mem): New function.
16788 (gen_movmem_ldrd_strd): Likewise.
16789 * config/arm/arm.md (movmemqi): Update condition and code.
16790 (unaligned_loaddi, unaligned_storedi): New patterns.
16791
0baddc45
RO		 167922013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16793
16794 * config.gcc: Obsolete *-*-solaris2.9*.
16795 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16796
aa9d5bcf
RB		 167972013-05-16 Richard Biener <rguenther@suse.de>
16798
16799 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16800 earlier, after GRAPHITE transforms and IV canonicalization.
16801
5a0f4dd3
JJ		 168022013-05-16 Jakub Jelinek <jakub@redhat.com>
16803
16804 * omp-low.c (extract_omp_for_data): For collapsed loops,
16805 if at least one of the loops is known at compile time to
16806 iterate zero times, set count to 0.
16807 (expand_omp_regimplify_p): New function.
16808 (expand_omp_for_generic): For collapsed loops, if at least
16809 one of the loops isn't known to iterate at least once,
16810 add runtime check with setting count to 0.
16811 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16812 For unsigned types if it isn't known at compile time that
16813 the loop will iterate at least once, add runtime check to bypass
16814 the whole loop if initial condition isn't true.
16815
e3753785
NS		 168162013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16817
16818 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16819
43bb4dd1
MG		 168202013-05-16 Marc Glisse <marc.glisse@inria.fr>
16821
16822 PR middle-end/57286
16823 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16824 transformations to avoid an infinite loop.
16825
3571dde6
MP		 168262013-05-16 Marek Polacek <polacek@redhat.com>
16827
16828 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16829
b8b3f0ca
LE		 168302013-05-15 Leif Ekblad <leif@rdos.net>
16831
16832 * config/i386/i386.c (ix86_decompose_address): Use
16833 DEFAULT_TLS_SEG_REG to access TLS segment register.
16834 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16835 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16836 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16837
e299a383
RS		 168382013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16839
16840 PR target/57260
16841 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16842 sibling calls to functions that would normally be lazily bound,
16843 unless $gp is call-clobbered.
16844
e7208ea3
UB		 168452013-05-15 Uros Bizjak <ubizjak@gmail.com>
16846
19db293a 16847 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 16848 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 16849 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16850 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 16851 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16852 of TARGET_3DNOW.
16853 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16854
e7413f3d
AS		 168552013-05-15 Andreas Schwab <schwab@suse.de>
16856
16857 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16858 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16859 third operand.
16860
ad4db775
TJ		 168612013-05-15 Teresa Johnson <tejohnson@google.com>
16862
16863 * loop-unroll.c (report_unroll_peel): Check decision before
16864 emitting unroll/peel message.
16865
af205f67
TJ		 168662013-05-15 Teresa Johnson <tejohnson@google.com>
16867
16868 * function.h (has_bb_partition): New rtl_data flag.
16869 (bb_reorder_complete): Ditto.
16870 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16871 instead of flag_reorder_blocks_and_partition.
16872 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16873 with some enhancements.
16874 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16875 * bb-reorder.c (connect_traces): Check for has_bb_partition
16876 instead of flag_reorder_blocks_and_partition.
16877 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16878 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16879 verify_hot_cold_block_grouping.
16880 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16881
9adcfa3c
RR		 168822013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16883
16884 PR target/19599
16885 * config/arm/predicates.md (call_insn_operand): New predicate.
16886 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16887 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16888 if insn is not a tail call.
16889 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16890 registers.
16891 * config/arm/arm.h (enum reg_class): New caller save register class.
16892 (REG_CLASS_NAMES): Likewise.
16893 (REG_CLASS_CONTENTS): Likewise.
16894 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16895 without decls.
16896
ccb3ad87
RB		 168972013-05-15 Richard Biener <rguenther@suse.de>
16898
16899 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16900 of MSG_OPTIMIZED_LOCATIONS.
16901 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16902 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16903 message.
16904 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16905 of MSG_OPTIMIZED_LOCATIONS.
16906 (execute_vect_slp): Likewise.
16907 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16908 (vect_create_cond_for_alias_checks): Likewise.
16909 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16910 (vect_recog_widen_mult_pattern): Likewise.
16911 (vect_recog_widen_sum_pattern): Likewise.
16912 (vect_recog_over_widening_pattern): Likewise.
16913 (vect_recog_widen_shift_pattern): Likewise.
16914 (vect_recog_vector_vector_shift_pattern): Likewise.
16915 (vect_recog_divmod_pattern): Likewise.
16916 (vect_recog_mixed_size_cond_pattern): Likewise.
16917 (vect_recog_bool_pattern): Likewise.
16918 (vect_pattern_recog_1): Likewise.
16919
48b1474e
MJ		 169202013-05-15 Martin Jambor <mjambor@suse.cz>
16921
16922 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16923 non-functions to builtin_unreachable.
16924 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16925 correct when new direct edges were discovered.
16926
9de04252
MJ		 169272013-05-15 Martin Jambor <mjambor@suse.cz>
16928
16929 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16930 header, print symbol order instead of node uid, print more information
16931 about indirect edge targets.
16932 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16933 uids.
16934 (ipa_make_edge_direct_to_target): Likewise.
16935 (remove_described_reference): Likewise.
16936 (propagate_controlled_uses): Likewise.
16937 (ipa_print_node_params): Also print symbol order.
16938 (ipcp_transform_function): Print symbol order instead of node uids.
16939 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16940 (cgraph_get_create_real_symbol_node): Likewise.
16941 * ipa-cp.c (print_lattice): Likewise.
16942 (print_all_lattices): Likewise.
16943 (determine_versionability): Likewise.
16944 (initialize_node_lattices): Likewise.
16945 (estimate_local_effects): Likewise.
16946 (update_profiling_info): Likewise.
16947 (create_specialized_node): Likewise.
16948 (perhaps_add_new_callers): Likewise.
16949 (decide_about_value): Likewise.
16950 (decide_whether_version_node): Likewise.
16951 (identify_dead_nodes): Likewise.
16952 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16953 (dump_inline_summary): Likewise.
16954 (estimate_node_size_and_time): Likewise.
16955 (inline_analyze_function): Likewise.
16956 * ipa-inline.c (report_inline_failed_reason): Likewise.
16957 (want_early_inline_function_p): Likewise.
16958 (edge_badness): Likewise.
16959 (update_edge_key): Likewise.
16960 (inline_small_functions): Likewise. Add dumping of order to two other
16961 dumps.
16962 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16963 instead of node uids.
16964 (propagate_pure_const): Likewise.
16965 (propagate_pure_const): Likewise.
16966 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16967 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16968 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16969 of node uids.
16970 * tree-pretty-print.c (dump_function_header): Likewise.
16971 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16972 Print symbol order instead of node uids.
16973
1dd03b91
AK		 169742013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16975
16976 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16977 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16978
fdf6a7b9
RB		 169792013-05-15 Richard Biener <rguenther@suse.de>
16980
16981 PR tree-optimization/57275
16982 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16983 return value for fail to do runtime alias checks for gather loads.
16984
2d6e4603
JH		 169852013-05-15 Jan Hubicka <jh@suse.cz>
16986
16987 PR lto/57038
16988 PR lto/47375
e7208ea3
UB		 16989 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16990 weakrefs are not external.
16991 (lto_symtab_merge_decls): Fix thinko when dealing with
16992 non-lto_symtab decls.
2d6e4603
JH		 16993 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16994 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16995 * varpool.c (dump_varpool_node): Dump more flags.
16996
83f44b39
GG		 169972013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16998
16999 * config/i386/i386.c (processor_alias_table): Add instruction
17000 FSGSBASE for AMD bdver3 architecture.
17001
2e55d062
JJ		 170022013-05-14 Jakub Jelinek <jakub@redhat.com>
17003
17004 * tree.c (warn_deprecated_use): Print file:line using locus color.
17005 * diagnostic.c (diagnostic_report_current_module): Print file:line
17006 and file:line:column using locus color.
17007
0bfdb81e
MS		 170082013-05-14 Mike Stump <mikestump@comcast.net>
17009
17010 * gdbinit.in: Add __null.
17011
a508ef22
MS		 170122013-05-14 Mike Stump <mikestump@comcast.net>
17013
17014 * recog.h: Rename struct recog_data to Recog_data.
17015 * recog.c: Likewise.
17016 * reload.c (can_reload_into): Likewise.
17017 * config/picochip/picochip.c: Likewise.
17018
e7180acb
MS		 170192013-05-14 Mike Stump <mikestump@comcast.net>
17020
17021 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17022
61216c26
SB		 170232013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17024
fd6beed4
SB		 17025 * resource.h (struct resources): Remove unch_memory member.
17026 (CLEAR_RESOURCE): Don't clear unch_memory.
17027 * resource.c (mark_referenced_resources): Don't set it.
17028 (mark_set_resources): Likewise.
17029 (mark_target_live_regs): Don't clear it.
17030 (init_resource_info): Likewise.
17031 * reorg.c (resource_conflicts_p): Don't compare it.
17032 (redundant_insn): Don't set it.
17033
61216c26
SB		 17034 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17035 Remove prototypes.
17036 * emit-rtl.c (next_label): Remove unused function.
17037 (skip_consecutive_labels, link_cc0_insns): Move to ...
17038 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17039 only place where these functions are used, and make them static.
17040
418d1b87
MG		 170412013-05-14 Marc Glisse <marc.glisse@inria.fr>
17042
17043 * fold-const.c (fold_negate_expr): Handle vectors.
17044 (fold_truth_not_expr): Make it static.
17045 (fold_invert_truthvalue): New static function.
17046 (invert_truthvalue_loc): Handle vectors. Do not call
17047 fold_truth_not_expr directly.
17048 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17049 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17050 (fold_comparison): Handle vector constants.
17051 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17052 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17053 * tree.h (fold_truth_not_expr): Remove declaration.
17054
fc21784d
JG		 170552013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17056
17057 * config/aarch64/aarch64-simd.md
17058 (aarch64_vcond_internal<mode>): Rename to...
17059 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17060 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17061 float modes. Clarify all iterator modes.
17062 (vcond<mode><mode>): Use new name for vcond expanders.
17063 (vcond<v_cmp_result><mode>): Likewise.
17064 (vcondu<mode><mode>: Likewise.
17065 * config/aarch64/iterators.md (VDQF_COND): New.
17066
d4c52634
MG		 170672013-05-14 Marc Glisse <marc.glisse@inria.fr>
17068
17069 PR bootstrap/57266
17070 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17071 variable for the shift amount. Check that we shift by non-negative
17072 amounts.
17073
2b261262
CLT		 170742013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17075
17076 PR target/42017
17077 * config/arm/arm.h (EPILOGUE_USES): Only return true
17078 for LR_REGNUM after epilogue_completed.
17079
12211b99 170802013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 17081
17082 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 17083 is error_mark_node.
6e022d7b 17084
56cf7859
RO		 170852013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17086
17087 PR target/57261
17088 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17089 and Solaris 11+/x86 with gld.
17090 * configure: Regenerate.
17091
75776c6d
JJ		 170922013-05-14 Jakub Jelinek <jakub@redhat.com>
17093
17094 * expmed.c (expand_shift_1): Canonicalize rotates by
17095 constant bitsize / 2 to bitsize - 1.
7f998021 17096 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 17097 case ROTATERT>: Likewise.
17098
17099 Revert:
17100 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17101
17102 * config/i386/i386.md (rotateinv): New code attr.
17103 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17104 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17105 roll $31, %eax, etc.
17106
df35498a
RB		 171072013-05-14 Richard Biener <rguenther@suse.de>
17108
17109 PR middle-end/57235
17110 * tree-eh.c (sink_clobbers): Give up for successors with
17111 multiple predecessors and no virtual uses.
17112
cc6e7ece
EB		 171132013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17114
17115 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17116 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17117
f778dd4d
JJ		 171182013-05-14 Jakub Jelinek <jakub@redhat.com>
17119
17120 PR middle-end/57251
17121 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17122 the case when both op0 and op1 have VOIDmode.
17123
bad4df9b
KP		 171242013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17125
17126 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17127 in multiply-accumulate mode.
17128
56f3e9ac
GW		 171292013-05-13 Guozhi Wei <carrot@google.com>
17130
17131 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17132
a3d7ab92
KT		 171332013-05-13 Kai Tietz <ktietz@redhat.com>
17134
17135 PR target/56975
17136 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 17137 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 17138 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 17139 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 17140 (ix86_expand_prologue): Likewise.
17141 (ix86_expand_split_stack_prologue): Likewise.
17142 (legitimate_pic_address_disp_p): Likewise.
17143 (legitimize_pic_address): Likewise.
17144 (legitimize_tls_address): Likewise.
17145 (legitimize_pe_coff_symbol): Likewise.
17146 (output_pic_addr_const): Likewise.
17147 (construct_plt_address): Likewise.
17148 (ix86_expand_call): Likewise.
17149 (x86_output_mi_thunk): Likewise.
17150 (x86_function_profiler): Likewise.
17151
c59b7e28
SN		 171522013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17153
17154 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17155 similar switch cases.
17156 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17157 (aarch64_simd_mov_to_<mode>low): Delete.
17158 (aarch64_simd_mov_to_<mode>high): Delete.
17159 (move_lo_quad_<mode>): Add w<-r alternative.
17160 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17161 (aarch64_simd_mov_from_*): Update type attribute.
17162 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17163 statement.
17164
5f28524a
JH		 171652013-05-13 Jan Hubicka <jh@suse.cz>
17166
17167 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17168 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 17169 ix86_expand_epilogue, emit_i387_cw_initialization,
17170 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17171 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 17172
ae6fa899
JJ		 171732013-05-13 Jakub Jelinek <jakub@redhat.com>
17174
17175 PR tree-optimization/45216
17176 PR tree-optimization/57157
17177 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17178 the (-Y) & (B - 1) variant if OP is |.
17179 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17180 return shifted. Use (-op1) & (prec - 1) as other_amount
17181 instead of prec - op1.
17182
4502fe8d
MJ		 171832013-05-13 Martin Jambor <mjambor@suse.cz>
17184
17185 PR middle-end/42371
17186 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17187 (ipa_constant_data): New type.
17188 (ipa_jump_func): Use ipa_constant_data to hold information about
17189 constant jump functions.
17190 (ipa_get_jf_constant): Adjust to jump function type changes.
17191 (ipa_get_jf_constant_rdesc): New function.
17192 (ipa_param_descriptor): New field controlled_uses.
17193 (ipa_get_controlled_uses): New function.
17194 (ipa_set_controlled_uses): Likewise.
17195 * ipa-ref.h (ipa_find_reference): Declare.
17196 * ipa-prop.c (ipa_cst_ref_desc): New type.
17197 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17198 changes.
17199 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17200 New parameter cs. Adjust all callers.
17201 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17202 (remove_described_reference): New function.
17203 (jfunc_rdesc_usable): Likewise.
17204 (try_make_edge_direct_simple_call): Decrement controlled use count,
17205 attempt to remove reference if it hits zero.
17206 (combine_controlled_uses_counters): New function.
17207 (propagate_controlled_uses): Likewise.
17208 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17209 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17210 (ipa_print_node_params): Print described use counter.
17211 (ipa_write_jump_function): Adjust to jump function type changes.
17212 (ipa_read_jump_function): New parameter CS, pass it to
17213 ipa_set_jf_constant. Adjust caller.
17214 (ipa_write_node_info): Stream controlled use count
17215 (ipa_read_node_info): Likewise.
17216 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17217 asserting.
17218 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17219 count. Remove cloning-added reference if it reaches zero.
17220 * ipa-ref.c (ipa_find_reference): New function.
17221
0864bfc2
GG		 172222013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17223
e7208ea3 17224 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 17225 alignment values for AMD BD and BT architectures.
17226
640bfeb2
MG		 172272013-05-13 Marc Glisse <marc.glisse@inria.fr>
17228
17229 * tree-vect-generic.c (uniform_vector_p): Move ...
17230 * tree.c (uniform_vector_p): ... here.
17231 * tree.h (uniform_vector_p): Declare it.
17232 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17233 into a scalar.
17234
3a60f32b
JJ		 172352013-05-13 Jakub Jelinek <jakub@redhat.com>
17236
198fe1bf
JJ		 17237 PR tree-optimization/57230
17238 * tree-ssa-strlen.c (handle_char_store): Record length for
17239 array store from STRING_CST.
17240
3a60f32b
JJ		 17241 PR tree-optimization/57230
17242 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17243 check.
17244
566be57c
JR		 172452013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17246
17247 * config/epiphany/epiphany.c (epiphany_init): Check size of
17248 NUM_MODES_FOR_MODE_SWITCHING.
17249 (epiphany_expand_prologue):
17250 Remove CONFIG_REGNUM initial value handling code.
17251 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17252 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 17253 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 17254 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17255 Don't return 1 for FP_MODE_NONE.
17256 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17257 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17258 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17259 * config/epiphany/epiphany.md (save_config): New pattern.
17260
0f2c2331
UB		 172612013-05-12 Uros Bizjak <ubizjak@gmail.com>
17262
17263 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17264
5b3f93c7
UB		 172652013-05-10 Uros Bizjak <ubizjak@gmail.com>
17266
17267 * config/i386/i386.md (memory): Handle sseishft1.
17268 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17269 (*vec_extractv2di_1): Ditto.
17270
1f873f0e
VM		 172712013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17272
17273 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17274 saved registers.
17275
341427fa 172762013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 17277
17278 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17279 Add mthumb/march=armv7-a multilib.
17280 Add mthumb/march=armv7-r multilib.
17281 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17282
9e69bdde
RC		 172832013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17284
17285 * config/v850/t-rtems: Add more multilibs.
17286
9ff09a22
RB		 172872013-05-10 Richard Biener <rguenther@suse.de>
17288
17289 PR tree-optimization/57214
17290 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17291 not propagate from SSA names that occur in abnormal PHI nodes.
17292
a5e0cd1d
MG		 172932013-05-10 Marc Glisse <marc.glisse@inria.fr>
17294
17295 * stor-layout.c (element_precision): New function.
17296 * machmode.h (element_precision): Declare it.
17297 * tree.c (build_minus_one_cst): New function.
17298 (element_precision): Likewise.
17299 * tree.h (build_minus_one_cst): Declare new function.
17300 (element_precision): Likewise.
17301 * fold-const.c (operand_equal_p): Use element_precision.
17302 (fold_binary_loc): Handle vector types.
17303 * convert.c (convert_to_integer): Use element_precision.
17304 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17305 separately.
17306
cb2558bc
RS		 173072013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17308
17309 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17310 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17311 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17312 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17313 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17314 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17315 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17316 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17317 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17318 (Uuw8): New constraints.
17319 (Usb4): Move into alphabetical order.
17320 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17321 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17322 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17323 previously unnamed patterns.
17324 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17325 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17326 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17327 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17328 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17329 of specific lengths.
17330
cb3b8d33
JJ		 173312013-05-10 Jakub Jelinek <jakub@redhat.com>
17332
6f93c008
JJ		 17333 * config/i386/i386.md (rotateinv): New code attr.
17334 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17335 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17336 roll $31, %eax, etc.
17337
cb3b8d33
JJ		 17338 PR tree-optimization/45216
17339 PR tree-optimization/57157
17340 * tree-ssa-forwprop.c (simplify_rotate): New function.
17341 (ssa_forward_propagate_and_combine): Call it.
17342
afb119be
RB		 173432013-05-10 Richard Biener <rguenther@suse.de>
17344
17345 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17346 disable peeling when we version for aliasing.
17347 (vector_alignment_reachable_p): Honor explicit user alignment.
17348 (vect_supportable_dr_alignment): Likewise.
17349 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17350 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17351 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17352 then peeling to arrange for the cost-model check to come first.
17353
01ae4861
AM		 173542013-05-10 Alan Modra <amodra@gmail.com>
17355
17356 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17357 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17358 * configure: Regenerate.
17359
ebc9a431
AM		 173602013-05-10 Alan Modra <amodra@gmail.com>
17361
17362 PR target/55033
17363 * varasm.c (default_elf_select_section): Move !DECL_P check..
17364 (get_named_section): ..to here before calling get_section_name.
17365 Adjust assertion.
17366 (default_section_type_flags): Add DECL_P check.
17367 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17368 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17369
d4bca93c
JR		 173702013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17371
17372 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17373 When using gen_stack_adjust_str with a register offset, add a
17374 REG_FRAME_RELATED_EXPR note.
17375
60ca9a65
UB		 173762013-05-09 Uros Bizjak <ubizjak@gmail.com>
17377
17378 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17379 (*vec_extractv4si_zext_mem): Ditto.
17380 (*vec_extractv2di): Add 0->x and x->x alternatives.
17381 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17382 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17383
a3409c02
JM		 173842013-05-09 Jason Merrill <jason@redhat.com>
17385
0138d6b2
JM		 17386 N3639 C++1y VLA support
17387 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17388 DECL_VALUE_EXPR.
17389
a3409c02
JM		 17390 * tree.c (build_constructor_va): New.
17391 * tree.h: Declare it.
17392
66e6b990
MJ		 173932013-05-09 Martin Jambor <mjambor@suse.cz>
17394
17395 PR lto/57084
17396 * gimple-fold.c (canonicalize_constructor_val): Call
17397 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17398
64cfa6c0
JH		 173992013-05-09 Jan Hubicka <jh@suse.cz>
17400 Richard Biener <rguenther@suse.de>
17401
17402 PR lto/54095
17403 * symtab.c (symtab_make_decl_local): Do not add private names.
17404
c3167b00
JH		 174052013-05-09 Jan Hubicka <jh@suse.cz>
17406
17407 PR lto/54095
17408 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17409 (unlink_from_assembler_name_hash): Likewise.
17410 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17411 symtab_unregister_node, symtab_initialize_asm_name_hash,
17412 change_decl_assembler_name): Update.
17413
12dc6974
SN		 174142013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17415
17416 * config/aarch64/aarch64.md: New movtf split.
17417 (*movtf_aarch64): Update.
17418 (aarch64_movdi_tilow): Handle TF modes and rename to
17419 aarch64_movdi_<mode>low.
17420 (aarch64_movdi_tihigh): Handle TF modes and rename to
17421 aarch64_movdi_<mode>high
17422 (aarch64_movtihigh_di): Handle TF modes and rename to
17423 aarch64_mov<mode>high_di
17424 (aarch64_movtilow_di): Handle TF modes and rename to
17425 aarch64_mov<mode>low_di
17426 (aarch64_movtilow_tilow): Remove spurious whitespace.
17427 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17428 splits.
17429 (aarch64_print_operand): Update.
17430
227eb343
AM		 174312013-05-09 Alan Modra <amodra@gmail.com>
17432
17433 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17434 powerpc64le.
17435 * configure: Regenerate.
17436
0b013847
UB		 174372013-05-08 Uros Bizjak <ubizjak@gmail.com>
17438
17439 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17440 splitter preparation statements.
17441 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17442 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17443 adjust_address_nv.
17444
1dc3d6e9
BS		 174452013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17446
17447 * gimple-ssa-strength-reduction.c (count_candidates): Change
17448 return value to int.
17449 (analyze_candidates_and_replace): Change type of length to int.
17450
e61e7d28
UB		 174512013-05-08 Uros Bizjak <ubizjak@gmail.com>
17452
17453 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17454 (*vec_extract<mode>): Use VI12_128 mode iterator.
17455 (*vec_extract<mode>_mem): Ditto.
17456 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17457 attribute.
17458
4fbfcf44
DN		 174592013-05-08 Diego Novillo <dnovillo@google.com>
17460
17461 PR bootstrap/54659
17462
17463 Revert:
4fbfcf44
DN		 17464 2012-08-17 Diego Novillo <dnovillo@google.com>
17465
e61e7d28
UB		 17466 PR bootstrap/54281
17467 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17468 * config.in: Regenerate.
17469 * configure: Regenerate.
17470 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17471
702d8703
JH		 174722013-05-08 Jan Hubicka <jh@suse.cz>
17473
17474 PR lto/54095
17475 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17476 * cgraph.h (symtab_node_base): Add unique_name.
17477 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17478 input_overwrite_node, input_varpool_node): Stream unique_name.
17479 * cgraphclones.c (cgraph_create_virtual_clone,
17480 cgraph_function_versioning): Set unique_name.
17481 * ipa.c (function_and_variable_visibility): Set unique_name.
17482
8b28cf47
BS		 174832013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17484
17485 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17486 (alloc_cand_and_find_basis): Restrict conditional candidate
17487 processing to CAND_MULTs.
17488
e86074fd
JH		 174892013-05-08 Jan Hubicka <jh@suse.cz>
17490
17491 PR lto/54095
17492 lto-symtab.c (lto_symtab_symbol_p): New function.
17493 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17494 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17495 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17496 Skip static symbols.
17497
44398cbe
PC		 174982013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17499
17500 PR tree-optimization/57200
17501 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17502 Only call inform if the preceding warning_at returns true.
17503
f6bc1c4a
HS		 175042013-05-07 Han Shen <shenhan@google.com>
17505
17506 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17507 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17508 * common.opt (fstack-protector-strong): New option.
17509 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17510 * doc/invoke.texi (Optimization Options): Document
17511 "-fstack-protector-strong".
17512 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17513
4ffecb1f
SB		 175142013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17515
17516 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17517
5a107a0f
VM		 175182013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17519
17520 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17521 it is not lhs of the insn.
17522 (setup_reg_equiv): Remove insn having equiv memory which it is not
17523 lhs of the insn.
17524 * lra-constraints.c (process_address): Try to improve generation
17525 code for address base + disp.
17526 (lra_constraints): Make correct the code for checking insn setting
17527 up backward equivalence. Remove insn only if it is in the init
17528 insn list.
17529 * lra-eliminations.c (update_reg_eliminate): Change return value.
17530 (lra_eliminate): Use the result.
17531
3f5783ea
UB		 175322013-05-07 Uros Bizjak <ubizjak@gmail.com>
17533
17534 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17535 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17536 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17537 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17538 register target operands.
17539 (*vec_extractv8hi_sse2): New pattern.
17540 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17541 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17542 (*vec_extract<mode>_mem): New insn and split pattern.
17543
8a5800b8
CL		 175442013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17545
17546 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17547 (TARGET_ASAN_SHADOW_OFFSET): Define.
17548 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17549 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17550
7bf55a70
BS		 175512013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17552
17553 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17554 (incr_vec_index): Return -1 if increment not found.
17555 (create_add_on_incoming_edge): Assert if increment not found.
17556 (record_increment): Limit number of increments recorded.
17557 (all_phi_incrs_profitable): Return false if an increment not found.
17558 (replace_profitable_candidates): Don't process increments that were
17559 not recorded.
17560 (analyze_candidates_and_replace): Limit size of incr_vec.
17561
3f8825c0
RB		 175622013-05-07 Richard Biener <rguenther@suse.de>
17563
17564 * calls.c (special_function_p): setjmp-like functions are leaf.
17565 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17566 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17567
fd4842cd
SN		 175682013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17569
17570 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17571 (aarch64_simd_mov<mode>): New expander.
17572 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17573 (aarch64_simd_mov_to_<mode>high): Likewise.
17574 (aarch64_simd_mov_from_<mode>low): Likewise.
17575 (aarch64_simd_mov_from_<mode>high): Likewise.
17576 (aarch64_dup_lane<mode>): Update.
17577 (aarch64_dup_lanedi): New instruction pattern.
17578 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17579 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17580
a7a7d10e
BS		 175812013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17582
17583 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17584 (replace_mult_candidate): Remove unnecessary argument; remove
17585 unnecessary parameter from call to introduce_cast_before_cand.
17586 (replace_unconditional_candidate): Remove unnecessary parameter
17587 from call to replace_mult_candidate.
17588 (replace_conditional_candidate): Likewise.
17589 (insert_initializers): Use make_temp_ssa_name.
17590 (introduce_cast_before_cand): Remove unnecessary argument; use
17591 make_temp_ssa_name.
17592 (replace_one_candidate): Remove unnecessary argument; remove
17593 unnecessary parameter from calls to introduce_cast_before_cand.
17594 (replace_profitable_candidates): Remove unnecessary parameters
17595 from calls to replace_one_candidate.
17596
29105868
BS		 175972013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17598
17599 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17600 phi def as possibly hiding a basis for a CAND_ADD whose operands
17601 have been commuted in the analysis.
17602 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17603
4095f9fa
N		 176042013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17605
17606 * config/aarch64/aarch64.md
17607 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17608 shift value between 0-4.
17609
39e843e8
RB		 176102013-05-07 Richard Biener <rguenther@suse.de>
17611
17612 * double-int.h (rshift): New overload.
17613 * double-int.c (rshift): New function.
17614 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17615 (create_reference_ops_from_ref): Remove.
17616 (vn_reference_insert): Use shared ops for constructing the
17617 reference and copy it.
17618
0a1a83cb
RB		 176192013-05-07 Richard Biener <rguenther@suse.de>
17620
17621 PR middle-end/57190
17622 * tree-eh.c (sink_clobbers): Properly propagate
17623 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17624
43303d6f
JJ		 176252013-05-07 Jakub Jelinek <jakub@redhat.com>
17626
ba7e83f8
JJ		 17627 PR tree-optimization/57149
17628 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17629 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17630 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17631 uninit_undefined_value_p instead of ssa_undefined_value_p.
17632
43303d6f
JJ		 17633 PR debug/57184
17634 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17635 for modifier == EXPAND_INITIALIZER.
17636
14523c25
AB		 176372013-05-07 Anton Blanchard <anton@samba.org>
17638
17639 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17640 for powerpc64 little endian.
17641 * configure: Regenerate.
17642
cb7c8be9
GS		 176432013-05-06 Graham Stott <grahams@btinternet.com>
17644
17645 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17646 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17647 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17648 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17649
176502013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 17651
17652 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17653 codes which allow non-lvalues.
17654
9a0ee7b0
MG		 176552013-05-06 Marc Glisse <marc.glisse@inria.fr>
17656
17657 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17658 components are all 1s.
17659 (integer_minus_onep): New function.
17660 * tree.h (integer_minus_onep): Declare it.
17661 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17662 integer_minus_onep instead of integer_all_onesp.
17663
f2c17ea9
OE		 176642013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17665
17666 PR target/52933
17667 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17668 variations of these patterns.
17669
f75e6a51
UB		 176702013-05-06 Uros Bizjak <ubizjak@gmail.com>
17671
17672 * config/i386/i386.md (isa): Add x64_sse4 member.
17673 (enabled): Handle x64_sse4.
17674 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17675 instruction for 64bit SSE4_1 targets. Update insn attributes.
17676 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17677 instruction for SSE4_1 targets. Update insn attributes.
17678 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17679 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17680 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17681 const_1 selector.
17682 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17683 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17684 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17685
a986d468
OE		 176862013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17687
17688 PR target/57108
17689 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17690
382522cb
MK		 176912013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17692
17693 * final.c (do_assembler_dialects): Don't handle curly braces and
17694 vertical bar escaped by % as dialect delimiters.
17695 (output_asm_insn): Print curly braces and vertical bar if escaped
17696 by % and ASSEMBLER_DIALECT defined.
17697 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17698 * doc/tm.texi: Regenerated.
17699
fb0d5c60
SB		 177002013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17701
fb0d5c60
SB		 17702 * config/mips/mips.c: Include tree-pass.h.
17703 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17704 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17705 (pass_mips_machine_reorg2): New machine specific pass.
17706 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17707 (mips_option_override): Register the new pass.
17708 * rtl.h (cleanup_barriers): Remove prototype.
17709 (dbr_schedule): Likewise.
17710 * jump.c (cleanup_barriers): Make static.
17711 * reorg.c (dbr_schedule): Likewise.
17712
aa06a978
RB		 177132013-05-06 Richard Biener <rguenther@suse.de>
17714
17715 PR tree-optimization/57185
17716 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17717 SSA names properly.
17718
0b953bec
UB		 177192013-05-06 Uros Bizjak <ubizjak@gmail.com>
17720
17721 PR target/57106
17722 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17723 between "first_arg" and "insn", not "last" and "insn".
17724
28708525
WS		 177252013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17726
17727 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17728 (find_candidates_in_block): Re-enable slsr_process_phi.
17729 (create_phi_basis): Fix double counting of candidate adjustment.
17730
0107dca2
RB		 177312013-05-06 Richard Biener <rguenther@suse.de>
17732
17733 PR middle-end/57147
17734 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17735 the edge is also fallthru, preserve it and just clear the
17736 abnormal flag.
17737 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17738 also complex, preserve that and just clear the fallthru flag.
17739 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17740 update virtual operands.
17741
470d4d13
AM		 177422013-05-06 Alan Modra <amodra@gmail.com>
17743
17744 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17745 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17746 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17747 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17748 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17749 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17750
8f1dbf8d
AM		 177512013-05-06 Alan Modra <amodra@gmail.com>
17752
17753 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17754 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17755 (DEFAULT_ASM_ENDIAN): Define.
17756 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17757 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17758 Update -K PIC clause from sysv4.h.
17759 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17760 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17761
54890767
AM		 177622013-05-06 Alan Modra <amodra@gmail.com>
17763
17764 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17765 twice for little-endian.
17766 (ashrdi3_no_power, ashrdi3): Support little-endian.
17767
2353515d
OE		 177682013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17769
17770 PR target/55303
17771 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17772 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17773 related expanders.
17774 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17775 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17776 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17777 New predicates.
17778
33e67557
SB		 177792013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17780 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17781
17782 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17783 * config/pa/pa.opt: Make mbig-switch a no-op.
17784 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17785 (CASE_VECTOR_MODE): Always return SImode.
17786 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17787 for the !TARGET_BIG_SWITCH case.
17788 * config/pa/pa-linux.h: Likewise.
17789 * config/pa/pa-openbsd.h: Likewise.
17790 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17791 * config/pa/pa.md (short_jump): Remove define_insn.
17792 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17793 (casesi0): Remove define_insn.
17794 (type): Remove btable_branch.
17795 (pa_combine_type): Likewise.
17796 (in_nullified_branch_delay): Likewise.
17797 (in_call_delay): Likewise.
17798 (define_delay): Likewise.
17799 (define_insn_reservation "Z3"): Likewise.
17800 (define_insn_reservation "Z4"): Likewise.
17801 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17802 (pa_adjust_insn_length): Remove adjustment for btable branches.
17803 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17804 and mno-big-switch
17805
3095685e
UB		 178062013-05-05 Uros Bizjak <ubizjak@gmail.com>
17807
17808 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17809 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17810 Add m->r,x alternatives.
17811 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17812 splitters using SWI48x mode iterator.
17813 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17814 TARGET_64BIT. Add m->x alternative.
17815 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17816 Add o->x alternative. Enable for TARGET_SSE.
17817 (sse_storeq): Remove expander.
17818 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17819 with memory input operand.
17820 (*vec_extractv2di_1 splitter): New.
17821 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17822 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17823
4b36ae28
SB		 178242013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17825
17826 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17827 (INT_LOWPART): Delete.
17828 (extract_MB): Adjust.
17829 (extract_ME): Adjust.
17830 (print_operand): Adjust.
17831
da226db2
SB		 178322013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17833
17834 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17835 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17836 (reg_or_logical_cint_operand, easy_fp_constant,
17837 logical_const_operand): Delete "CONST_DOUBLE" case.
17838 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17839 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 17840 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 17841 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17842 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17843 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17844 test.
17845 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17846 CONST_DOUBLE DImode/VOIDmode case.
17847 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17848 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17849 CONST_DOUBLE VOIDmode case.
17850 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17851 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17852 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17853 Delete CONST_DOUBLE case.
17854 (splitters for mov FMOVE64 const_double): Delete
17855 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17856 "HOST_BITS_PER_WIDE_INT >= 64" test.
17857 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17858 case.
17859 (mov DI const_double): Delete.
17860
40de22d6
JJ		 178612013-05-04 Jakub Jelinek <jakub@redhat.com>
17862
17863 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17864 on op shows all bits zero in mode of a lowpart subreg, return zero.
17865
5ec6aff2
MM		 178662013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17867
17868 PR target/57150
17869 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17870 to save TFmode registers and DImode to save TImode registers for
17871 caller save operations.
17872 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17873 mark being partially clobbered since they only use the first
17874 double word.
17875
17876 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17877 and TDmode only use the upper 64-bits of each VSX register.
17878
2cefad90
BS		 178792013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17880
17881 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17882 (find_candidates_in_block): Disable slsr_process_phi.
17883
d6d7eee1
GW		 178842013-05-03 Guozhi Wei <carrot@google.com>
17885
17886 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 17887 constructor to ...
17888 (build_init_ctor): ... here.
d6d7eee1 17889
9b92d12b
BS		 178902013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17891
17892 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17893 (slsr_cand_d): Redefine def_phi.
17894 (stride_status, phi_adjust_status, count_phis_status): New enums.
17895 (find_phi_def): New.
17896 (find_basis_for_base_expr): New.
17897 (find_basis_for_candidate): Handle hidden bases.
17898 (alloc_cand_and_find_basis): Handle phi candidates.
17899 (slsr_process_phi): New.
17900 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17901 (create_mul_imm_cand): Likewise.
17902 (create_add_ssa_cand): Exclude phi base candidates.
17903 (create_add_imm_cand): Likewise.
17904 (slsr_process_cast): Likewise.
17905 (slsr_process_copy): Likewise.
17906 (find_candidates_in_block): Handle phi candidates.
17907 (dump_candidate): Likewise.
17908 (unconditional_cands): Delete.
17909 (unconditional_cands_with_known_stride_p): Delete.
17910 (phi_dependent_cand_p): New.
17911 (cand_increment): Handle phi-dependent candidates.
17912 (replace_dependent): Delete.
17913 (replace_mult_candidate): New.
17914 (replace_unconditional_candidate): New.
17915 (incr_vec_index): Move to avoid forward reference.
17916 (create_add_on_incoming_edge): New.
17917 (create_phi_basis): New.
17918 (replace_dependents): Delete.
17919 (replace_conditional_candidate): New.
17920 (phi_add_costs): New.
17921 (replace_uncond_cands_and_profitable_phis): New.
17922 (record_increment): Handle phi adjustments.
17923 (record_phi_increments): New.
17924 (record_increments): Handle phi adjustments.
17925 (phi_incr_cost): New.
17926 (lowest_cost_path): Handle phis.
17927 (total_savings): Likewise.
17928 (analyze_increments): Likewise.
17929 (ncd_with_phi): New.
17930 (ncd_of_cand_and_phis): New.
17931 (nearest_common_dominator_for_cands): Handle phi increments.
17932 (all_phi_incrs_profitable): New.
17933 (replace_profitable_candidates): Handle phi-dependent candidates.
17934 (analyze_candidates_and_replace): Likewise.
17935
68f073d4
TJ		 179362013-05-03 Teresa Johnson <tejohnson@google.com>
17937
17938 PR bootstrap/57154
17939 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17940 do not exceed REG_BR_PROB_BASE.
17941
a4ee7cb9
JL		 179422013-05-03 Jeff Law <law@redhat.com>
17943
ade67f70 17944 PR tree-optimization/57144
a4ee7cb9
JL		 17945 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17946 operand of the condition will bit into the new type when eliminating
17947 a cast feeding a condition.
17948
47954c4e
JJ		 179492013-05-03 Jakub Jelinek <jakub@redhat.com>
17950
17951 PR rtl-optimization/57130
3095685e
UB		 17952 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17953 of COMPARE as in_code to the recursive call if needed.
47954c4e 17954
3c21604f
UB		 179552013-05-03 Uros Bizjak <ubizjak@gmail.com>
17956
17957 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17958 (enabled): Handle new members.
17959 * config/i386/sse.md (*vec_concatv2si): Merge from
17960 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17961 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17962
12211b99 179632013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 17964
17965 PR tree-optimization/57027
17966 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17967 for fnms opportunity, check we got the prerequisite kind
17968 of tree / gimple before using accessor functions.
17969
07bfc9ec
RB		 179702013-05-03 Richard Biener <rguenther@suse.de>
17971
17972 * double-int.h (lshift): New overload without precision
17973 and arith argument.
17974 (operator *=, operator +=, operator -=): Move ...
17975 * double-int.c (operator *=, operator +=, operator -=): ... here
17976 and implement more efficiently.
17977 (mul_double_with_sign): Remove.
17978 (lshift_double): Adjust to take unsinged shift argument, push
17979 dispatching code to callers.
17980 (mul_double_wide_with_sign): Add early out for callers that
17981 are not interested in high parts or overflow.
17982 (lshift): New function.
17983 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17984 dispatch code here.
17985 (lrotate, rrotate): Use logical shifts.
17986 * expr.c (get_inner_reference): Use lshift.
17987 * fixed-value.c (do_fixed_divide): Likewise.
17988 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17989 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17990 (indirect_refs_may_alias_p): Likewise.
17991 (stmt_kills_ref_p_1): Likewise.
17992
7769bb64
VP		 179932013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17994
17995 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17996
1d0c8e5c
VP		 179972013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17998
17999 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18000 scalar form of FABD instruction.
18001
f15feaf9
VM		 180022013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18003
18004 * lra-constraints.c (process_alt_operands): Add checking alt
18005 number to choose the best alternative.
18006
d90e76d4
RB		 180072013-05-02 Richard Biener <rguenther@suse.de>
18008
18009 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18010 bitmap and its handling.
18011 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18012
9f8e7a96
RB		 180132013-05-02 Richard Biener <rguenther@suse.de>
18014
18015 PR middle-end/57140
18016 * tree-inline.c (copy_loops): Properly handle removed loops.
18017 (copy_cfg_body): Mark destination loops for fixup if source
18018 loops needed fixup.
18019
f3a81b39
GY		 180202013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18021
18022 PR target/56732
18023 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18024 generating simple_return for naked functions.
18025
7b920a9a
MJ		 180262013-05-02 Martin Jambor <mjambor@suse.cz>
18027
18028 PR middle-end/56988
18029 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18030 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18031 flags match.
18032 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18033 ipa_agg_replacement_value structures.
18034 (known_aggs_to_agg_replacement_list): Likewise.
18035 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18036 (read_agg_replacement_chain): Likewise.
18037 (ipcp_transform_function): Also check that by_ref flags match.
18038
2c41c19d
RB		 180392013-05-02 Richard Biener <rguenther@suse.de>
18040
18041 * graphds.h (struct graph): Add obstack member.
18042 * graphds.c (new_graph): Initialize obstack and allocate
18043 vertices from it.
18044 (add_edge): Allocate edge from the obstack.
3c21604f 18045 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 18046
8b47039c
TJ		 180472013-05-02 Teresa Johnson <tejohnson@google.com>
18048
18049 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18050 divides.
18051 * cfg.c (update_bb_profile_for_threading): Ditto.
18052 * tree-inline.c (copy_bb): Ditto.
18053 (copy_edges_for_bb): Ditto.
18054 (initialize_cfun): Ditto.
18055 (copy_cfg_body): Ditto.
18056 (expand_call_inline): Ditto.
18057 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18058 (estimate_node_size_and_time): Ditto.
18059 (inline_merge_summary): Ditto.
18060 * cgraphclones.c (cgraph_clone_edge): Ditto.
18061 (cgraph_clone_node): Ditto.
18062 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18063 (compute_trg_info): Ditto.
18064
da65928c
IB		 180652013-05-02 Ian Bolton <ian.bolton@arm.com>
18066
18067 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18068 S reg when fp attribute set.
18069 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18070
add3c965
IB		 180712013-05-02 Ian Bolton <ian.bolton@arm.com>
18072
18073 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18074 New pattern.
18075 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18076 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18077 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18078
fdd43ac4
RB		 180792013-05-02 Richard Biener <rguenther@suse.de>
18080
18081 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18082 (struct instantiate_cache_entry): New type.
18083 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18084 (struct instantiate_cache_type): New type.
18085 (set_instantiated_value, get_instantiated_value): Remove.
18086 (get_instantiated_value_entry): New function.
18087 (instantiate_scev_name): Use the new cache and adjust.
18088 (instantiate_scev_poly): Adjust.
18089 (instantiate_scev_binary): Likewise.
18090 (instantiate_array_ref): Likewise.
18091 (instantiate_scev_convert): Likewise.
18092 (instantiate_scev_not): Likewise.
18093 (instantiate_scev_3): Likewise.
18094 (instantiate_scev_2): Likewise.
18095 (instantiate_scev_r): Likewise.
18096 (instantiate_scev): Likewise.
18097 (resolve_mixers): Likewise.
18098
36ff9dfb
VM		 180992013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18100
18101 PR target/57091
18102 * lra-constraints.c (best_small_class_operands_num): Remove.
18103 (process_alt_operands): Remove small_class_operands_num. Take
18104 small classes operands into losers and only if the operand is not
18105 matched. Modify debugging output.
18106 (curr_insn_transform): Remove best_small_class_operands_num.
18107 Print insn name.
18108
36054fab
JG		 181092013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18110
18111 * config/aarch64/aarch64-builtins.c
18112 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18113 * config/aarch64/aarch64-simd-builtins.def
18114 (reduc_splus_): Add new modes.
18115 (reduc_uplus_): New.
18116 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18117 (reduc_uplus_v4sf): Likewise.
18118 (reduc_splus_v4sf): Likewise.
18119 (aarch64_addv<mode>): Likewise.
18120 (reduc_uplus_<mode>): Likewise.
18121 (reduc_splus_<mode>): Likewise.
18122 (aarch64_addvv2di): Likewise.
18123 (reduc_uplus_v2di): Likewise.
18124 (reduc_splus_v2di): Likewise.
18125 (aarch64_addvv2si): Likewise.
18126 (reduc_uplus_v2si): Likewise.
18127 (reduc_splus_v2si): Likewise.
18128 (reduc_<sur>plus_<mode>): New.
18129 (reduc_<sur>plus_v2di): Likewise.
18130 (reduc_<sur>plus_v2si): Likewise.
18131 (reduc_<sur>plus_v4sf): Likewise.
18132 (aarch64_addpv4sf): Likewise.
18133 * config/aarch64/arm_neon.h
18134 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18135 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18136 add UNSPEC_SADDV, UNSPEC_UADDV.
18137 (SUADDV): New.
18138 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18139
6dce23a8
JG		 181402013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18141
18142 * config/aarch64/arm_neon.h
18143 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18144
1598945b
JG		 181452013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18146
18147 * config/aarch64/aarch64-builtins
18148 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18149
998eaf97
JG		 181502013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18151
18152 * config/aarch64/aarch64-simd-builtins.def
18153 (reduc_smax_): New.
18154 (reduc_smin_): Likewise.
18155 (reduc_umax_): Likewise.
18156 (reduc_umin_): Likewise.
18157 (reduc_smax_nan_): Likewise.
18158 (reduc_smin_nan_): Likewise.
18159 (fmax): Remove.
18160 (fmin): Likewise.
18161 (smax): Update for V2SF, V4SF and V2DF modes.
18162 (smin): Likewise.
18163 (smax_nan): New.
18164 (smin_nan): Likewise.
18165 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18166 (<su><maxmin><mode>3): ...This, refactor.
18167 (s<maxmin><mode>3): New.
18168 (<maxmin_uns><mode>3): Likewise.
18169 (reduc_<maxmin_uns>_<mode>): Refactor.
18170 (reduc_<maxmin_uns>_v4sf): Likewise.
18171 (reduc_<maxmin_uns>_v2si): Likewise.
18172 (aarch64_<fmaxmin><mode>: Remove.
18173 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18174 new builtin names.
18175 (vmin<q>_f<32,64>): Likewise.
18176 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18177 (FMAXMIN): New.
18178 (su): Add mappings for smax, smin, umax, umin.
18179 (maxmin): New.
18180 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18181 (FMAXMIN): Rename as...
18182 (FMAXMIN_UNS): ...This.
18183 (maxminv): Remove.
18184 (fmaxminv): Likewise.
18185 (fmaxmin): Likewise.
18186 (maxmin_uns): New.
18187 (maxmin_uns_op): Likewise.
18188
bd11644e
JG		 181892013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18190
18191 * config/aarch64/arm_neon.h
18192 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18193 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18194 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18195
75dd5ace
JG		 181962013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18197
18198 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18199 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18200
7c19979f
JG		 182012013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18202
18203 * config/aarch64/aarch64-simd.md
18204 (vcond<mode>_internal): Handle special cases for constant masks.
18205 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18206 (vcondu<mode><mode>): Likewise.
18207 (vcond<v_cmp_result><mode>): New.
18208
bb60efd9
JG		 182092013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18210
18211 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18212 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18213 * config/aarch64/aarch64-simd-builtins.def
18214 (cmeq): Update to BUILTIN_VALLDI.
18215 (cmgt): Likewise.
18216 (cmge): Likewise.
18217 (cmle): Likewise.
18218 (cmlt): Likewise.
18219 * config/aarch64/arm_neon.h
18220 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18221 to builtins or C as appropriate.
18222
889b9412
JG		 182232013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18224
18225 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18226 (cmgeu): ...This.
18227 (cmhi): Rename to...
18228 (cmgtu): ...This.
18229 * config/aarch64/aarch64-simd.md
18230 (simd_mode): Add SF.
18231 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18232 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18233 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18234 (cstore<mode>_neg): ...This.
18235 * config/aarch64/iterators.md
18236 (VALLF): new.
18237 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18238 (COMPARISONS): New.
18239 (UCOMPARISONS): Likewise.
18240 (optab): Add missing comparisons.
18241 (n_optab): New.
18242 (cmp_1): Likewise.
18243 (cmp_2): Likewise.
18244 (CMP): Likewise.
18245 (cmp): Remove.
18246 (VCMP_S): Likewise.
18247 (VCMP_U): Likewise.
18248 (V_cmp_result): Add DF, SF modes.
18249 (v_cmp_result): Likewise.
18250 (v): Likewise.
18251 (vmtype): Likewise.
18252 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18253
0a7dbb76
GY		 182542013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18255
18256 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18257 define_insn to define_insn_and_split.
18258 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18259 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18260 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18261 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18262 (thumb2_negscc): Likewise.
18263
fb614ca6
GY		 182642013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18265
18266 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18267
9e64a0bf
GY		 182682013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18269
18270 * config/arm/thumb2.md: Remove trailing whitespaces.
18271
d6b28156
RS		 182722013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18273
18274 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18275 Use gen_int_mode rather than GEN_INT.
18276
f91674c3
L		 182772013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18278
b0dec607 18279 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 18280 debug_gimple_stmt.
18281
3551257c
RB		 182822013-04-30 Richard Biener <rguenther@suse.de>
18283
18284 PR middle-end/57122
3c21604f 18285 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 18286
4c1aff1c
RB		 182872013-04-30 Richard Biener <rguenther@suse.de>
18288
18289 PR middle-end/57107
18290 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18291
54b8379a
AB		 182922013-04-30 Andrey Belevantsev <abel@ispras.ru>
18293
18294 PR rtl-optimization/56957
18295 PR rtl-optimization/57105
54b8379a
AB		 18296 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18297 variable. Use just INSN_UID for determining whether an insn
18298 should be only disconnected from the insn stream.
18299 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18300
abc27962
JJ		 183012013-04-30 Jakub Jelinek <jakub@redhat.com>
18302
18303 PR tree-optimization/57104
18304 * tsan.c (instrument_expr): Don't instrument accesses to
18305 DECL_HARD_REGISTER VAR_DECLs.
18306
0fc822d0
RB		 183072013-04-30 Richard Biener <rguenther@suse.de>
18308
18309 * function.h (loops_for_fn): New inline function.
18310 (set_loops_for_fn): Likewise.
18311 * cfgloop.h (place_new_loop): Add struct function parameter.
18312 (get_loop): Likewise.
18313 (get_loops): Likewise.
18314 (number_of_loops): Likewise.
18315 (fel_next): Adjust.
18316 (fel_init): Likewise.
18317 * cfg.c (get_loop_copy): Adjust.
18318 * cfgloop.c (flow_loops_dump): Likewise.
18319 (record_loop_exits): Likewise.
18320 (verify_loop_structure): Likewise.
18321 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18322 (estimate_reg_pressure_cost): Likewise.
18323 (mark_loop_exit_edges): Likewise.
18324 * cfgloopmanip.c (place_new_loop): Likewise.
18325 (add_loop): Likewise.
18326 (duplicate_loop): Likewise.
18327 * graph.c (draw_cfg_nodes): Likewise.
18328 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18329 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18330 (extract_affine_chrec): Likewise.
18331 (build_scop_iteration_domain): Likewise.
18332 * graphite.c (graphite_initialize): Likewise.
18333 * ira-build.c (create_loop_tree_nodes): Likewise.
18334 (more_one_region_p): Likewise.
18335 (rebuild_regno_allocno_maps): Likewise.
18336 (mark_loops_for_removal): Likewise.
18337 (mark_all_loops_for_removal): Likewise.
18338 (remove_unnecessary_regions): Likewise.
18339 (ira_build): Likewise.
18340 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18341 * loop-init.c (fix_loop_structure): Likewise.
18342 (gate_rtl_move_loop_invariants): Likewise.
18343 (gate_rtl_unswitch): Likewise.
18344 (gate_rtl_unroll_and_peel_loops): Likewise.
18345 (rtl_doloop): Likewise.
18346 * lto-streamer-in.c (input_cfg): Likewise.
18347 * lto-streamer-out.c (output_cfg): Likewise.
18348 * modulo-sched.c (sms_schedule): Likewise.
18349 * predict.c (tree_estimate_probability): Likewise.
18350 (tree_estimate_probability_driver): Likewise.
18351 (estimate_loops): Likewise.
18352 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18353 (move_sese_region_to_fn): Likewise.
18354 (debug_loop_num): Likewise.
18355 * tree-chrec.c (chrec_evaluate): Likewise.
18356 (hide_evolution_in_other_loops_than_loop): Likewise.
18357 (chrec_component_in_loop_num): Likewise.
18358 (reset_evolution_in_loop): Likewise.
18359 (evolution_function_is_invariant_rec_p): Likewise.
18360 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18361 * tree-inline.c (copy_loops): Likewise.
18362 (copy_cfg_body): Likewise.
18363 (tree_function_versioning): Likewise.
18364 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18365 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18366 Likewise.
18367 (add_to_evolution_1): Likewise.
18368 (scev_const_prop): Likewise.
18369 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18370 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18371 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18372 (tree_ssa_lim_initialize): Likewise.
18373 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18374 (verify_loop_closed_ssa): Likewise.
18375 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18376 (tree_ssa_loop_im): Likewise.
18377 (tree_ssa_loop_unswitch): Likewise.
18378 (tree_vectorize): Likewise.
18379 (check_data_deps): Likewise.
18380 (tree_ssa_loop_ivcanon): Likewise.
18381 (tree_ssa_loop_bounds): Likewise.
18382 (tree_complete_unroll): Likewise.
18383 (tree_complete_unroll_inner): Likewise.
18384 (tree_parallelize_loops): Likewise.
18385 (tree_ssa_loop_prefetch): Likewise.
18386 (tree_ssa_loop_ivopts): Likewise.
18387 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18388 * tree-vectorizer.c (vectorize_loops): Likewise.
18389
37953bd3
MF		 183902013-04-29 Mike Frysinger <vapier@gentoo.org>
18391
18392 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18393 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18394 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18395 with EABI_LINK_SPEC.
18396
f9ed28db
UB		 183972013-04-29 Uros Bizjak <ubizjak@gmail.com>
18398
18399 PR target/44578
18400 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18401 alternative.
18402
deca73f5
VM		 184032013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18404
18405 PR target/57097
37953bd3 18406 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 18407 using memory for pseudos. Print cost dump for alternatives.
18408 Modify cost values for conflicts with early clobbers.
18409 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18410
fdca7d03
UB		 184112013-04-29 Uros Bizjak <ubizjak@gmail.com>
18412
18413 PR target/57098
18414 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18415
0b064172
IB		 184162013-04-29 Ian Bolton <ian.bolton@arm.com>
18417
18418 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18419 from/to S register.
18420 (movdi_aarch64): Support LDR/STR from/to D register.
18421
473cec55
IB		 184222013-04-29 Ian Bolton <ian.bolton@arm.com>
18423
18424 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18425 or higher by default.
18426
a6f30e66
RB		 184272013-04-29 Richard Biener <rguenther@suse.de>
18428
18429 PR middle-end/57075
18430 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18431 even if not adding abnormal edges for calls that can make
18432 abnormal gotos.
18433
0c2b2040
RB		 184342013-04-29 Richard Biener <rguenther@suse.de>
18435
18436 PR middle-end/57103
18437 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18438 TREE_BLOCK.
18439 (move_stmt_r): Remove redundant checking.
18440
f41f80f9
TJ		 184412013-04-29 Teresa Johnson <tejohnson@google.com>
18442
18443 PR bootstrap/57077
18444 * basic-block.h (apply_scale): New function.
18445 (apply_probability): Use apply_scale.
18446 * gimple-streamer-in.c (input_bb): Ditto.
18447 * lto-streamer-in.c (input_cfg): Ditto.
18448 * lto-cgraph.c (merge_profile_summaries): Ditto.
18449 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18450 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 18451 (copy_edges_for_bb): Ditto.
18452 (copy_cfg_body): Ditto.
18453
315bbd2e
TV		 184542013-04-29 Tom de Vries <tom@codesourcery.com>
18455
18456 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18457 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18458 (tail_merge_optimize): Handle current_loops == NULL.
18459
ebbd90d8
JL		 184602013-04-26 Jeff Law <law@redhat.com>
18461
18462 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18463 (simplify_cond_using_ranges): Generalize code to simplify
18464 COND_EXPRs where one argument is a constant and the other
18465 is an SSA_NAME created by an integral type conversion.
18466
8b9b57eb
KT		 184672013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18468
18469 * config/arm/arm.md (store_minmaxsi): Use only when
18470 optimize_insn_for_size_p.
18471
9498e5dc
CB		 184722013-04-29 Christian Bruel <christian.bruel@st.com>
18473
18474 PR target/57108
18475 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18476
6093bc06
RB		 184772013-04-29 Richard Biener <rguenther@suse.de>
18478
18479 PR middle-end/57089
fdca7d03
UB		 18480 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18481 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 18482 (expand_omp_for_generic): Properly add loops.
18483 (expand_omp_for_static_nochunk): Likewise.
18484 (expand_omp_for_static_chunk): Likewise.
18485 (expand_omp_for): For the degenerate case fixup loops.
18486 (expand_omp_sections): Fix default bb placement in loops.
18487 (expand_omp_atomic_pipeline): Properly add loops.
18488
84aacbfd
KT		 184892013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18490
18491 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18492
8dee4479
TV		 184932013-04-29 Tom de Vries <tom@codesourcery.com>
18494
18495 * tree-ssa-tail-merge.c: Update header comment.
18496
47934dc4
JG		 184972013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18498
18499 * config/aarch64/arm_neon.h
18500 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18501 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18502 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18503 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18504 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18505 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18506 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18507 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18508
384be29f
JG		 185092013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18510
18511 * config/aarch64/aarch64-simd.md
18512 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18513 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18514 fix_trunc, fixuns_trunc.
18515 (ftrunc<VDQF:mode>2): New.
18516 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18517 (fix_trunc_optab): New.
18518
0386b123
JG		 185192013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18520
18521 * config/aarch64/aarch64-builtins.c
18522 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18523 iceilf, lround, iroundf.
18524
00fcb892
UB		 185252013-04-29 Uros Bizjak <ubizjak@gmail.com>
18526
18527 PR target/54349
18528 * config/i386/i386.h (enum ix86_tune_indices)
18529 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18530 New, split from X86_TUNE_INTER_UNIT_MOVES.
18531 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18532 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18533 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18534 (TARGET_INTER_UNIT_MOVES): Remove.
18535 * config/i386/i386.c (initial_ix86_tune_features): Update.
18536 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18537 (ix86_expand_convert_uns_didf_sse): Use
18538 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18539 (ix86_expand_vector_init_one_nonzero): Ditto.
18540 (ix86_expand_vector_init_interleave): Ditto.
18541 (inline_secondary_memory_needed): Return true for moves from SSE class
18542 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18543 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18544 * config/i386/constraints.md (Yi, Ym): Depend on
18545 TARGET_INTER_UNIT_MOVES_TO_VEC.
18546 (Yj, Yn): New constraints.
18547 * config/i386/i386.md (*movdi_internal): Change constraints of
18548 operand 1 from Yi to Yj and from Ym to Yn.
18549 (*movsi_internal): Ditto.
18550 (*movdf_internal): Ditto.
18551 (*movsf_internal): Ditto.
18552 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18553 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18554 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18555 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18556 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18557 * config/i386/sse.md (movdi_to_sse): Ditto.
18558 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18559 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18560 TARGET_INTER_UNIT_MOVES.
18561 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18562 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18563 instead of TARGET_INTER_UNIT_MOVES.
18564 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18565 operand 1 from Yi to Yj and from Ym to Yn.
18566
4c871069
JG		 185672013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18568
18569 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18570 (float_truncate_hi_): Likewise.
18571 (float_extend_lo_): Likewise.
18572 (float_truncate_lo_): Likewise.
18573 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18574 (aarch64_float_extend_lo_v2df): Likewise.
18575 (vec_unpacks_hi_v4sf): Likewise.
18576 (aarch64_float_truncate_lo_v2sf): Likewise.
18577 (aarch64_float_truncate_hi_v4sf): Likewise.
18578 (vec_pack_trunc_v2df): Likewise.
18579 (vec_pack_trunc_df): Likewise.
18580
1709ff9b
JG		 185812013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18582
18583 * config/aarch64/aarch64-builtins.c
18584 (aarch64_fold_builtin): Fold float conversions.
18585 * config/aarch64/aarch64-simd-builtins.def
18586 (floatv2si, floatv4si, floatv2di): New.
18587 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18588 * config/aarch64/aarch64-simd.md
18589 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18590 * config/aarch64/iterators.md (FLOATUORS): New.
18591 (optab): Add float, floatuns.
18592 (su_optab): Likewise.
18593
ce966824
JG		 185942013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18595
18596 * config/aarch64/aarch64-builtins.c
18597 (aarch64_builtin_vectorized_function): Use new names for
18598 fcvt builtins.
18599 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18600 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18601 (fcvtzu): Split as...
18602 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18603 (fcvtas): Split as...
18604 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18605 (fcvtau): Split as...
18606 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18607 (fcvtps): Split as...
18608 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18609 (fcvtpu): Split as...
18610 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18611 (fcvtms): Split as...
18612 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18613 (fcvtmu): Split as...
18614 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18615 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18616 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18617 (lfrintnusf, lfrintnudf): Likewise.
18618 * config/aarch64/aarch64-simd.md
18619 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18620 define_insn.
18621 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18622 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18623 (fcvt_pattern): Likewise.
18624
b9de24fe
JG		 186252013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18626
18627 * config/aarch64/aarch64-simd.md
18628 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18629 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18630
77a205be
JG		 186312013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18632
18633 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18634 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18635 (vrnd<a,m,n,p>_f32): Implement using builtins.
18636 (vrnd<i,x><q>_f<32, 64>): New.
18637
0659ce6f
JG		 186382013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18639
18640 * config/aarch64/aarch64-builtins.c
18641 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18642 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18643 (frintz): Rename to...
18644 (btrunc): ...this.
18645 (frintp): Rename to...
18646 (ceil): ...this.
18647 (frintm): Rename to...
18648 (floor): ...this.
18649 (frinti): Rename to...
18650 (nearbyint): ...this.
18651 (frintx): Rename to...
18652 (rint): ...this.
18653 (frinta): Rename to...
18654 (round): ...this.
18655 * config/aarch64/aarch64-simd.md
18656 (aarch64_frint<frint_suffix><mode>): Delete.
18657 (<frint_pattern><mode>2): Convert to insn.
18658 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18659 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18660 (frint_pattern): Likewise.
18661 (frint_suffix): Likewise.
18662
ea78906a
RB		 186632013-04-29 Richard Biener <rguenther@suse.de>
18664
18665 PR tree-optimization/57081
18666 * loop-init.c: Include tree-flow.h.
18667 (loop_optimizer_finalize): Free number of iteration estimates.
18668 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18669
baee1763
JJ		 186702013-04-29 Jakub Jelinek <jakub@redhat.com>
18671
94dc5332
JJ		 18672 PR tree-optimization/57083
18673 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18674 non-singleton shift count range, zero extend low_bound for uns case.
18675
baee1763
JJ		 18676 * config/i386/predicates.md (general_vector_operand): New predicate.
18677 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18678 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18679 if they aren't nonimmediate operands. If their original values
18680 satisfy const_vector_equal_evenodd_p, don't shift them.
18681 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18682 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18683 if not nonimmediate_operand.
18684 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18685 instead of register_operand.
18686 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18687
a9073727 186882013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 18689
18690 * stor-layout.c (finalize_size_functions): Allocate a structure and
18691 reset cfun before dumping the functions.
18692
ba8011e6
JJ		 186932013-04-27 Jakub Jelinek <jakub@redhat.com>
18694
d6fde69e
JJ		 18695 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18696
ba8011e6
JJ		 18697 PR target/56866
18698 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18699 use xop_pmacsdqh if uns_p.
18700 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18701 the immediate rotate count.
18702
2c62cbaa
VM		 187032013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18704
18705 * rtl.h (struct rtx_def): Add comment for field jump.
18706 (LRA_SUBREG_P): New macro.
18707 * recog.c (register_operand): Check LRA_SUBREG_P.
18708 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18709 stack frame.
18710 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18711 (lra_final_code_change): Skip subreg change for operators.
18712 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18713 if there are no operand changes.
18714 * lra-constraints.c (curr_insn_set): New.
18715 (match_reload): Set LRA_SUBREG_P.
18716 (emit_spill_move): Ditto.
18717 (check_and_process_move): Use curr_insn_set. Process only single
18718 set insns. Don't initialize sec_mem_p and change_p.
18719 (simplify_operand_subreg): Use LRA_SUBREG_P.
18720 (reg_in_class_p): New function.
18721 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18722 of #ifdef. Add code to remove cycling.
18723 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18724 non-null disp. Reload inner instead of disp when base and index
18725 are null. Try to put lo_sum into register.
18726 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18727 (check_and_process_move): Move code for move cost check to
18728 simple_move_p. Remove equiv_substitution.
18729 (simple_move_p): New function.
18730 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18731 curr_insn_set. Call check_and_process_move only for single set
18732 insns. Use the new function. Move call of check_and_process_move
18733 after operand equiv substitution and address process.
18734
e7d764f3
JJ		 187352013-04-26 Jakub Jelinek <jakub@redhat.com>
18736
18737 PR go/57045
18738 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18739 with nonlocal goto receivers or returns twice calls, ignore
18740 unininitialized values from abnormal edges to nl goto receiver
18741 or returns twice call.
18742
41e10689
JJ		 187432013-04-26 Jakub Jelinek <jakub@redhat.com>
18744
18745 PR tree-optimization/57051
18746 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18747 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18748 bitsize.
18749
d7ed20db
RB		 187502013-04-26 Richard Biener <rguenther@suse.de>
18751
18752 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18753 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18754 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18755 (fixup_loop_arrays_after_move): New function.
18756 (move_sese_region_to_fn): Properly outline the loop tree parts
18757 of the SESE region.
18758
df93505e
UB		 187592013-04-26 Uros Bizjak <ubizjak@gmail.com>
18760
18761 * config/i386/i386.md (type, unit): Fix long lines.
18762
dd366ec3
RB		 187632013-04-26 Richard Biener <rguenther@suse.de>
18764
18765 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18766 (lto-streamer-out.o): Likewise.
18767 * cfgloop.c (init_loops_structure): Export, add struct function
18768 argument and adjust.
18769 (flow_loops_find): Adjust.
18770 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18771 (init_loops_structure): Declare.
18772 * lto-streamer-in.c: Include cfgloop.h.
18773 (input_cfg): Input the loop tree.
18774 * lto-streamer-out.c: Include cfgloop.h.
18775 (output_cfg): Output the loop tree.
18776 (output_struct_function_base): Do not drop PROP_loops.
18777
a9e0d843
RB		 187782013-03-26 Richard Biener <rguenther@suse.de>
18779
18780 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18781 (pass_build_cfg): Provide PROP_loops.
18782 (move_sese_region_to_fn): Remove loops that are outlined into fn
18783 for now.
18784 * tree-inline.c: Include cfgloop.h.
18785 (initialize_cfun): Do not drop PROP_loops.
18786 (copy_loops): New function.
18787 (copy_cfg_body): Copy loop structure.
18788 (tree_function_versioning): Initialize destination loop tree.
18789 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18790 (pass_parallelize_loops): Do IL verification.
18791 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18792 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18793 the CFG make sure we fixup loops as well.
18794 * tree-ssa-tail-merge.c: Include cfgloop.h.
18795 (replace_block_by): When merging loop latches mark loops for fixup.
18796 * lto-streamer-out.c (output_struct_function_base): Drop
18797 PROP_loops for now.
18798 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18799 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18800 * ipa-split.c: Include cfgloop.h.
18801 (split_function): Add the new return block to the loop tree root.
18802 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18803 whether we have removed the forwarder block.
18804 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18805 * cfgloop.h (place_new_loop): Declare.
18806 * cfgloopmanip.c (place_new_loop): Export.
18807 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18808 (tree-switch-conversion.o): Likewise.
18809 (tree-complex.o): Likewise.
18810 (tree-inline.o): Likewise.
18811 (tree-ssa-tailmerge.o): Likewise.
18812 (ipa-split.o): Likewise.
18813 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18814 (tree-ssa-copy.o): Likewise.
18815 * tree-switch-conversion.c: Include cfgloop.h
18816 (process_switch): If we emit a bit-test cascade, schedule loops
18817 for fixup.
18818 * tree-complex.c: Include cfgloop.h.
18819 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18820 * asan.c: Include cfgloop.h.
18821 (create_cond_insert_point): Properly add new basic-blocks to
18822 loops, schedule loop fixup.
18823 * cfgloop.c (verify_loop_structure): Check that looks are not
18824 marked for fixup.
18825 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18826 to loops.
18827 (expand_omp_for_generic): Likewise.
18828 (expand_omp_sections): Likewise.
18829 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18830 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18831 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18832 is initialized, not when loops are present.
18833 * tree-parloops.c (parallelize_loops): Remove checking here.
18834 * passes.c (init_optimization_passes): Schedule a copy-propagation
18835 pass before complete unrolling of inner loops.
18836
e78e8a0b
JJ		 188372013-04-26 Jakub Jelinek <jakub@redhat.com>
18838
a2e836b2
JJ		 18839 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18840 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18841 (colorize_init): Add argument to _WIN32 version.
18842 * toplev.c: Include diagnostic-color.h.
18843 (process_options): Default to -fdiagnostics-color=auto if
18844 GCC_COLORS env var is in the environment.
18845 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 18846 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18847 env var is in the environment, the default is auto rather than never.
a2e836b2 18848
e78e8a0b
JJ		 18849 * diagnostic.h (file_name_as_prefix): Add context argument.
18850 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18851 the string as locus.
18852 * langhooks.c (lhd_print_error_function): Adjust caller.
18853
013e5ef9
LC		 188542013-04-25 Lawrence Crowl <crowl@google.com>
18855
18856 * var-tracking.c (shared_hash_def::htab):
18857 Change type to hash_table. Update dependent calls and types.
18858
4a8fb1a1
LC		 188592013-04-25 Lawrence Crowl <crowl@google.com>
18860
18861 * Makefile.in: Update as needed below.
18862
18863 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18864 Move declaration to after the type's method definitons.
18865
18866 * attribs.c (htab_t scoped_attributes::attribute_hash):
18867 Change type to hash_table. Update dependent calls and types.
18868
18869 * bitmap.c (htab_t bitmap_desc_hash):
18870 Change type to hash_table. Update dependent calls and types.
18871
18872 * cselib.c (htab_t cselib_hash_table):
18873 Change type to hash_table. Update dependent calls and types.
18874
18875 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18876 (hash_string_slot_node): Move implementation into lto-streamer.h
18877 struct string_slot_hasher.
18878 (eq_string_slot_node): Likewise.
18879
18880 * data-streamer-out.c: Update output_block::string_hash_table
18881 dependent calls and types.
18882
18883 * dwarf2cfi.c (htab_t trace_index):
18884 Change type to hash_table. Update dependent calls and types.
18885
18886 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18887 Change type to hash_table. Update dependent calls and types.
18888 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18889 (htab_t optimize_external_refs::map): Likewise.
18890 (htab_t output_comp_unit::extern_map): Likewise.
18891 (htab_t output_comdat_type_unit::extern_map): Likewise.
18892 (htab_t output_macinfo::macinfo_htab): Likewise.
18893 (htab_t optimize_location_lists::htab): Likewise.
18894 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18895
18896 * except.c (htab_t ehspec_hash_type):
18897 Change type to hash_table. Update dependent calls and types.
18898 (assign_filter_values::ttypes): Likewise.
18899 (assign_filter_values::ehspec): Likewise.
18900 (sjlj_assign_call_site_values::ar_hash): Likewise.
18901 (convert_to_eh_region_ranges::ar_hash): Likewise.
18902
18903 * gcse.c (htab_t pre_ldst_table):
18904 Change type to hash_table. Update dependent calls and types.
18905
18906 * ggc-common.c (htab_t saving_htab):
18907 Change type to hash_table. Update dependent calls and types.
18908 (htab_t loc_hash): Likewise.
18909 (htab_t ptr_hash): Likewise.
18910 (call_count): Rename ggc_call_count.
18911 (call_alloc): Rename ggc_call_alloc.
18912 (loc_descriptor): Rename make_loc_descriptor.
18913 (add_statistics): Rename ggc_add_statistics.
18914
18915 * ggc-common.c (saving_htab):
18916 Change type to hash_table. Update dependent calls and types.
18917
18918 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18919 (push_gimplify_context): Likewise.
18920 (pop_gimplify_context): Likewise.
18921 (struct gimple_temp_hash_elt): Added.
18922 (struct gimplify_hasher): Likewise.
18923 (struct gimplify_ctx.temp_htab):
18924 Change type to hash_table. Update dependent calls and types.
18925
18926 * gimple-fold.c: Include gimplify-ctx.h.
18927
18928 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18929 Change type to hash_table. Update dependent calls and types.
18930 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18931 avoid potential global name collision.
18932
18933 * gimplify.c: Include gimplify-ctx.h.
18934 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18935 (htab_t gimplify_ctx::temp_htab):
18936 Update dependent calls and types for new type hash_table.
18937 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18938 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18939
18940 * gimplify-ctx.h: New.
18941 (struct gimple_temp_hash_elt): Move from gimplify.c.
18942 (class gimplify_hasher): New.
18943 (struct gimplify_ctx): Move from gimple.h.
18944 (htab_t gimplify_ctx::temp_htab):
18945 Change type to hash_table. Update dependent calls and types.
18946
18947 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18948 (htab_t ivs_params::newivs_index):
18949 Change type to hash_table. Update dependent calls and types.
18950 (htab_t ivs_params::params_index): Likewise.
18951 (htab_t print_generated_program::params_index): Likewise.
18952 (htab_t gloog::newivs_index): Likewise.
18953 (htab_t gloog::params_index): Likewise.
18954
18955 * graphite.c: Include graphite-htab.h.
18956 4htab_t graphite_transform_loops::bb_pbb_mapping):
18957 Change type to hash_table. Update dependent calls and types.
18958
18959 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18960 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18961 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18962
18963 * graphite-dependences.c: Include graphite-htab.h.
18964 (loop_is_parallel_p): Change hash table type of parameter.
18965
18966 * graphite-htab.h: New.
18967 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18968 (extern find_pbb_via_hash): Move from graphite-poly.h.
18969 (extern loop_is_parallel_p): Move from graphite-poly.h.
18970 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18971
18972 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18973 (extern loop_is_parallel_p): Move to graphite-htab.h.
18974 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18975
18976 * haifa-sched.c (htab_t delay_htab):
18977 Change type to hash_table. Update dependent calls and types.
18978 (htab_t delay_htab_i2): Likewise.
18979
18980 * ira-color.c (htab_t allocno_hard_regs_htab):
18981 Change type to hash_table. Update dependent calls and types.
18982
18983 * ira-costs.c (htab_t cost_classes_htab):
18984 Change type to hash_table. Update dependent calls and types.
18985
18986 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18987 Change type to hash_table. Update dependent calls and types.
18988
18989 * loop-iv.c (htab_t bivs):
18990 Change type to hash_table. Update dependent calls and types.
18991
18992 * loop-unroll.c (htab_t opt_info::insns_to_split):
18993 Change type to hash_table. Update dependent calls and types.
18994 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18995
18996 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18997 (struct string_slot_hasher): New.
18998 (htab_t output_block::string_hash_table):
18999 Change type to hash_table. Update dependent calls and types.
19000
19001 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19002 (htab_t file_name_hash_table):
19003 Change type to hash_table. Update dependent calls and types.
19004
19005 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19006 calls and types.
19007
19008 * lto-streamer.c (htab_t tree_htab):
19009 Change type to hash_table. Update dependent calls and types.
19010
19011 * omp-low.c: Include gimplify-ctx.h.
19012
19013 * passes.c (htab_t name_to_pass_map):
19014 Change type to hash_table. Update dependent calls and types.
19015 (pass_traverse): Rename to passes_pass_traverse.
19016
19017 * plugin.c (htab_t event_tab):
19018 Change type to hash_table. Update dependent calls and types.
19019
19020 * postreload-gcse.c (htab_t expr_table):
19021 Change type to hash_table. Update dependent calls and types.
19022 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19023
19024 * sese.c (debug_rename_map_1): Make extern.
19025 (htab_t copy_bb_and_scalar_dependences::rename_map):
19026 Change type to hash_table. Update dependent calls and types.
19027
19028 * sese.h (extern debug_rename_map): Move to .c file.
19029
19030 * store-motion.c (htab_t store_motion_mems_table):
19031 Change type to hash_table. Update dependent calls and types.
19032
19033 * trans-mem.c (htab_t tm_new_mem_hash):
19034 Change type to hash_table. Update dependent calls and types.
19035
19036 * tree-browser.c (htab_t TB_up_ht):
19037 Change type to hash_table. Update dependent calls and types.
19038
19039 * tree-cfg.c (htab_t discriminator_per_locus):
19040 Change type to hash_table. Update dependent calls and types.
19041
19042 * tree-complex.c: Include tree-hasher.h
19043 (htab_t complex_variable_components):
19044 Change type to hash_table. Update dependent calls and types.
19045
19046 * tree-eh.c (htab_t finally_tree):
19047 Change type to hash_table. Update dependent calls and types.
19048
19049 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19050 struct int_tree_hasher.
19051 (extern int_tree_map_eq): Likewise.
19052 (uid_decl_map_hash): Removed.
19053 (extern decl_tree_map_eq): Likewise.
19054
19055 * tree-hasher.h: New.
19056 (struct int_tree_hasher): New.
19057 (typedef int_tree_htab_type): New.
19058
19059 * tree-inline.c: Include gimplify-ctx.h.
19060
19061 * tree-mudflap.c: Include gimplify-ctx.h.
19062
19063 * tree-parloops.c: Include tree-hasher.h.
19064 (htab_t eliminate_local_variables_stmt::decl_address):
19065 Change type to hash_table. Update dependent calls and types.
19066 (htab_t separate_decls_in_region::decl_copies): Likewise.
19067
19068 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19069 Change type to hash_table. Update dependent calls and types.
19070
19071 * tree-sra.c (candidates):
19072 Change type to hash_table. Update dependent calls and types.
19073
19074 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19075 in tree-flow.h.
19076 (int_tree_map_hash): Likewise.
19077
19078 * tree-ssa-dom.c (htab_t avail_exprs):
19079 Change type to hash_table. Update dependent calls and types.
19080
19081 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19082 Change type to hash_table. Update dependent calls and types.
19083
19084 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19085 Change type to hash_table. Update dependent calls and types.
19086
19087 * tree-ssa-phiopt.c (seen_ssa_names):
19088 Change type to hash_table. Update dependent calls and types.
19089
19090 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19091 Change type to hash_table. Update dependent calls and types.
19092
19093 * tree-ssa-uncprop.c (equiv):
19094 Change type to hash_table. Update dependent calls and types.
19095
c5a44004
JJ		 190962013-04-25 Jakub Jelinek <jakub@redhat.com>
19097
19098 PR rtl-optimization/57003
19099 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19100 call note_stores with kill_clobbered_value callback again after
19101 killing regs_invalidated_by_call.
19102
09962a4a
JG		 191032013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19104
19105 * config/aarch64/aarch64-simd.md
19106 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19107 (aarch64_simd_bsl<mode>): Likewise.
19108 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19109
ea28bb0b
MP		 191102013-04-25 Marek Polacek <polacek@redhat.com>
19111
19112 PR tree-optimization/57066
3c21604f 19113 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 19114
96659611
JG		 191152013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19116
19117 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19118
9697e620
JG		 191192013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19120
19121 * config/aarch64/aarch64-builtins.c
19122 (aarch64_fold_builtin): New.
19123 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19124 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19125 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19126 * config/aarch64/arm_neon.h
19127 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19128
0ac198d3
JG		 191292013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19130 Tejas Belagod <tejas.belagod@arm.com>
19131
19132 * config/aarch64/aarch64-builtins.c
19133 (aarch64_gimple_fold_builtin): New.
19134 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19135 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19136 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19137 (addvv4sf): Update.
19138 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19139
58cff58c
N		 191402013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19141
df93505e 19142 * config/aarch64/aarch64.md
58cff58c
N		 19143 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19144
7e0228bf
N		 191452013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19146
19147 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19148 (*ngcsi_uxtw): New pattern.
19149
5819f96f 191502013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 19151 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 19152
19153 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19154 (TB_DREG): Add T_V4HF.
19155 (v4hf_UP): New macro.
19156 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 19157 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19158 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19159 built-ins.
df93505e 19160 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19161 (arm_vector_mode_supported_p): Handle V4HF.
19162 (arm_mangle_map): Handle V4HFmode.
19163 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19164 * config/arm/arm_neon_builtins.def: Add entries for
19165 vcvtv4hfv4sf, vcvtv4sfv4hf.
19166 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19167 (neon_vcvtv4hfv4sf): Likewise.
19168 * config/arm/neon-gen.ml: Handle half-precision floating point
19169 features.
19170 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19171 * config/arm/arm_neon.h: Regenerate.
19172 * config/arm/neon.ml (type elts): Add F16.
19173 (type vectype): Add T_float16x4, T_floatHF.
19174 (type vecmode): Add V4HF.
19175 (type features): Add Requires_FP_bit feature.
19176 (elt_width): Handle F16.
19177 (elt_class): Likewise.
19178 (elt_of_class_width): Likewise.
19179 (mode_of_elt): Refactor.
19180 (type_for_elt): Handle F16, fix error messages.
19181 (vectype_size): Handle T_float16x4.
19182 (vcvt_sh): New function.
19183 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19184 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19185 (string_of_mode): Handle V4HF.
19186 * doc/arm-neon-intrinsics.texi: Regenerate.
19187
1ef395e4
JG		 191882013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19189
19190 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19191 format specifier in 'X' case.
19192
41c34e94
AM		 191932013-04-25 Alan Modra <amodra@gmail.com>
19194
19195 PR target/57052
19196 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19197 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19198 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19199 Repeat for many other rotate/shift and mask patterns using subregs.
19200 Name lshiftrt insns.
19201 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19202 on WORDS_BIG_ENDIAN.
19203
b9a7eb5d
AM		 192042013-04-25 Alan Modra <amodra@gmail.com>
19205
19206 * config.gcc: Support little-endian powerpc-linux targets.
19207 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19208 (LINK_OS_LINUX_SPEC): Define.
19209 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19210 Preserve MASK_LITTLE_ENDIAN.
19211 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19212 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19213 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19214 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19215 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19216 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19217 Correct fp word order for little-endian. Don't shift toc entries
19218 smaller than a word for little-endian.
19219 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19220 (bswapdi2 splits): Correct low-part subreg for little-endian.
19221 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19222 low/high where such is correct only for be.
19223 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19224 little-endian for -mcall-aixdesc.
19225
87f73374
AM		 192262013-04-25 Alan Modra <amodra@gmail.com>
19227
19228 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19229 replace_equiv_address_nv.
19230
cabf91cd
AM		 192312013-04-25 Alan Modra <amodra@gmail.com>
19232
19233 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19234
0ae24cc8
VM		 192352013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19236
19237 Revert:
19238 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19239 * rtl.h (struct rtx_def): ...
cabf91cd 19240
77bce07c
VM		 192412013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19242
19243 PR rtl-optimizations/57046
19244 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19245 for multi-reg splits.
19246
0db63e7f
L		 192472013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19248
19249 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19250
3362b6b6
SA		 192512013-04-24 Sterling Augustine <saugustine@google.com>
19252
19253 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19254 (comp_dir_string, debug_str_dwo_section): New.
19255 (DEBUG_STR_DWO_SECTION): Rename to ...
19256 (DEBUG_DWO_STR_SECTION): ... this.
19257 (DEBUG_NORM_STR_SECTION): Delete.
19258 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19259 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19260 (find_AT_string): Move most logic to ...
19261 (find_AT_string_in_table): ... here. New.
19262 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19263 add_skeleton_AT_string. Delete logic.
19264 (output_skeleton_debug_sections): Remove call to
19265 add_top_level_skeleton_die_attrs.
19266 (add_comp_dir_attribute): Move logic to comp_dir_string.
19267 (dwarf2out_init): Initialize debug_str_dwo_section.
19268 (output_indirect_string): Call find_string_form.
19269 (output_indirect_strings): Rewrite.
19270 (prune_unused_types): Empty skeleton_debug_str_hash.
19271 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19272 (dwarf2out_finish): Call output_indirect_strings.
19273
e93e18e9
PC		 192742013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19275
19276 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19277
f6ce35ac
VM		 192782013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19279
cabf91cd 19280 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 19281 (LRA_SUBREG_P): New macro.
19282 * recog.c (register_operand): Check LRA_SUBREG_P.
19283 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19284 stack frame.
19285 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19286 (lra_final_code_change): Skip subreg change for operators.
19287 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19288 if there are no operand changes.
19289 * lra-constraints.c (curr_insn_set): New.
19290 (match_reload): Set LRA_SUBREG_P.
19291 (emit_spill_move): Ditto.
19292 (check_and_process_move): Use curr_insn_set. Process only single
19293 set insns. Don't initialize sec_mem_p and change_p.
19294 (simplify_operand_subreg): Use LRA_SUBREG_P.
19295 (reg_in_class_p): New function.
19296 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19297 of #ifdef. Add code to remove cycling.
19298 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19299 non-null disp. Reload inner instead of disp when base and index
19300 are null. Try to put lo_sum into register.
19301 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 19302 (check_and_process_move): Move code for move cost check to
f6ce35ac 19303 simple_move_p. Remove equiv_substitution.
cabf91cd 19304 (simple_move_p): New function.
f6ce35ac
VM		 19305 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19306 curr_insn_set. Call check_and_process_move only for single set
19307 insns. Use the new function. Move call of check_and_process_move
19308 after operand equiv substitution and address process.
19309
38047d90
JG		 193102013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19311
19312 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19313 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19314 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19315
13f39b2e
PC		 193162013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19317
19318 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19319
97c116dc
MP		 193202013-04-24 Marek Polacek <polacek@redhat.com>
19321
19322 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19323 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19324 (select_loops_exit_conditions): Likewise.
19325 (number_of_iterations_for_all_loops): Likewise.
19326 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19327 (scev_analysis): Likewise.
19328
83082391 193292013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 19330 Chao-ying Fu <fu@mips.com>
83082391 19331
cabf91cd
AM		 19332 * config/mips/micromips.md (jraddiusp): New pattern.
19333 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19334 instruction if possible.
83082391 19335
19e34aa2
AM		 193362013-04-24 Alan Modra <amodra@gmail.com>
19337
19338 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19339
fdb6603c
JB		 193402013-04-24 Julian Brown <julian@codesourcery.com>
19341 Chung-Lin Tang <cltang@codesourcery.com>
19342
19343 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 19344 dependency behavior in enumeration type DIE generation. Add TODO note
19345 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 19346
bf190e8d
LC		 193472013-04-23 Lawrence Crowl <crowl@google.com>
19348
19349 * Makefile.in: Update as needed below.
19350
19351 * hash-table.h (class hash_table):
19352 Correct many methods with parameter types compare_type to the correct
19353 value_type. (Correct code was unlikely to notice the change.)
19354 (hash_table::elements_with_deleted) New.
19355 (class hashtable::iterator): New.
19356 (hashtable::begin()): New.
19357 (hashtable::end()): New.
19358 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19359
19360 * statistics.c (statistics_hashes):
19361 Change type to hash_table. Update dependent calls and types.
19362
19363 * tree-into-ssa.c (var_infos):
19364 Change type to hash_table. Update dependent calls and types.
19365
19366 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19367 Change type to hash_table. Update dependent calls and types.
19368
19369 * tree-ssa-loop-im.c (struct mem_ref.refs):
19370 Change type to hash_table. Update dependent calls and types.
19371
19372 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19373 Change type to hash_table. Update dependent calls and types.
19374
19375 * tree-ssa-sccvn.c (vn_tables_s::nary):
19376 Change type to hash_table. Update dependent calls and types.
19377 (vn_tables_s::phis): Likewise.
19378 (vn_tables_s::references): Likewise.
19379
19380 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19381 (vn_reference_eq): Update parameter and return types.
19382
19383 * tree-ssa-structalias.c (pointer_equiv_class_table):
19384 Change type to hash_table. Update dependent calls and types.
19385 (location_equiv_class_table): Likewise.
19386
19387 * tree-vect-data-refs.c: Consequential changes for making
19388 peeling a hash_table.
19389
19390 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19391 (destroy_loop_vec_info): Dependent hash_table update.
19392
19393 * tree-vectorizer.h (peeling_htab):
19394 Change type to hash_table. Update dependent calls and types.
19395
d70a81dd
SC		 193962013-04-23 Shiva Chen <shiva0217@gmail.com>
19397
cabf91cd
AM		 19398 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19399 to check the register content is equal or not.
19400 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19401 to assign register content record.
19402 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 19403 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 19404 * lra-int.h (struct lra_reg): Add offset member.
19405 (lra_reg_val_equal_p): New static inline function.
19406 (lra_update_reg_val_offset): New static inline function.
19407 (lra_assign_reg_val): New static inline function.
19408 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19409 to assign register content record.
19410 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 19411
b894a1f3
CM		 194122013-04-23 Catherine Moore <clm@codesourcery.com>
19413
19414 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19415 operands. Record compression.
19416
ecd14de9
XDL		 194172013-04-23 Xinliang David Li <davidxl@google.com>
19418
19419 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19420
92e776e9
RB		 194212013-04-23 Richard Biener <rguenther@suse.de>
19422
19423 PR middle-end/57036
19424 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19425 parameter, only add abnormal goto edges from the copied body
19426 if the call could perform abnormal gotos.
19427 (copy_cfg_body): Adjust.
19428
a15ee567
SN		 194292013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19430
19431 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19432
08c52234
AS		 194332013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19434
19435 * coretypes.h (gimple_stmt_iterator): Add struct to make
19436 compatible with C.
19437
999c1171
RB		 194382013-04-23 Richard Biener <rguenther@suse.de>
19439
19440 PR tree-optimization/57026
19441 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19442 from SSA names occuring in abnormal PHI nodes.
19443
53669259
AK		 194442013-04-22 Andi Kleen <ak@linux.intel.com>
19445
19446 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19447
1a0ad150
AK		 194482013-04-22 Andi Kleen <ak@linux.intel.com>
19449
19450 * lto/lto.c (print_lto_report_1): Declare early.
19451 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19452
057f8f20
AK		 194532013-04-22 Andi Kleen <ak@linux.intel.com>
19454
19455 * common.opt (-flto-report-wpa): Add.
19456 * doc/invoke.texi (-flto-report-wpa): Add.
19457 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19458 (lto_main): dito.
19459
473b1e05
XDL		 194602013-04-22 Xinliang David Li <davidxl@google.com>
19461
19462 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19463 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19464 * Makefile.in: New dependency
19465
cabf91cd 19466 David Daney <ddaney.cavm@gmail.com>
b1485a33 19467
cabf91cd
AM		 19468 * configure.ac (gcc_cv_as_micromips_support): Use the
19469 --fatal-warnings option.
19470 * configure: Regenerate.
b1485a33 19471
829d0168
MP		 194722013-04-22 Marek Polacek <polacek@redhat.com>
19473
19474 PR sanitizer/56990
19475 * tsan.c (instrument_expr): Don't instrument expression
19476 in case its size is zero.
19477
6d9b7208
UB		 194782013-04-22 Uros Bizjak <ubizjak@gmail.com>
19479
19480 PR target/57032
19481 Revert:
19482 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19483
19484 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19485
ea679d55
JG		 194862013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19487
19488 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19489 (gimple_stmt_iterator): New typedef.
19490 * gimple.h (gimple_stmt_iterator): Rename to...
19491 (gimple_stmt_iterator_d): ... This.
19492 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19493 trees be valid for GIMPLE and GENERIC.
19494 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19495 * gimple-fold.c (gimple_fold_call): Call target hook
19496 gimple_fold_builtin.
19497 * hooks.c (hook_bool_gsiptr_false): New.
19498 * hooks.h (hook_bool_gsiptr_false): New.
19499 * target.def (fold_stmt): New.
19500 * doc/tm.texi: Regenerate.
19501
88a581da
VM		 195022013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19503
19504 PR target/57018
19505 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19506 a set sp if no stack realignment.
19507
92be22dc
NC		 195082013-04-22 Nick Clifton <nickc@redhat.com>
19509
19510 * config.gcc (tilegx-linux): Extend extra_objs rather than
19511 overwriting it.
19512 (tilepro-linux): Likewise.
19513
0ddec79f
JG		 195142013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19515
19516 * config/aarch64/aarch64-builtins.c
19517 (CF): Remove.
19518 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19519 (VAR<1-12>): Add MAP parameter.
19520 (BUILTIN_*): Likewise.
19521 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19522 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19523 (aarch64_ushl_n<mode>): Likewise.
19524 (aarch64_sshr_n<mode>): Likewise.
19525 (aarch64_ushr_n<mode>): Likewise.
19526 (aarch64_<maxmin><mode>): Likewise.
19527 (aarch64_sqrt<mode>): Likewise.
19528 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19529 (vshr<q>_n_*): Likewise.
19530
0050faf8
JG		 195312013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19532
19533 * config/aarch64/aarch64-builtins.c
19534 (aarch64_simd_builtin_type_mode): Handle SF types.
19535 (sf_UP): Define.
19536 (BUILTIN_GPF): Define.
19537 (aarch64_init_simd_builtins): Handle SF types.
19538 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19539 (frecps): Likewise.
19540 (frecpx): Likewise.
19541 * config/aarch64/aarch64-simd.md
19542 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19543 (aarch64_frecpe<mode>): New.
19544 (aarch64_frecps<mode>): Likewise.
19545 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19546 (v8type): Add frecp<esx>.
19547 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19548 (aarch64_frecps<mode>): Likewise.
19549 * config/aarch64/iterators.md (FRECP): New.
19550 (frecp_suffix): Likewise.
19551 * config/aarch64/arm_neon.h
19552 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19553
0fad3dbc 195542013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 19555
19556 PR target/56995
19557 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19558 (REG_CLASS_NAMES): Idem.
19559 (REG_CLASS_CONTENTS): Idem.
19560 (REGCLASS_HAS_FP_REG): Idem.
19561 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19562 (sh_conditional_register_usage): Idem.
19563
3e8a33f9
JL		 195642013-04-21 Jeff Law <law@redhat.com>
19565
19566 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19567 (ssa_forward_propagate_and_combine): Use it.
19568
f38e1b0a
VM		 195692013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19570
19571 * lra.c: Update the flow chart diagram.
19572
682303da
VM		 195732013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19574
19575 PR rtl-optimization/56847
19576 * lra-constraints.c (process_alt_operands): Discourage alternative
19577 with non-matche doffsettable memory constraint fro memory with
19578 known offset.
19579
f6b64c35
RB		 195802013-04-19 Richard Biener <rguenther@suse.de>
19581
19582 PR tree-optimization/56982
19583 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19584 function.
19585 * gimplify.c (gimplify_call_expr): Notice special calls.
19586 (gimplify_modify_expr): Likewise.
19587 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19588 abnormal control flow receivers.
19589 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19590 in the same way as cfun->has_nonlocal_labels.
19591 (gimple_purge_dead_abnormal_call_edges): Likewise.
19592 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19593 receivers start a basic-block.
19594
01d8bf07
RB		 195952013-04-19 Richard Biener <rguenther@suse.de>
19596
19597 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19598 member ...
19599 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19600 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19601 (SLP_TREE_LOAD_PERMUTATION): Add.
19602 (vect_transform_slp_perm_load): Adjust prototype.
19603 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19604 (vect_free_slp_instance): Likewise.
19605 (vect_create_new_slp_node): Likewise.
19606 (vect_supported_slp_permutation_p): Remove.
19607 (vect_slp_rearrange_stmts): Adjust.
19608 (vect_supported_load_permutation_p): Likewise. Inline
19609 vect_supported_slp_permutation_p here.
19610 (vect_analyze_slp_instance): Compute load permutations per
19611 slp node instead of per instance.
19612 (vect_get_slp_defs): Adjust.
19613 (vect_transform_slp_perm_load): Likewise.
19614 (vect_schedule_slp_instance): Remove redundant code.
19615 (vect_schedule_slp): Remove hack for PR56270, add it ...
19616 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19617 CSE loads for SLP. Adjust.
19618
ede22fc3
GY		 196192013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19620
19621 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19622 spelling in two comments.
19623
67bc84fb
GY		 196242013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19625
19626 PR target/56797
19627 * config/arm/arm.c (load_multiple_sequence): Require SP
19628 as base register for loads if SP is in the register list.
19629
e248d83f
MJ		 196302013-04-19 Martin Jambor <mjambor@suse.cz>
19631
19632 PR tree-optimization/56718
19633 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19634 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19635 and made public. Adjusted all callers.
19636 (ipa_intraprocedural_devirtualization): New function.
19637 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19638 (ipa_intraprocedural_devirtualization): Likewise.
19639 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19640
4891e8f8
RB		 196412013-04-19 Richard Biener <rguenther@suse.de>
19642
19643 PR tree-optimization/57000
19644 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19645
dad89f7c
TG		 196462013-04-19 Terry Guo <terry.guo@arm.com>
19647
19648 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19649 Replace with ...
19650 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19651 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19652 (cortex_m4_fmacs): Use new reservations.
19653 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19654
72ea0d47
VM		 196552013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19656
f1e6512c 19657 PR rtl-optimization/56999
72ea0d47
VM		 19658 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19659 related code.
19660 (lra_coalesce): Remove split_origin_bitmap and related code.
19661 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19662 ranges if necessary.
19663
780a5b71
UB		 196642013-04-18 Uros Bizjak <ubizjak@gmail.com>
19665
19666 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19667 New array.
19668 (ix86_expand_call): Remove clobbered_registers array and use
19669 x86_64_ms_sysv_extra_clobbered_registers instead.
19670 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19671 Declare here.
19672 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19673 predicate.
19674 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19675 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19676 (*call_value_rex64_ms_sysv): Ditto.
19677
6f5a366a
CC		 196782013-04-18 Cary Coutant <ccoutant@google.com>
19679
19680 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19681 parent instead of die_mark.
19682
475b8f37
DN		 196832013-04-18 Diego Novillo <dnovillo@google.com>
19684
19685 * gimple.c (create_gimple_tmp): New.
19686 (get_expr_type): New.
19687 (build_assign): New.
19688 (build_type_cast): New.
19689 * gimple.h (enum ssa_mode): Define.
19690 (gimple_seq_set_location): New.
19691 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 19692 to use build_assign and build_type_cast.
475b8f37 19693
08940f33
RB		 196942013-04-18 Richard Biener <rguenther@suse.de>
19695
19696 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19697 handle negative step. Remove redundant checks.
19698 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19699 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19700 for negative step and grouped loads fail to vectorize.
19701
0e0f87d4
SB		 197022013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19703
19704 * emit-rtl.c (reset_insn_used_flags): New function.
19705 (reset_all_used_flags): Use it.
19706 (verify_insn_sharing): New function.
19707 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19708
4c445590
JJ		 197092013-04-18 Jakub Jelinek <jakub@redhat.com>
19710
19711 PR tree-optimization/56984
19712 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19713 and (x >> M) >= N don't register any assertion if N << M is the
19714 minimum value.
19715
6873ecab
SB		 197162013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19717
19718 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19719 do not delete_insn insns that have not yet been emitted, only
19720 unlink them with remove_insn.
19721 * df-scan.c (df_insn_delete): Revert r197492.
19722
3ccb989e
SB		 197232013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19724
19725 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19726 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19727
197282013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 19729
19730 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19731 define_insn_and_split.
19732 (and_scc,ior_scc,negscc): Likewise.
19733 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19734
3ccb989e 197352013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 19736
19737 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19738 can benefit from using a sequence of LDRD instructions in epilogue
19739 instead of a single LDM instruction.
19740
6d10a203
MLI		 197412013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19742
19743 PR 45688
19744 * doc/extend.texi: Fix typo.
19745
6983e6b5
RB		 197462013-04-17 Richard Biener <rguenther@suse.de>
19747
19748 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19749 (vect_build_slp_tree): ... here.
19750 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19751 match. Remove special-casing of mismatched complex loads.
19752 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19753 re-try the match with swapped commutative operands.
19754 (vect_supported_load_permutation_p): Remove special-casing of
19755 mismatched complex loads.
19756 (vect_analyze_slp_instance): Adjust.
19757
ef23e6a2
RB		 197582013-04-17 Richard Biener <rguenther@suse.de>
19759
19760 PR rtl-optimization/56921
19761 * cfgloop.h (struct loop): Add simple_loop_desc member.
19762 (struct niter_desc): Mark with GTY(()).
19763 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19764 * loop-iv.c (get_simple_loop_desc): Likewise.
19765 (free_simple_loop_desc): Likewise.
19766
19767 Revert
19768 2013-04-16 Richard Biener <rguenther@suse.de>
19769
19770 PR rtl-optimization/56921
19771 * loop-init.c (pass_rtl_move_loop_invariants): Add
19772 TODO_do_not_ggc_collect to todo_flags_finish.
19773 (pass_rtl_unswitch): Same.
19774 (pass_rtl_unroll_and_peel_loops): Same.
19775 (pass_rtl_doloop): Same.
19776
fc6f94f5
EB		 197772013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19778
19779 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19780 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19781 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19782 references.
19783 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19784 * tree-streamer.c (record_common_node): Adjust reference in comment.
19785
10a88311
TG		 197862013-04-17 Terry Guo <terry.guo@arm.com>
19787
19788 * config/arm/cortex-m4.md: Add a new bypass.
19789
6d9b7208 197902013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 19791
19792 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19793 New pattern.
19794 (*subs_<optab><mode>_multp2): New pattern.
19795 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19796 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19797
6d9b7208 197982013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 19799
19800 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19801 (*subs_mul_imm_<mode>): New pattern.
19802
18a6701e
DE		 198032013-04-16 David Edelsohn <dje.gcc@gmail.com>
19804
19805 PR target/56948
19806 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19807 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19808 (vsx_movti_32bit): Change j->wa to O->wa.
19809
07c37b2f
RB		 198102013-04-16 Richard Biener <rguenther@suse.de>
19811
19812 PR rtl-optimization/56921
19813 * loop-init.c (pass_rtl_move_loop_invariants): Add
19814 TODO_do_not_ggc_collect to todo_flags_finish.
19815 (pass_rtl_unswitch): Same.
19816 (pass_rtl_unroll_and_peel_loops): Same.
19817 (pass_rtl_doloop): Same.
19818
0e0f87d4 198192013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 19820
19821 * config/arm/arm.c (emit_multi_reg_push): New declaration
19822 for an existing function.
19823 (arm_emit_strd_push): New function.
19824 (arm_expand_prologue): Used here.
19825 (arm_emit_ldrd_pop): New function.
19826 (arm_expand_epilogue): Used here.
19827 (arm_get_frame_offsets): Update condition.
19828 (arm_emit_multi_reg_pop): Add a special case for load of a single
19829 register with writeback.
19830
5e8e2af4
UB		 198312013-04-16 Uros Bizjak <ubizjak@gmail.com>
19832
19833 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19834 description.
19835
9fd9ccf7
RB		 198362013-04-16 Richard Biener <rguenther@suse.de>
19837
19838 PR tree-optimization/56756
19839 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19840 (first_mem_ref_loc): New.
19841 (execute_sm): Place the load temporarily before a previous
19842 access instead of in the latch edge to ensure its SSA dependencies
19843 are defined at points dominating the load.
19844
96fba521
SB		 198452013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19846
4c8af858
SB		 19847 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19848 correct fix by moving header and footer insn to the footer of
19849 the merged basic block. Clear BB_END of the merged-away block.
19850
96fba521
SB		 19851 PR middle-end/43631
19852 * emit-rtl.c (make_note_raw): New function.
19853 (link_insn_into_chain): New static inline function.
19854 (add_insn): Use it.
19855 (add_insn_before, add_insn_after): Factor insn chain linking code...
19856 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19857 using link_insn_into_chain.
19858 (note_outside_basic_block_p): New helper function for emit_note_after
19859 and emit_note_before.
19860 (emit_note_after): Use nobb variant of add_insn_after if the note
19861 should not be contained in a basic block.
19862 (emit_note_before): Use nobb variant of add_insn_before if the note
19863 should not be contained in a basic block.
19864 (emit_note_copy): Use make_note_raw.
19865 (emit_note): Likewise.
19866 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19867 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19868 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19869 the moved barrier the tail of the basic block it follows.
19870 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19871
7b8265ba
JJ		 198722013-04-15 Jakub Jelinek <jakub@redhat.com>
19873
19874 PR tree-optimization/56962
19875 * gimple-ssa-strength-reduction.c (record_increment): Only set
19876 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19877 either rhs1 or rhs2 is equal to c->base_expr.
19878
5185d248
RB		 198792013-04-15 Richard Biener <rguenther@suse.de>
19880
19881 PR tree-optimization/56933
19882 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19883 member.
19884 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19885 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19886 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19887 dependence check ...
19888 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19889 ... here.
19890 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19891 GROUP_READ_WRITE_DEPENDENCE.
19892
a24243a0
AK		 198932013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19894
19895 * emit-rtl.c (reset_all_used_flags): New function.
19896 (verify_rtl_sharing): Call reset_all_used_flags before and after
19897 performing the checks.
19898
1c50eada
KT		 198992013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19900
19901 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19902 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19903 * config/arm/constraints.md (De): New constraint.
19904 * config/arm/neon.md (anddi3_neon): Delete.
19905 (neon_vand<mode>): Expand to standard anddi3 pattern.
19906 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19907 Move earlier in the file.
19908 (neon_inv_logic_op2): Likewise.
19909 (arm_anddi_operand_neon): New predicate.
19910
e927b6ad
RO		 199112013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19912
19913 * configure.ac (gcc_cv_ld_as_needed): Set
19914 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19915 Use -z ignore, -z record on *-*-solaris2*.
19916 (HAVE_LD_AS_NEEDED): Update comment.
19917 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19918 * configure: Regenerate.
19919 * config.in: Regenerate.
19920 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19921 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19922 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19923 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19924 equivalents. Fix markup.
19925 * doc/tm.texi: Regenerate.
19926
e0ea8797
AH		 199272013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19928
19929 * config/i386/i386.opt: New option mstack-protector-guard=.
19930 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19931 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19932 TARGET_SSP_TLS_GUARD.
19933 * config/i386/i386.c (ix86_option_override_internal): Set
19934 ix86_stack_protector_guard.
19935 * config/i386/i386.md (stack_protect_set): Enable for
19936 TARGET_SSP_TLS_GUARD only.
19937 (stack_protect_set_<mode>): Ditto.
19938 (stack_protect_test): Ditto.
19939 (stack_protect_test_<mode>): Ditto.
19940 * doc/invoke.texi (i386 Option): Document.
19941
811b72f9
EB		 199422013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19943
19944 PR target/56890
19945 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19946 (S_MODES): Set H_MODE bit.
19947 (SF_MODES): Set only S_MODE and SF_MODE bits.
19948 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19949 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19950 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19951 <MODE_FLOAT>: Likewise.
19952
5529fdd6
JY		 199532013-04-15 Joey Ye <joey.ye@arm.com>
19954
19955 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19956
517b1da2
JY		 199572013-04-15 Joey Ye <joey.ye@arm.com>
19958
19959 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19960 for real far jump.
19961 (thumb_far_jump_used_p): Count instruction size and set
19962 far_jump_used.
19963
01007ae0
EB		 199642013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19965
19966 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19967 * resource.c (mark_target_live_regs): Reformat conditional block.
19968
c46f6580
SB		 199692013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19970
19971 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19972 notes, they are emitted only just before final.
19973 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19974
1f397f45
SB		 199752013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19976
19977 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19978 * cfgrtl.c (delete_insn): Call it here instead.
19979 * lra-spills.c (lra_final_code_change): Use delete_insn.
19980 * haifa-sched.c (sched_remove_insn): Likewise.
19981 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19982 returning to the nop pool.
19983 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19984 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19985
58a51369
SB		 199862013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19987
19988 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19989 * doc/tm.texi: Regenerated.
19990
33159866
UB		 199912013-04-12 Uros Bizjak <ubizjak@gmail.com>
19992
19993 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19994 QImode checks.
19995
226e378f
SB		 199962013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19997
19998 * df-core.c (df_find_def): Compare register numbers.
19999 (df_find_use): Likewise.
20000
fafb9b18
VM		 200012013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20002
20003 PR target/56903
20004 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20005 lra_in_progress for return.
20006
9a946fd6
GY		 200072013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20008
20009 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20010 define_insn into define_insn_and_split and emit movsicc patterns.
20011
41b83758
GY		 200122013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20013
20014 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20015
d3afd9aa
RB		 200162013-04-12 Richard Biener <rguenther@suse.de>
20017
20018 * tree-pass.h (TODO_do_not_ggc_collect): New.
20019 * passes.c (execute_one_ipa_transform_pass): Honor
20020 TODO_do_not_ggc_collect.
20021 (execute_one_pass): Likewise.
20022
20023 Revert
20024 2013-04-10 Richard Biener <rguenther@suse.de>
20025
20026 * passes.c (init_optimization_passes): Remove reload pass.
20027 * ira.c (do_reload): Merge into ...
20028 (ira): ... this.
20029 (rest_of_handle_reload): Remove.
20030 (pass_reload): Likewise.
20031 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20032 of reload for vzeroupper pass placement.
20033
06f9b387
JJ		 200342013-04-12 Jakub Jelinek <jakub@redhat.com>
20035
20036 PR tree-optimization/56918
20037 PR tree-optimization/56920
20038 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20039 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20040 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20041 use wide_mul_with_sign method.
20042
953094d2
RB		 200432013-04-12 Richard Biener <rguenther@suse.de>
20044
20045 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20046 not be considered a gimple constant.
20047
bb506982
MG		 200482013-04-12 Marc Glisse <marc.glisse@inria.fr>
20049
20050 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20051 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20052 operations.
20053
4b84d650
JJ		 200542013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20055 Jakub Jelinek <jakub@redhat.com>
20056
20057 * opts.c: Include diagnostic-color.h.
20058 (common_handle_option): Handle OPT_fdiagnostics_color_.
20059 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20060 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20061 (diagnostic-color.o): New.
20062 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20063 (diagnostic_color_rule): New enum.
20064 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20065 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20066 the location string.
20067 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20068 either NULL, or color kind.
20069 * diagnostic-color.c: New file.
20070 * diagnostic-color.h: New file.
20071 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20072 arguments.
20073 * doc/invoke.texi (-fdiagnostics-color): Document.
20074 * pretty-print.h (pp_show_color): Define.
20075 (struct pretty_print_info): Add show_color field.
20076 * diagnostic.c: Include diagnostic-color.h.
20077 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20078 macros. Colorize error:, warning: etc. strings and also the location
20079 string.
20080 (diagnostic_show_locus): Colorize the caret line.
20081 * pretty-print.c: Include diagnostic-color.h.
20082 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20083 inside of %< %> quotes or quoted through q format modifier.
20084
067a1e71
AK		 200852013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20086
33159866 20087 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 20088
33159866 200892013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 20090
20091 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20092 code in CC_NZ mode.
20093 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20094 pattern.
20095
7b55f98f
MP		 200962013-04-11 Marek Polacek <polacek@redhat.com>
20097
20098 PR tree-optimization/48184
33159866 20099 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 20100
966b587e
EB		 201012013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20102
20103 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20104 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20105 (skip_simple_arithmetic): Tidy up.
20106 * tree.h (skip_simple_constant_arithmetic): Declare.
20107
33159866 201082013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 20109
20110 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20111
1d42c1ec
RB		 201122013-04-11 Richard Biener <rguenther@suse.de>
20113
20114 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20115 generate vector constants.
20116
4ba5ea11
RB		 201172013-04-11 Richard Biener <rguenther@suse.de>
20118
20119 PR tree-optimization/56878
20120 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20121 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20122 New function.
20123 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20124 Prefer to align the DR with the most invariant base address.
20125
f0defe58
SKS		 201262013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20127
20128 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20129 comment.
20130
d07458be
JG		 201312013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20132
20133 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20134 floating-point vector comparisons against 0.
20135
146b8692
JJ		 201362013-04-11 Jakub Jelinek <jakub@redhat.com>
20137
20138 PR tree-optimization/56899
20139 * fold-const.c (extract_muldiv_1): Apply distributive law
20140 only if TYPE_OVERFLOW_WRAPS (ctype).
20141
b8578ff7
BC		 201422013-04-11 Bin Cheng <bin.cheng@arm.com>
20143
20144 PR target/56124
20145 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20146 loading has side effect.
20147
0ea8a6f9
SB		 201482013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20149
20150 * config/sparc/sparc.c: Include tree-pass.h.
20151 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20152 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20153 head of file. Change return type. Split off gate function.
20154 (sparc_gate_work_around_errata): New function.
20155 (pass_work_around_errata): New pass definition.
20156 (insert_pass_work_around_errata) New pass insert definition to
20157 insert pass_work_around_errata just after delayed-branch scheduling.
20158 (sparc_option_override): Insert the pass.
20159 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20160
42e37616
DM		 201612013-04-10 David S. Miller <davem@davemloft.net>
20162
89deeb3b
DM		 20163 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20164 or -mcpu=hypersparc.
20165
42e37616
DM		 20166 * target.def (cstore_mode): New hook.
20167 * target.h: Include insn-codes.h
20168 * targhooks.c: Likewise.
20169 (default_cstore_mode): New function.
20170 * targhooks.h: Declare it.
20171 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20172 * doc/tm.texi: Rebuild.
20173 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20174 target hook, rather than inspecting the insn_data.
20175 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20176 (TARGET_CSTORE_MODE): Redefine.
20177 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20178 result patterns.
20179 * config/sparc/predicates.md (cstore_result_operand): New special
20180 predicate.
20181 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20182 Use it for operand 0.
20183 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20184 (*snesi_special): Likewise.
20185 (*snesi_zero): Likewise.
20186 (*seqsi_zero): Likewise.
20187 (*sltu_insn): Likewise.
20188 (*sgeu_insn): Likewise.
20189 (*seqdi_special): Make operand 0 and comparison operation be of
20190 DImode.
20191 (*snedi_special): Likewise.
20192 (*snedi_special_vis3): Likewise.
20193 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20194 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20195 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20196 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20197 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20198 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20199 (*sltu_extend_sp64): Likewise.
20200 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20201 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20202 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20203 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20204 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20205
95ca411e
YZ		 202062013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20207
20208 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20209 (aarch64_start_file): Use the new function.
20210
6782438d 202112013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 20212 Jason Merrill <jason@redhat.com>
6782438d
SKS		 20213
20214 * common.opt: Add -gdwarf.
20215 * opts.c (common_handle_option): Handle it.
20216 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20217
bb313b93
RB		 202182013-04-10 Richard Biener <rguenther@suse.de>
20219
20220 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20221 (execute_one_ipa_transform_pass): But unconditionally here.
20222 (execute_one_pass): And here.
20223 (init_optimization_passes): Remove reload pass.
20224 * tree-pass.h (TODO_ggc_collect): Remove.
20225 (pass_reload): Likewise.
20226 * ira.c (do_reload): Merge into ...
20227 (ira): ... this.
20228 (rest_of_handle_reload): Remove.
20229 (pass_reload): Likewise.
20230 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20231 of reload for vzeroupper pass placement.
20232 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20233 and todo_flags_finish of all passes.
20234
793d9a16
RB		 202352013-04-10 Richard Biener <rguenther@suse.de>
20236
20237 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20238 first_const_oprnd field, rename first_def_type to first_op_type.
20239 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20240 (vect_get_and_check_slp_defs): Always use the type of the
20241 operand. Allow mixed vect_external_def, vect_constant_def types.
20242 (vect_get_constant_vectors): Handle mixed vect_external_def,
20243 vect_constant_def types.
20244
12211b99 202452013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 20246
20247 PR tree-optimization/55524
20248 * tree-ssa-math-opts.c
20249 (convert_mult_to_fma): Don't use an fms construct
20250 when we don't have an fms operation, but fnma, and it looks
20251 likely that we'll be able to use the latter.
20252
12211b99 202532013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 20254
20255 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20256 function.
20257 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20258 inline fail caused by overwritable functions.
20259
34ab4a5b
CJW		 202602013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20261
20262 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20263 unnecessary bits in the constant power of two case.
20264
abf9bfbc
RB		 202652013-04-10 Richard Biener <rguenther@suse.de>
20266
20267 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20268 broken code swapping operands.
20269 (vect_build_slp_tree): Do not compute load permutations here.
20270 (vect_analyze_slp_instance): Compute load permutations here,
20271 after building the SLP tree.
20272
f408477e
CB		 202732013-04-09 Christian Bruel <christian.bruel@st.com>
20274
20275 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20276 of next/prev_real_insn.
20277
5ac42672
JH		 202782013-04-09 Jan Hubicka <jh@suse.cz>
20279
abf9bfbc
RB		 20280 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20281 Drop aliased parameter.
5ac42672
JH		 20282 (function_and_variable_visibility): Do not handle alias pairs.
20283 * cgraph.c (varpool_externally_visible_p): Update prototype.
20284 * varpool.c (varpool_add_new_variable): Update.
20285
5017f1d2
KT		 202862013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20287
20288 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20289
48eecbee
SB		 202902013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20291
a949cf1c
SB		 20292 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20293
48eecbee
SB		 20294 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20295
75ef8e3d
MP		 202962013-04-09 Marek Polacek <polacek@redhat.com>
20297
20298 PR tree-optimization/48762
33159866 20299 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 20300
23847df4
RB		 203012013-04-09 Richard Biener <rguenther@suse.de>
20302
20303 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20304 dealing with cost.
20305 (vect_build_slp_tree): Likewise.
20306 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20307 calculating the cost of a SLP instance.
20308 (vect_analyze_slp_instance): Use it from here, after building
20309 the SLP tree.
20310
ea3a0fde
JJ		 203112013-04-09 Jakub Jelinek <jakub@redhat.com>
20312
20313 PR middle-end/56883
20314 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20315 expand_omp_for_static_chunk): Use simple_p = true in
20316 force_gimple_operand_gsi calls when assigning to addressable decls.
20317
a32dfe9d
JL		 203182013-04-09 Jeff Law <law@redhat.com>
20319
20320 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20321 when the boolean was created by converting a wider object which
20322 had a boolean range.
20323
d755c7ef
RB		 203242013-04-09 Richard Biener <rguenther@suse.de>
20325
20326 * tree-vectorizer.h (slp_void_p): Remove.
20327 (slp_tree): Typedef before _slp_tree declaration.
20328 (struct _slp_tree): Use a vector of slp_tree as children.
20329 (vect_get_place_in_interleaving_chain): Remove.
20330 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20331 Move ...
20332 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20333 and make static.
20334 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20335 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20336 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20337 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20338 Use slp_node instead of slp_void_p and adjust.
20339
3d741091
RB		 203402013-04-09 Richard Biener <rguenther@suse.de>
20341
20342 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20343 work that is not necessary.
20344
39307ba7
JJ		 203452013-04-09 Jakub Jelinek <jakub@redhat.com>
20346
20347 PR tree-optimization/56854
20348 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20349 forward into clobber stmts if it would change MEM_REF lhs into
20350 non-MEM_REF.
20351
343881fd
MK		 203522013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20353
20354 * tree.c (type_hash_lookup, type_hash_add): Make static.
20355 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20356
3922658a
RB		 203572013-04-09 Richard Biener <rguenther@suse.de>
20358
20359 * tree.h (unsave_expr_now): Remove.
20360 * tree-inline.c (mark_local_for_remap_r): Remove.
20361 (unsave_expr_1): Likewise.
20362 (unsave_r): Likewise.
20363 (unsave_expr_now): Likewise.
20364 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20365 (propagate_tree_value): Likewise.
20366
9fb6b620
SB		 203672013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20368
20369 * doc/rtl.texi (sequence): Rewrite documentation to match the
20370 current use of SEQUENCE rtl objects.
20371 * rtl.def (SEQUENCE): Likewise.
20372
20373 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20374 Update documentation.
20375 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20376 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20377
20378 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20379
8ddb5a29
TJ		 203802013-04-08 Teresa Johnson <tejohnson@google.com>
20381
20382 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20383 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 20384 methods.
8ddb5a29
TJ		 20385 (estimate_edge_size_and_time): Add comment to suggest using rounding
20386 methods.
20387 (estimate_node_size_and_time): Ditto.
20388 (remap_edge_change_prob): Use helper rounding divide methods.
20389 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20390 (gimple_mod_pow2_value_transform): Ditto.
20391 (gimple_mod_subtract_transform): Ditto.
20392 (gimple_ic_transform): Ditto.
20393 (gimple_stringops_transform): Ditto.
20394 * stmt.c (conditional_probability): Ditto.
20395 (emit_case_dispatch_table): Ditto.
20396 * lto-cgraph.c (merge_profile_summaries): Ditto.
20397 * tree-optimize.c (execute_fixup_cfg): Ditto.
20398 * cfgcleanup.c (try_forward_edges): Ditto.
20399 * cfgloopmanip.c (scale_loop_profile): Ditto.
20400 (loopify): Ditto.
20401 (duplicate_loop_to_header_edge): Ditto.
20402 (lv_adjust_loop_entry_edge): Ditto.
20403 * tree-vect-loop.c (vect_transform_loop): Ditto.
20404 * profile.c (compute_branch_probabilities): Ditto.
20405 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20406 * lto-streamer-in.c (input_cfg): Ditto.
20407 * gimple-streamer-in.c (input_bb): Ditto.
20408 * ipa-cp.c (update_profiling_info): Ditto.
20409 (update_specialized_profile): Ditto.
20410 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20411 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 20412 rounding methods.
8ddb5a29
TJ		 20413 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20414 (compute_trg_info): Ditto.
20415 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20416 (purge_dead_edges): Ditto.
20417 * loop-unswitch.c (unswitch_loop): Ditto.
20418 * cgraphclones.c (cgraph_clone_edge): Ditto.
20419 (cgraph_clone_node): Ditto.
20420 * tree-inline.c (copy_bb): Ditto.
20421 (copy_edges_for_bb): Ditto.
20422 (initialize_cfun): Ditto.
20423 (copy_cfg_body): Ditto.
20424 (expand_call_inline): Ditto.
20425
661e6bd7
KT		 204262013-04-08 Kai Tietz <ktietz@redhat.com>
20427
20428 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20429 TARGET_CYGWIN64 by TARGET_64BIT.
20430
105766f3
JR		 204312013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20432
20433 * config/epiphany/epiphany.md (GPR_1): New constant.
20434 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20435 * config/epiphany/epiphany.c (gen_compare_reg):
20436 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20437 is already in place.
20438 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20439 Don't require being called during rtl expansion; If y operlaps r0,
20440 return 0.
20441 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20442 (epiphany_expand_epilogue): Likewise.
20443
8afab237
JR		 20444 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20445 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20446 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20447
fa7707d6
JR		 20448 * config/epiphany/constraints.md (CnL): New constraint.
20449 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20450 * config/epiphany/predicates.md (add_operand): Allow 1024.
20451
5138e18d
JR		 20452 * config/epiphany/epiphany.md (logical_op): New code iterator.
20453 (op_mnc): New code attribute.
20454 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20455 (mov_f+1, mov_f+2): New peephole2 patterns.
20456
2ccc703d
JR		 20457 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20458 (cstoresi4): Also allow re-use of zero result when doing a NE
20459 comparison to a non-zero operand.
aefb0819 20460 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20461
093ac0a5
JR		 20462 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20463 Use gen_addsi3_i / gen_subsi3_i.
20464
f223bb13
JJ		 204652013-04-08 Jakub Jelinek <jakub@redhat.com>
20466
20467 PR c++/34949
20468 PR c++/50243
20469 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20470 contain anything but clobbers, at most one __builtin_stack_restore,
20471 optionally debug stmts and final resx, and if it has at least one
20472 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20473 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20474 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20475 which isn't defaut definition, remove them.
20476 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20477 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20478 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20479 with MEM_REF LHS with SSA_NAME address.
20480
4481581f
JL		 204812013-04-08 Jeff Law <law@redhat.com>
20482
20483 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20484
451dabda
RB		 204852013-04-08 Richard Biener <rguenther@suse.de>
20486
20487 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20488 extra newline.
20489 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20490 determined vector type.
20491 (vect_analyze_data_refs): Likewise.
20492 (vect_get_new_vect_var): Adjust.
20493 (vect_create_destination_var): Preserve SSA name versions.
20494 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20495 not dump anything here.
20496
3b088b47
JR		 204972013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20498
20499 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20500 Add member lr_slot_known.
20501 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20502 if necessary.
20503 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20504 Remove code that sets lr_slot_offset according to what a previous
20505 version of epiphany_emit_save_restore used to do.
20506 (epiphany_emit_save_restore): When doing an lr save or restore,
20507 set/verify lr_slot_known and lr_slot_offset.
20508
d8484d41
XQ		 205092013-04-08 Xinyu Qi <xyqi@marvell.com>
20510
33159866
UB		 20511 PR target/54338
20512 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 20513 in ALL_REGS.
20514
481be1c4
RB		 205152013-04-08 Richard Biener <rguenther@suse.de>
20516
20517 * alias.c (find_base_term): Fix thinko in previous change.
20518
401f3a81
JJ		 205192013-04-08 Jakub Jelinek <jakub@redhat.com>
20520
20521 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20522 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20523 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20524 if possible to compute val.
20525 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20526 For QImode integers don't require anything about precision. Use
20527 const_with_all_bytes_same to find out if the constant doesn't have
20528 repeated bytes in it.
20529
03ed99a8
AK		 205302013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20531
20532 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20533 within mode size.
20534
781b2e62
MP		 205352013-04-08 Marek Polacek <polacek@redhat.com>
20536
20537 PR rtl-optimization/48182
20538 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20539 value to 1.
20540
27e430a2
JDA		 205412013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20542
20543 PR target/55487
20544 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20545 nuses, make sure we have a label.
20546
4902aa64
BS		 205472013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20548
20549 PR target/56843
20550 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20551 (rs6000_emit_swdiv_low_precision): Remove.
20552 (rs6000_emit_swdiv): Rewrite to handle between one and four
20553 iterations of Newton-Raphson generally; modify required number of
20554 iterations for some cases.
20555 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20556
7bca81dc
SB		 205572013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20558
20559 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20560 set-but-unused variable.
20561
20562 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20563 basic blocks of released function bodies garbage-collectable.
20564
20565 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20566 (struct rtl_opt_pass): Add TODO_df_finish.
20567
20568 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20569
4542a38a
GY		 205702013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20571
20572 * config/arm/constraints.md (q): New constraint.
20573 * config/arm/ldrdstrd.md: New file.
20574 * config/arm/arm.md (ldrdstrd.md) New include.
20575 (arm_movdi): Use "q" instead of "r" constraint
20576 for double-word memory access.
20577 (movdf_soft_insn): Likewise.
20578 * config/arm/vfp.md (movdi_vfp): Likewise.
20579 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20580 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 20581 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20582 (mem_ok_for_ldrd_strd): Likewise.
20583 (output_move_double): Update assertion.
20584
2385b218
GY		 205852013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20586
20587 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20588
75fe1cb5
GY		 205892013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20590
20591 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20592 define_insn_and_split.
20593 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20594
dd660e8e
GY		 205952013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20596
20597 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20598 define_insn_and_split.
33159866 20599 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 20600 (shiftsi3_compare): New pattern.
20601 (rrx): New pattern.
20602 * config/arm/unspecs.md (UNSPEC_RRX): New.
20603
045e472c
GY		 206042013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20605
20606 * config/arm/arm.md (negdi_extendsidi): New pattern.
20607 (negdi_zero_extendsidi): Likewise.
20608
3f3bf1a8
GY		 206092013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20610
20611 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20612 define_insn_and_split.
20613 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20614 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20615
b0b49556
GY		 206162013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20617
20618 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20619 define_insn_and_split.
20620 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20621 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20622
d633dd84
GY		 206232013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20624
20625 * config/arm/arm.md (subsi3_carryin): New pattern.
20626 (subsi3_carryin_const): Likewise.
20627 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20628 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20629
f4499066
GY		 206302013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20631
20632 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20633
ceef6fd9
GY		 206342013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20635
20636 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20637 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20638
ddbdd8a7
KT		 206392013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20640
20641 * config/arm/arm.c (arm_expand_builtin): Change fcode
20642 type to unsigned int.
20643
8456d78a
RR		 206442013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20645
20646 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20647
526c230f
ILT		 206482013-04-04 Ian Lance Taylor <iant@google.com>
20649
33159866
UB		 20650 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20651 language standard.
526c230f 20652
3cfbe04d
SB		 206532013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20654
20655 PR middle-end/56729
20656 * df-scan.c (df_insn_delete): Disable failing assert.
20657
dfa3f8d0
KT		 206582013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20659
20660 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20661 New function prototype.
20662 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20663 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20664 (arm_builtin_vectorized_function): New function.
20665
f7837758
KT		 206662013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20667
20668 * config/arm/arm_neon_builtins.def: New file.
20669 * config/arm/arm.c (neon_builtin_data): Move contents to
20670 arm_neon_builtins.def.
20671 (enum arm_builtins): Include neon builtin definitions.
20672 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 20673 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 20674
39c1b6db
MP		 206752013-04-04 Marek Polacek <polacek@redhat.com>
20676
20677 PR tree-optimization/48186
20678 * predict.c (maybe_hot_frequency_p): Return false if
20679 HOT_BB_FREQUENCY_FRACTION is 0.
20680 (cgraph_maybe_hot_edge_p): Likewise.
20681
314f64eb
RB		 206822013-04-04 Richard Biener <rguenther@suse.de>
20683
20684 PR tree-optimization/56826
20685 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20686 more accurately.
20687
51a905b2
RB		 206882013-04-04 Richard Biener <rguenther@suse.de>
20689
20690 PR tree-optimization/56213
20691 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 20692 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 20693
f1bf4f3a
RB		 206942013-04-04 Richard Biener <rguenther@suse.de>
20695
20696 PR tree-optimization/56837
20697 * tree-loop-distribution.c (classify_partition): For non-zero
20698 values require that the value has the same precision as its
20699 mode to be useful as memset value.
20700
0bca7ded
NC		 207012013-04-03 Nick Clifton <nickc@redhat.com>
20702
33159866 20703 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 20704 (fmssf4): Use fmsf.s on E3V5 architectures.
20705 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20706 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20707
b4019227
JL		 207082013-04-03 Jeff Law <law@redhat.com>
20709
20710 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20711 (lra-eliminations.o): Likewise.
20712
f57ddb5b
TJ		 207132013-04-03 Teresa Johnson <tejohnson@google.com>
20714
20715 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 20716 compute_working_sets here from profile.c.
f57ddb5b 20717 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 20718 (gcov_working_set_t): Moved typedef here from basic-block.h
20719 (compute_working_set): Declare.
f57ddb5b
TJ		 20720 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20721 (get_working_sets): Renamed from compute_working_set,
33159866 20722 replace most of body with call to new compute_working_sets.
f57ddb5b 20723 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 20724 to get_working_sets.
20725 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 20726 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 20727 to get_working_sets.
f57ddb5b
TJ		 20728 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20729 * gcov-dump.c (dump_working_sets): New function.
20730
12211b99 207312013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 20732
20733 * hwint.c (sext_hwi, zext_hwi): New functions.
20734 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20735 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20736 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20737 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20738 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20739 (sext_hwi, zext_hwi): New functions.
0bca7ded 20740
be672e08
JL		 207412013-04-03 Jeff Law <law@redhat.com>
20742
20743 PR tree-optimization/56799
33159866
UB		 20744 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20745 back test for widening conversion erroneously dropped in prior change.
be672e08 20746
9d821fa5
KT		 207472013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20748
20749 PR target/56809
20750 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20751 instead of next_real_insn.
20752
71c581e7
MP		 207532013-04-03 Marek Polacek <polacek@redhat.com>
20754
20755 PR sanitizer/55702
33159866 20756 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 20757
4a32ef80
KT		 207582013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20759
20760 PR target/56809
20761 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20762 next_real_insn.
20763 (thumb1_output_casesi): Likewise.
20764 (thumb2_output_casesi): Likewise.
20765
1b2253d4
RB		 207662013-04-03 Richard Biener <rguenther@suse.de>
20767
20768 PR tree-optimization/56817
20769 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20770 Split out ...
20771 (tree_unroll_loops_completely_1): ... new function to manually
20772 walk the loop tree, properly defering outer loops of unrolled
20773 loops to later iterations.
20774
38000232
MG		 207752013-04-03 Marc Glisse <marc.glisse@inria.fr>
20776
20777 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20778 (vectorizable_load): Likewise.
20779 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20780 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20781
3fa35298
MG		 207822013-04-03 Marc Glisse <marc.glisse@inria.fr>
20783
20784 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20785 BIT_FIELD_REF.
20786
b3d45ff0
UW		 207872013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20788
20789 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20790
ec9202a8
BC		 207912013-04-03 Bin Cheng <bin.cheng@arm.com>
20792
20793 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20794
6805bd36
MG		 207952013-04-03 Marc Glisse <marc.glisse@inria.fr>
20796
20797 PR tree-optimization/56790
33159866
UB		 20798 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20799 folding.
6805bd36 20800
da694a77
MG		 208012013-04-03 Marc Glisse <marc.glisse@inria.fr>
20802
20803 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20804 Handle VEC_MERGE.
20805 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20806 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20807 equal arguments.
20808
4e7d7b3d
JJ		 208092013-04-03 Jakub Jelinek <jakub@redhat.com>
20810
20811 PR c/19449
20812 * tree.h (force_folding_builtin_constant_p): New decl.
20813 * builtins.c (force_folding_builtin_constant_p): New variable.
20814 (fold_builtin_constant_p): Fold immediately also if
20815 force_folding_builtin_constant_p.
20816
e6c9d234
RB		 208172013-04-03 Richard Biener <rguenther@suse.de>
20818
20819 PR tree-optimization/56812
20820 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20821 DRs of the same interleaving chain are independent.
20822
984d07dd
JM		 208232013-04-02 Jason Merrill <jason@redhat.com>
20824
20825 * gdbinit.in (pbb): Use debug fn.
20826
622849c9
LC		 208272013-04-02 Lawrence Crowl <crowl@google.com>
20828
20829 * sese.h (struct ivtype_map_elt_s): Remove unused.
20830 (extern debug_ivtype_map): Remove unused.
20831 (extern eq_ivtype_map_elts): Remove unused.
20832 * sese.c (debug_ivtype_map): Removed unused.
20833 (debug_ivtype_map_1): Removed unused.
20834 (debug_ivtype_elt): Remove unused.
20835 (eq_ivtype_map_elts): Remove unused.
20836
20837
82c0e1a0
KT		 208382013-04-02 Kai Tietz <ktietz@redhat.com>
20839
20840 PR target/52790
20841 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20842 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 20843 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20844 function.
82c0e1a0
KT		 20845 (legitimize_pe_coff_symbol): Likewise.
20846 (is_imported_p): New helper-function.
20847 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20848 for Windows x64 targets.
20849 (ix86_expand_prologue): Optimize for pe-coff targets.
20850 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20851 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20852 medium/large code-model.
20853 (legitimize_pic_address): Likewise.
20854 (legitimize_tls_address): Likewise.
20855 (ix86_expand_call): Likewise.
20856 (x86_output_mi_thunk): Likewise.
20857 (get_dllimport_decl): Add new beimport argument.
20858 (construct_plt_address): Don't assert for x64 pe-coff targets.
20859 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20860 targets.
20861 (SYMBOL_FLAG_STUBVAR): New macro.
20862 (SYMBOL_REF_STUBVAR_P): Likewise.
20863 * config/i386/winnt.c (stub_list): New structure.
20864 (stub_head): New local variable.
20865 (i386_pe_record_stub): New function.
20866 (i386_pe_file_end): Emit refptr-stubs.
20867
5d751b0c
JJ		 208682013-04-02 Jakub Jelinek <jakub@redhat.com>
20869
09bb4c99
JJ		 20870 PR rtl-optimization/56745
20871 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20872 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20873
a1d8947a
JJ		 20874 PR c++/34949
20875 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20876 and both of them are MEM_REFs, just compare first argument for
20877 equality and attempt to deal even with differing offsets.
20878
5d751b0c
JJ		 20879 PR c++/34949
20880 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20881 of gimple_clobber_p to be MEM_REF.
20882 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20883 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20884 after gimplification.
20885 * asan.c (get_mem_ref_of_assignment): Don't instrument
20886 gimple_clobber_p stmts.
20887 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20888 gimple_clobber_p stmt if they have MEM_REF lhs and
20889 are dead because of another gimple_clobber_p stmt.
20890 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20891 gimple_clobber_p stmts like debug stmts.
20892 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20893 that refer to unused VAR_DECLs or uninitialized values.
20894 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20895 gimple_clobber_p stmts if they refer to removed parameters.
20896 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20897 formatting.
20898
e594716a
UB		 208992013-04-02 Uros Bizjak <ubizjak@gmail.com>
20900
20901 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20902 using SWI48 mode attribute.
20903
7a80735b
WM		 209042013-04-02 Wei Mi <wmi@google.com>
20905
20906 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20907 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20908 *<rotate_insn><mode>3_mask in i386.md.
20909
f423a9e4
AI		 209102013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20911
20912 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20913
90eb75f2
RB		 209142013-04-02 Richard Biener <rguenther@suse.de>
20915
20916 PR tree-optimization/56778
20917 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20918 Runtime alias tests are not supported for gather loads.
20919 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20920 stmts referenced from SSA operands before updating SSA form.
20921
d8c69a92
IC		 209222013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20923 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20924
20925 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 20926 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20927 * config/arm/cortex-a53.md: New file.
20928 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20929 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20930 * config/arm/arm.c (arm_issue_rate): Likewise.
20931 * config/arm/arm-tune.md: Regenerate
20932 * config/arm/arm-tables.opt: Regenerate.
20933 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 20934
239eb04c
ZC		 209352013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20936
20937 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20938 non-static link.
20939
c902d3c8
SN		 209402013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20941
d8c69a92
IC		 20942 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20943 scalar load/store operations using B/H registers.
c902d3c8
SN		 20944 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20945
051d0e2f
SN		 209462013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20947
20948 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20949 scalar move.
20950 * config/aarch64/aarch64.c
20951 (aarch64_simd_scalar_immediate_valid_for_move): New.
20952 * config/aarch64/aarch64-protos.h
20953 (aarch64_simd_scalar_immediate_valid_for_move): New.
20954 * config/aarch64/constraints.md (Dh, Dq): New.
20955 * config/aarch64/iterators.md (hq): New.
20956
0ee1e3d9
EB		 209572013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20958
20959 * reorg.c (get_branch_condition): Deal with conditional returns.
20960 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20961
136fb3f7
WM		 209622013-04-01 Wei Mi <wmi@google.com>
20963
20964 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20965 Truncate operand 2 using %b asm operand modifier.
20966 (*<shift_insn><mode>3_mask): Ditto.
20967 (*<rotate_insn><mode>3_mask): Ditto.
20968
6388c738
SB		 209692013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20970
20971 PR middle-end/56798
20972 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20973
f7a4d826
KK		 209742013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20975
20976 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20977 of next_real_insn.
20978 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20979
4b943a49
LC		 209802013-03-30 Lawrence Crowl <crowl@google.com>
20981
20982 * dse.c (clear_alias_sets): Remove never set.
20983 (disqualified_clear_alias_sets): Remove never set.
20984 (clear_alias_mode_pool): Remove never set.
20985 (dse_step0): Remove condition that is never true.
20986 (canon_address): Remove condition that is never true.
20987 (dse_step7): Remove condition that is never true.
20988 (rest_of_handle_dse): Remove condition that is never true.
20989 (rest_of_handle_dse::did_global): Remove never read from above.
20990 (dse_step2_spill): Remove never called from above.
20991 (dse_step5_spill): Remove never called from above.
20992
39718607
SB		 209932013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20994
da5c6bde
SB		 20995 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20996 JUMP_TABLE_DATA changes.
20997 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20998 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20999 (Insns) <jump_table_data>: New entry.
21000 * doc/tm.texi: Regenerate.
21001
39718607
SB		 21002 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21003
21004 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21005 for table jump at the end of a basic block using tablejump_p.
21006 * targhooks.c (default_invalid_within_doloop): Likewise.
21007 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21008 target hook implementation that is identical to the default hook.
21009 (rs6000_invalid_within_doloop): Remove.
21010
21011 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21012 unused variable from tablejump_p call.
21013
21014 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21015 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21016 (INSN_DELETED_P): Likewise.
21017 (emit_jump_table_data): New prototype.
21018 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21019 after 4th as unused.
21020 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21021 * sched-vis.c (print_insn): Likewise.
21022 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21023 insn for compatibility with back ends that use next_active_insn to
21024 identify jump table data.
21025 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21026 (remove_insn): Likewise.
21027 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21028 to be emitted.
21029 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21030 (emit_jump_table_data): New function.
21031
21032 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21033 basic block, a JUMP_TABLE_DATA never is.
21034 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21035 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21036 off from code handling real insns.
21037 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21038 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21039 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21040 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21041 is not a NONDEBUG_INSN_P.
21042 * ira-costs.c (scan_one_insn): Likewise.
21043 * jump.c (mark_all_labels): Likewise.
21044 (mark_jump_label_1): Likewise.
21045 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21046 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21047 a basic block.
21048 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21049 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21050 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21051 JUMP_TABLE_DATA_P insns.
21052 (calculate_elim_costs_all_insns): Likewise.
21053 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21054 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21055 (delete_output_reload): Code style fixups.
21056 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21057 insn flags on this non-insn.
21058 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21059 as scheduling barriers, for pre-change compatibility.
21060 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21061 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21062
21063 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21064 redundant JUMP_TABLE_DATA_P test.
21065 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21066 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21067 (frv_for_each_packet): Likewise.
21068 * config/i386/i386.c (min_insn_size): Likewise.
21069 (ix86_avoid_jump_mispredicts): Likewise.
21070 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21071 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21072 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21073 (mips16_insn_length): Robustify.
21074 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21075 (mips16_split_long_branches): Likewise.
21076 * config/pa/pa.c (pa_combine_instructions): Likewise.
21077 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21078 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21079 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21080 as contributing to pool range lengths.
21081 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21082 Remove redundant JUMP_TABLE_DATA_P test.
21083 (sh_loop_align): Likewise.
21084 (split_branches): Likewise.
21085 (sh_insn_length_adjustment): Likewise.
21086 * config/spu/spu.c (get_branch_target): Likewise.
21087
0208f7da
JH		 210882013-03-29 Jan Hubicka <jh@suse.cz>
21089
21090 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21091 gcov streaming; stream hot bb threshold to ltrans.
21092 * predict.c (get_hot_bb_threshold): Break out from ....
21093 (maybe_hot_count_p): ... here.
21094 (set_hot_bb_threshold): New function.
21095 * lto-section-in.c (lto_section_name): Add profile.
21096 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21097 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21098 and data-streamer.h
21099 (histogram_entry): New structure.
21100 (histogram, histogram_pool): New global vars.
21101 (histogram_hash): New structure.
21102 (histogram_hash::hash): New method.
21103 (histogram_hash::equal): Likewise.
21104 (account_time_size): New function.
21105 (cmp_counts): New function.
21106 (dump_histogram): New function.
21107 (ipa_profile_generate_summary): New function.
21108 (ipa_profile_write_summary): New function.
21109 (ipa_profile_read_summary): New function.
21110 (ipa_profile): Decide on threshold.
e594716a
UB		 21111 (pass_ipa_profile): Add ipa_profile_write_summary and
21112 ipa_profile_read_summary.
0208f7da
JH		 21113 * Makefile.in (ipa.o): Update dependencies.
21114 * lto-streamer.h (LTO_section_ipa_profile): New section.
21115
5a6ccc94
GDR		 211162013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21117
21118 * tree.h (VAR_P): New.
21119
39385fa6
PC		 211202013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21121
21122 PR lto/56777
21123 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21124
34f0d87a
SB		 211252013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21126
21127 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21128 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21129 (control_flow_insn_p): Likewise.
21130 * cfgrtl.c (duplicate_insn_chain): Likewise.
21131 * final.c (get_attr_length_1): Likewise.
21132 (shorten_branches): Likewise.
21133 (final_scan_insn): Likewise.
21134 * function.c (instantiate_virtual_regs): Likewise.
21135 * gcse.c (insert_insn_end_basic_block): Likewise.
21136 * ira-costs.c (scan_one_insn): Likewise.
21137 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21138 * lra.c (check_rtl): Likewise.
21139 * reload1.c (elimination_costs_in_insn): Likewise.
21140 * reorg.c (follow_jumps): Likewise.
21141
21142 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21143 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21144 (thumb_far_jump_used_p): Likewise.
21145 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21146 (workaround_speculation): Likewise.
21147 (add_sched_insns_for_speculation): Likewise.
21148 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21149 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21150 (frv_for_each_packet): Likewise.
21151 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21152 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21153 (final_emit_insn_group_barriers): Likewise.
21154 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21155 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21156 (mips16_insn_length): Likewise.
21157 * config/pa/pa.c (pa_reorg): Likewise.
21158 (pa_combine_instructions): Likewise.
21159 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21160 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21161 (sh_reorg): Likewise.
21162 (split_branches): Likewise.
21163 * config/spu/spu.c (get_branch_target): Likewise.
21164
21165 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21166 JUMP_TABLE_DATA_P.
21167
4ac761b0
KY		 211682013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21169
39385fa6 21170 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 21171 Fix declaration name.
21172
58a49189
LC		 211732013-03-28 Lawrence Crowl <crowl@google.com>
21174
21175 * graphds.h (struct graph.indicies): Remove unused.
21176 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21177 (SCOP_ORIGINAL_PDDRS): Remove unused.
21178 * sese.h (extern insert_loop_close_phis): Removed unused.
21179 (extern insert_guard_phis): Removed unused.
21180 (extern ivtype_map_elt_info): Removed unused.
21181 (new_ivtype_map_elt): Removed unused.
21182 * sese.c (ivtype_map_elt_info): Removed unused.
21183
7b3b6ae4
LC		 211842013-03-28 Lawrence Crowl <crowl@google.com>
21185
21186 * Makefile.in: Add several missing include dependences.
21187 (DUMPFILE_H): New.
21188 (test-dump.o): New. This object is not added to any executable,
21189 but is present for ad-hoc testing.
21190 * bitmap.c
21191 (debug (const bitmap_head_def &)): New.
21192 (debug (const bitmap_head_def *)): New.
21193 * bitmap.h
21194 (extern debug (const bitmap_head_def &)): New.
21195 (extern debug (const bitmap_head_def *)): New.
21196 * cfg.c
21197 (debug (edge_def &)): New.
21198 (debug (edge_def *)): New.
21199 * cfghooks.c
21200 (debug (basic_block_def &)): New.
21201 (debug (basic_block_def *)): New.
21202 * dumpfile.h
21203 (dump_node (const_tree, int, FILE *)): Correct source file.
21204 * dwarf2out.c
21205 (debug (die_struct &)): New.
21206 (debug (die_struct *)): New.
21207 * dwarf2out.h
21208 (extern debug (die_struct &)): New.
21209 (extern debug (die_struct *)): New.
21210 * gimple-pretty-print.c
21211 (debug (gimple_statement_d &)): New.
21212 (debug (gimple_statement_d *)): New.
21213 * gimple-pretty-print.h
21214 (extern debug (gimple_statement_d &)): New.
21215 (extern debug (gimple_statement_d *)): New.
21216 * ira-build.c
21217 (debug (ira_allocno_copy &)): New.
21218 (debug (ira_allocno_copy *)): New.
21219 (debug (ira_allocno &)): New.
21220 (debug (ira_allocno *)): New.
21221 * ira-int.h
21222 (extern debug (ira_allocno_copy &)): New.
21223 (extern debug (ira_allocno_copy *)): New.
21224 (extern debug (ira_allocno &)): New.
21225 (extern debug (ira_allocno *)): New.
21226 * ira-lives.c
21227 (debug (live_range &)): New.
21228 (debug (live_range *)): New.
21229 * lra-int.h
21230 (debug (lra_live_range &)): New.
21231 (debug (lra_live_range *)): New.
21232 * lra-lives.c
21233 (debug (lra_live_range &)): New.
21234 (debug (lra_live_range *)): New.
21235 * omega.c
21236 (debug (omega_pb_d &)): New.
21237 (debug (omega_pb_d *)): New.
21238 * omega.h
21239 (extern debug (omega_pb_d &)): New.
21240 (extern debug (omega_pb_d *)): New.
21241 * print-rtl.c
21242 (debug (const rtx_def &)): New.
21243 (debug (const rtx_def *)): New.
21244 * print-tree.c
21245 (debug_tree (tree): Move within file.
21246 (debug_raw (const tree_node &)): New.
21247 (debug_raw (const tree_node *)): New.
21248 (dump_tree_via_hooks (const tree_node *, int)): New.
21249 (debug (const tree_node &)): New.
21250 (debug (const tree_node *)): New.
21251 (debug_verbose (const tree_node &)): New.
21252 (debug_verbose (const tree_node *)): New.
21253 (debug_head (const tree_node &)): New.
21254 (debug_head (const tree_node *)): New.
21255 (debug_body (const tree_node &)): New.
21256 (debug_body (const tree_node *)): New.
21257 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21258 (debug (vec<tree, va_gc> &)): New.
21259 (debug (vec<tree, va_gc> *)): New.
21260 * rtl.h
21261 (extern debug (const rtx_def &)): New.
21262 (extern debug (const rtx_def *)): New.
21263 * sbitmap.c
21264 (debug_raw (simple_bitmap_def &)): New.
21265 (debug_raw (simple_bitmap_def *)): New.
21266 (debug (simple_bitmap_def &)): New.
21267 (debug (simple_bitmap_def *)): New.
21268 * sbitmap.h
21269 (extern debug (simple_bitmap_def &)): New.
21270 (extern debug (simple_bitmap_def *)): New.
21271 (extern debug_raw (simple_bitmap_def &)): New.
21272 (extern debug_raw (simple_bitmap_def *)): New.
21273 * sel-sched-dump.c
21274 (debug (vinsn_def &)): New.
21275 (debug (vinsn_def *)): New.
21276 (debug_verbose (vinsn_def &)): New.
21277 (debug_verbose (vinsn_def *)): New.
21278 (debug (expr_def &)): New.
21279 (debug (expr_def *)): New.
21280 (debug_verbose (expr_def &)): New.
21281 (debug_verbose (expr_def *)): New.
21282 (debug (vec<rtx> &)): New.
21283 (debug (vec<rtx> *)): New.
21284 * sel-sched-dump.h
21285 (extern debug (vinsn_def &)): New.
21286 (extern debug (vinsn_def *)): New.
21287 (extern debug_verbose (vinsn_def &)): New.
21288 (extern debug_verbose (vinsn_def *)): New.
21289 (extern debug (expr_def &)): New.
21290 (extern debug (expr_def *)): New.
21291 (extern debug_verbose (expr_def &)): New.
21292 (extern debug_verbose (expr_def *)): New.
21293 (extern debug (vec<rtx> &)): New.
21294 (extern debug (vec<rtx> *)): New.
21295 * sel-sched-ir.h
21296 (_list_iter_cond_expr): Make inline instead of static.
21297 * sreal.c
21298 (debug (sreal &)): New.
21299 (debug (sreal *)): New.
21300 * sreal.h
21301 (extern debug (sreal &)): New.
21302 (extern debug (sreal *)): New.
21303 * tree.h
21304 (extern debug_raw (const tree_node &)): New.
21305 (extern debug_raw (const tree_node *)): New.
21306 (extern debug (const tree_node &)): New.
21307 (extern debug (const tree_node *)): New.
21308 (extern debug_verbose (const tree_node &)): New.
21309 (extern debug_verbose (const tree_node *)): New.
21310 (extern debug_head (const tree_node &)): New.
21311 (extern debug_head (const tree_node *)): New.
21312 (extern debug_body (const tree_node &)): New.
21313 (extern debug_body (const tree_node *)): New.
21314 (extern debug (vec<tree, va_gc> &)): New.
21315 (extern debug (vec<tree, va_gc> *)): New.
21316 * tree-cfg.c
21317 (debug (struct loop &)): New.
21318 (debug (struct loop *)): New.
21319 (debug_verbose (struct loop &)): New.
21320 (debug_verbose (struct loop *)): New.
21321 * tree-dump.c: Add header dependence.
21322 * tree-flow.h
21323 (extern debug (struct loop &)): New.
21324 (extern debug (struct loop *)): New.
21325 (extern debug_verbose (struct loop &)): New.
21326 (extern debug_verbose (struct loop *)): New.
21327 * tree-data-ref.c
21328 (debug (data_reference &)): New.
21329 (debug (data_reference *)): New.
21330 (debug (vec<data_reference_p> &)): New.
21331 (debug (vec<data_reference_p> *)): New.
21332 (debug (vec<ddr_p> &)): New.
21333 (debug (vec<ddr_p> *)): New.
21334 * tree-data-ref.h
21335 (extern debug (data_reference &)): New.
21336 (extern debug (data_reference *)): New.
21337 (extern debug (vec<data_reference_p> &)): New.
21338 (extern debug (vec<data_reference_p> *)): New.
21339 (extern debug (vec<ddr_p> &)): New.
21340 (extern debug (vec<ddr_p> *)): New.
21341 * tree-ssa-alias.c
21342 (debug (pt_solution &)): New.
21343 (debug (pt_solution *)): New.
21344 * tree-ssa-alias.h
21345 (extern debug (pt_solution &)): New.
21346 (extern debug (pt_solution *)): New.
21347 * tree-ssa-alias.c
21348 (debug (_var_map &)): New.
21349 (debug (_var_map *)): New.
21350 (debug (tree_live_info_d &)): New.
21351 (debug (tree_live_info_d *)): New.
21352 * tree-ssa-alias.h
21353 (extern debug (_var_map &)): New.
21354 (extern debug (_var_map *)): New.
21355 (extern debug (tree_live_info_d &)): New.
21356 (extern debug (tree_live_info_d *)): New.
21357
be77e1e5
JH		 213582013-03-28 Jan Hubicka <jh@suse.cz>
21359
21360 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21361
777e6976
IB		 213622013-03-28 Ian Bolton <ian.bolton@arm.com>
21363
21364 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21365 record only when desired or required.
21366
9cd347ae
UB		 213672013-03-28 Uros Bizjak <ubizjak@gmail.com>
21368
21369 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21370 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21371
33159866 213722013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 21373
21374 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21375 (*andsi3_compare0_uxtw): New pattern.
21376 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21377 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21378
89ab31c1
JH		 213792013-03-28 Jan Hubicka <jh@suse.cz>
21380
21381 * data-streamer-in.c (streamer_read_gcov_count): New function.
21382 * gimple-streamer-out.c: Include value-prof.h.
21383 (output_gimple_stmt): Output histogram.
21384 (output_bb): Use streamer_write_gcov_count.
21385 * value-prof.c: Include data-streamer.h
21386 (dump_histogram_value): Add HIST_TYPE_MAX.
21387 (stream_out_histogram_value): New function.
21388 (stream_in_histogram_value): New function.
21389 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21390 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21391 * data-streamer-out.c (streamer_write_gcov_count): New function.
21392 (streamer_write_gcov_count_stream): New function.
21393 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21394 (lto_output_node): Likewise.
21395 (input_node, input_edge): Likewise.
21396 * lto-streamer-out.c (output_cfg): Update streaming.
21397 * lto-streamer-in.c (input_cfg): Likewise.
21398 * data-streamer.h (streamer_write_gcov_count,
21399 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21400 * gimple-streamer-in.c: Include value-prof.h
21401 (input_gimple_stmt): Input histograms.
21402 (input_bb): Update profile streaming.
21403
e594716a 214042013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 21405
33159866 21406 * genmodes.c (emit_max_int): New function.
8697be17 21407 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 21408 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21409 Added doc.
8697be17 21410 * machmode.def: Fixed comment.
89ab31c1 21411
e594716a 214122013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 21413
21414 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 21415 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21416
74946978
MP		 214172013-03-28 Marek Polacek <polacek@redhat.com>
21418 Richard Biener <rguenther@suse.de>
21419
21420 PR tree-optimization/56695
21421 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21422 build signed result of a vector comparison.
21423 * tree-cfg.c (verify_gimple_comparison): Check that a result
21424 of a vector comparison has signed type.
21425
a64b9c26
RB		 214262013-03-28 Richard Biener <rguenther@suse.de>
21427
21428 PR tree-optimization/37021
21429 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21430 do not restrict gaps between groups.
21431 * tree-vect-stmts.c (vectorizable_load): Properly account for
21432 a gap between groups.
21433
a9dc2a2f
EB		 214342013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21435
21436 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21437 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21438 is not enabled.
21439
53cb97f9
GP		 214402013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21441
21442 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21443 * doc/extend.texi (Named Address Spaces): Ditto.
21444 (Variable Attributes): Ditto.
21445
b802ae5c
KT		 214462013-03-27 Kai Tietz <ktietz@redhat.com>
21447
eddae10a
KT		 21448 * config.build: Add support for cygwin x64 target.
21449 * config.gcc: Likewise.
21450 * config.host: Likewise.
21451 * configure.ac: Likewise
21452 * configure: Regenerated.
b802ae5c 21453
371e77e3 214542013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 21455
21456 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21457 * config/i386/t-cygwin-w64: New file.
21458 * config/i386/cygwin-w64.h: New file.
21459 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21460 and add support for x64-cygwin target.
21461 (CPP_SPEC): Likewise.
21462 (CXX_WRAP_SPEC_LIST): Undefine before define.
21463 (LIBGCJ_SONAME): Use 15 as version.
21464
f49b33cb
RB		 214652013-03-27 Richard Biener <rguenther@suse.de>
21466
21467 PR tree-optimization/56716
21468 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21469 dumping for ref nodes.
21470
b37a6ce5
MJ		 214712013-03-27 Martin Jambor <mjambor@suse.cz>
21472
21473 PR tree-optimization/55334
21474 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21475 restricted pointers to arrays.
21476
9469b9b2
GDR		 214772013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21478
21479 * Makefile.in (.SUFFIXES): Add .cc.
21480 (.c.o): Apply same recipe for implicit rule .cc.o.
21481
7d24f650
RB		 214822013-03-27 Richard Biener <rguenther@suse.de>
21483
21484 PR tree-optimization/37021
21485 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21486 REALPART/IMAGPART_EXPRs around the supported refs.
21487 * tree-ssa-structalias.c (find_func_aliases): Assume that
21488 floating-point values are not used to transfer pointers.
21489
2f251a05
AI		 214902013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21491
f43245d1
UB		 21492 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21493 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21494 * doc/tm.texi: Regenerate.
21495 * targhooks.h (default_has_ifunc_p): New.
21496 * targhooks.c (default_has_ifunc_p): Ditto.
21497 * config/linux-protos.h: New file.
21498 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21499 hook for linux which disables support of indirect functions in android.
21500 * config/linux-android.c: New file.
21501 * config/t-linux-android.c: Ditto.
21502 * config.gcc: Added new object file linux-android.o.
21503 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21504 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21505 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 21506 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21507 doesn't support indirect functions.
21508 * configure: Regenerate.
21509
78b4e425
BC		 215102013-03-27 Bin Cheng <bin.cheng@arm.com>
21511
21512 PR target/56102
21513 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21514 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21515 mult-word mode.
21516
a5ba7b92
AK		 215172013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21518
21519 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21520
be624986
TG		 215212013-03-27 Terry Guo <terry.guo@arm.com>
21522
21523 * config/arm/arm-cores.def: Added core cortex-r7.
21524 * config/arm/arm-tune.md: Regenerated.
21525 * config/arm/arm-tables.opt: Regenerated.
21526 * doc/invoke.texi: Added entry for core cortex-r7.
21527
0a514f47
WL		 215282013-03-27 Walter Lee <walt@tilera.com>
21529
21530 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21531 double-decrement of next_scratch_regno.
21532
9332b0d2
WL		 215332013-03-27 Walter Lee <walt@tilera.com>
21534
801d9b2a 21535 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 21536 input operands.
21537 (insn_v1mulus): Ditto.
21538 (insn_v2muls): Ditto.
21539
f54ea5dd
WL		 215402013-03-27 Walter Lee <walt@tilera.com>
21541
e594716a 21542 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 21543 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21544
bf60f4ca
WL		 215452013-03-27 Walter Lee <walt@tilera.com>
21546
e594716a 21547 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 21548 (*sibcall_value): Ditto.
21549
e3b51eeb
WL		 215502013-03-27 Walter Lee <walt@tilera.com>
21551
21552 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21553 (insn_mnz_v8qi): ... this ...
21554 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21555 vector equivalent.
21556 (insn_v<n>mnz): Replaced by ...
21557 (insn_v1mnz): ... this ...
21558 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21559 equivalent.
21560 (insn_mz_<mode>): Replaced by ...
21561 (insn_mz_v8qi): ... this ...
21562 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21563 vector equivalent.
21564 (insn_v<n>mz): Replaced by ...
21565 (insn_v1mz): ... this ...
21566 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21567 equivalent.
21568
4fc7b145
EB		 215692013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21570
21571 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21572
6f33817e
RM		 215732013-03-26 Roland McGrath <mcgrathr@google.com>
21574
21575 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21576 than fprintf with a non-constant, non-format string.
21577
93a4145b
UB		 215782013-03-26 Uros Bizjak <ubizjak@gmail.com>
21579
21580 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21581 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21582 operand 0 predicate.
21583 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21584 attribute. Use general_x64nomem_operand as operand 1 predicate.
21585 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21586 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21587 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21588 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21589 (mov<mode>_insv_1): Remove expander. Merge insn with
21590 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21591 Use general_x64nomem_operand as operand 1 predicate.
21592 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21593 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21594 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21595 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21596 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21597 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21598 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21599 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21600 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21601 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21602 (general_x64nomem_operand): Ditto.
21603
c6a9ed5a 216042013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 21605
21606 * config/rtems.opt: Add -pthread option.
21607
31b0a960
RB		 216082013-03-26 Richard Biener <rguenther@suse.de>
21609
93a4145b
UB		 21610 * alias.c (find_base_term): Avoid redundant and not used recursion.
21611 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 21612 (true_dependence_1): Compute and pass base terms to base_alias_check.
21613 (write_dependence_p): Likewise.
21614 (may_alias_p): Likewise.
21615
79517551
SN		 216162013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21617
21618 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21619 PC-relative load in SI modes and above only.
21620
a76213b9
XQ		 216212013-03-26 Xinyu Qi <xyqi@marvell.com>
21622
21623 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 21624 * config/arm/iwmmxt.md (WCGR0): Update.
21625 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21626
37ff9355
UB		 216272013-03-26 Uros Bizjak <ubizjak@gmail.com>
21628
21629 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21630 Use x64 and nox64 isa attributes.
21631
30b0317c
RB		 216322013-03-26 Richard Biener <rguenther@suse.de>
21633
21634 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21635 alignment computations and rely on get_object_alignment_1
21636 for the !TYPE_P case.
21637 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21638
cbcdb140
WL		 216392013-03-26 Walter Lee <walt@tilera.com>
21640
21641 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21642 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21643
dd884525 216442013-03-25 Jeff Law <law@redhat.com>
be672e08 21645
33159866
UB		 21646 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21647 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 21648 wrong version of prior patch.
21649
0051d3ec
WL		 216502013-03-25 Walter Lee <walt@tilera.com>
21651
21652 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21653 TILEGX_INSN_SHUFFLEBYTES1.
21654 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21655 shufflebytes1.
21656 (tilegx_builtins): Ditto.
21657 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21658
065a3605
WL		 216592013-03-25 Walter Lee <walt@tilera.com>
21660
21661 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 21662 (floatunssisf2): New pattern.
21663 (floatsidf2): New pattern.
21664 (floatunssidf2): New pattern.
065a3605 21665
5b2a3c82
WL		 216662013-03-25 Walter Lee <walt@tilera.com>
21667
21668 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21669 tests for constraint J, K, N, P.
21670
192ea533
WL		 216712013-03-25 Walter Lee <walt@tilera.com>
21672
21673 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21674 Use indirect/pcrel encoding.
21675 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21676 Ditto.
21677
855e0d0b
SE		 216782013-03-25 Steve Ellcey <sellcey@mips.com>
21679
21680 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21681 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21682 * config/mips/mips.c (mips_option_override): Set IMADD default.
21683 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21684 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21685 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21686 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21687 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21688
39e45653
JL		 216892013-03-25 Jeff Law <law@redhat.com>
21690
21691 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21692 slightly to avoid creating and folding useless trees. Simplify
21693 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21694
7482c470
UB		 216952013-03-25 Uros Bizjak <ubizjak@gmail.com>
21696
21697 * config/i386/i386.md (*zero_extendsidi2): Merge with
21698 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21699 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21700 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21701 targets. Clarify comment.
21702
4a53743e
MJ		 217032013-03-25 Martin Jambor <mjambor@suse.cz>
21704
21705 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21706 pass-through jump functions differently.
21707 (ipa_read_jump_function): Likewise. Also use setter functions to set
21708 up jump functions.
21709
162712de
MJ		 217102013-03-25 Martin Jambor <mjambor@suse.cz>
21711
21712 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21713 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21714 process it.
21715 (ipa_get_indirect_edge_target): New function.
21716 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21717 ipa_get_indirect_edge_target. Update all callers.
21718 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21719 ipa_get_indirect_edge_target_1 instead of calling
21720 ipa_get_indirect_edge_target.
21721 (create_specialized_node): Pass aggvlas to
21722 ipcp_discover_new_direct_edges.
21723
a5a4c20a
KT		 217242013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21725
21726 * config/arm/arm.md (f_sels, f_seld): New types.
21727 (*cmov<mode>): New pattern.
21728 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21729 predicate.
21730
f992b9fc
KT		 217312013-03-25 Kai Tietz <ktietz@redhat.com>
21732
21733 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21734 POSIX-printf for mingw-hosted builds.
21735
c16fd676
RB		 217362013-03-25 Richard Biener <rguenther@suse.de>
21737
21738 PR middle-end/56694
21739 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21740 must-not-throw stmt location.
21741
5ad29f12
KT		 217422013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21743
21744 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21745 Emit load-acquire versions when acq is true.
21746 (arm_emit_store_exclusive): Add rel parameter.
21747 Emit store-release versions when rel is true.
21748 (arm_split_compare_and_swap): Use acquire-release instructions
21749 instead.
21750 of barriers when appropriate.
21751 (arm_split_atomic_op): Likewise.
21752 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21753 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21754 (VUNSPEC_SLX): Likewise.
21755 (VUNSPEC_LDA): Likewise.
21756 (VUNSPEC_STL): Likewise.
21757 * config/arm/sync.md (atomic_load<mode>): New pattern.
21758 (atomic_store<mode>): Likewise.
21759 (arm_load_acquire_exclusive<mode>): Likewise.
21760 (arm_load_acquire_exclusivesi): Likewise.
21761 (arm_load_acquire_exclusivedi): Likewise.
21762 (arm_store_release_exclusive<mode>): Likewise.
21763
03a7dddb
CM		 217642013-03-25 Catherine Moore <clm@codesourcery.com>
21765
21766 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21767 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21768 * config/mip/predicates.md (lwsp_swsp_operand,
21769 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21770 sb16_operand, db4_operand, db7_operand, ib3_operand,
21771 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21772 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21773 andi16_operand): New predicates.
21774 * config/mips/mips.md (compression): New attribute.
21775 (enabled): New attribute.
21776 (length): Consider compression in computing length.
21777 (shift_compression): New code attribute.
21778 (*add<mode>3): New operands. Record compression.
21779 (sub<mode>3): Likewise.
21780 (one_cmpl<mode>2): Likewise.
21781 (*and<mode>3): Likewise.
21782 (*ior<mode>3): Likewise.
21783 (unnamed pattern for xor): Likewise.
21784 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21785 (*<optab><mode>3): Likewise.
21786 (*mov<mode>_internal: Likewise.
21787 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21788 (mips_unsigned_immediate_p): New.
21789 (umips_lwsp_swsp_address_p): New.
21790 (m16_based_address_p): New.
21791 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21792 (mips_unsigned_immediate_p): New prototype.
21793 (lwsp_swsp_address_p): New prototype.
21794 (m16_based_address_p): New prototype.
21795 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21796 (mips_signed_immediate_p): New function.
21797 (m16_based_address_p): New function.
21798 (lwsp_swsp_address_p): New function.
21799 (mips_print_operand_punctuation): Recognize short delay slot insns
21800 for microMIPS.add<mode>3"
21801
f35c297f 218022013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 21803
f35c297f
KT		 21804 PR target/56720
21805 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21806 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21807
051b9446
RB		 218082013-03-25 Richard Biener <rguenther@suse.de>
21809
21810 PR tree-optimization/56689
21811 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21812 any edge.
21813
374001cb
RB		 218142013-03-25 Richard Biener <rguenther@suse.de>
21815
21816 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21817 of bitmap.
21818 (memory_references): Likewise.
21819 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21820 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21821 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21822 (gather_mem_refs_in_loops): Fold into ...
21823 (analyze_memory_references): ... this. Move initialization
21824 to tree_ssa_lim_initialize.
21825 (fill_always_executed_in): Rename to ...
21826 (fill_always_executed_in_1): ... this.
21827 (fill_always_executed_in): Move contains_call computation to
21828 this new function from ...
21829 (tree_ssa_lim_initialize): ... here.
21830 (tree_ssa_lim): Call fill_always_executed_in.
21831
57895947
EB		 218322013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21833
21834 * postreload.c (reload_combine): Fix code detecting returns.
21835
1e1b18c1
EB		 218362013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21837
21838 * function.c (emit_use_return_register_into_block): On cc0 targets,
21839 do not emit the sequence between cc0 setter and user.
21840
9216baf1
KT		 218412013-03-25 Kai Tietz <ktietz@redhat.com>
21842
1e1b18c1
EB		 21843 * config/i386/predicates.md (local_symbolic_operand): Interpret
21844 dll-imported symbols as none-local.
9216baf1 21845
76421b44
RB		 218462013-03-25 Richard Biener <rguenther@suse.de>
21847
21848 * tree-ssa-loop-im.c (struct depend): Remove.
21849 (struct lim_aux_data): Make depends a vec of gimples.
21850 (free_lim_aux_data): Adjust.
21851 (add_dependency): Likewise.
21852 (set_level): Likewise.
21853
d154bfa2
RB		 218542013-03-25 Richard Biener <rguenther@suse.de>
21855
21856 PR middle-end/56434
21857 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21858 the pointer returned by calls with ECF_MALLOC set.
21859
2cd9804e
UB		 218602013-03-24 Uros Bizjak <ubizjak@gmail.com>
21861
acef5fe0
UB		 21862 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21863
218642013-03-24 Uros Bizjak <ubizjak@gmail.com>
21865
21866 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 21867 using MMXMODE mode iterator.
21868 (*move<mode>_internal): Merge with *movv2sf_internal and
21869 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21870
225ccc68
SB		 218712013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21872
85c0f02d
SB		 21873 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21874 (record_last_mem_set_info): Likewise.
21875
225ccc68
SB		 21876 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21877 of XNEWVEC followed by memset.
21878 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21879
b64925dc
SB		 218802013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21881
21882 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21883 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21884 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21885 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21886 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21887 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21888 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21889 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21890 BARRIER_P instead of GET_CODE.
21891
b9aaf52e
EB		 218922013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21893
21894 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21895 inaccuracy in the probing code.
21896
21897 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21898 (ctrapdi4): Likewise.
21899
66de4d7c
EB		 219002013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21901
21902 * calls.c (expand_call): Add missing guard to code handling return
21903 of non-BLKmode structures in MSB.
21904 * function.c (expand_function_end): Likewise.
21905
4f25c8fb
EB		 219062013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21907
21908 * combine.c (try_combine): Adjust comment. Do not add the set of
21909 insn #0 if the destination indirectly is set or dies in insn #2.
21910 Tidy up code to distribute a new note.
21911
23b7850d
UB		 219122013-03-22 Uros Bizjak <ubizjak@gmail.com>
21913
21914 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21915 also for alternatives 16 and 17.
21916
bed852cf
UB		 219172013-03-22 Uros Bizjak <ubizjak@gmail.com>
21918
21919 * config/i386/sse.md (*mov<mode>_internal): Merge with
21920 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21921 Emit insn template depending on type attribute. Use
21922 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21923 movd instead of movq mnemonic for interunit moves. Rewrite mode
21924 attribute calculation. Remove unit attribute calculation.
21925 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21926 Set prefix_data16 attribute for DImode ssemov types.
21927 Use Ym instead of y for SSE-MMX conversion alternatives.
21928 Reorder operand constraints.
21929
1e8a7937
SB		 219302013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21931
21932 * df.h (df_insn_delete): Adjust prototype.
21933 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21934 and let it decide whether mark the basic block dirty.
21935 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21936 * df-scan.c (df_insn_info_delete): New helper function, split
21937 off from df_insn_delete.
21938 (df_scan_free_bb_info): Use it.
21939 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21940 Likewise.
21941 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21942 that the insn is actually an insn and it has a non-NULL basic block.
21943 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21944
e14d094c
RB		 219452013-03-22 Richard Biener <rguenther@suse.de>
21946
21947 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21948 dep_ref members.
21949 (mem_ref_alloc): Do not allocate them.
21950 (refs_independent_p): Do not query or maintain a cache.
21951
e9cf7316
RB		 219522013-03-22 Richard Biener <rguenther@suse.de>
21953
21954 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21955 (gather_mem_refs_in_loops): Do not compute it.
21956 (analyze_memory_references): Do not allocate it.
21957 (tree_ssa_lim_finalize): Do not free it.
21958 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21959
f046e81b
RB		 219602013-03-22 Richard Biener <rguenther@suse.de>
21961
21962 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 21963
a05c0ddf
IB		 219642013-03-22 Ian Bolton <ian.bolton@arm.com>
21965
21966 * config/aarch64/aarch64.c (aarch64_print_operand): New
21967 format specifier for printing a constant in hex.
21968 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21969 format specifier for printing second operand.
21970
c00217fc
RB		 219712013-03-22 Richard Biener <rguenther@suse.de>
21972
21973 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21974 bitmaps.
21975 (gather_mem_refs_in_loops): Perform store accumulation here.
21976 (create_vop_ref_mapping_loop): Remove.
21977 (create_vop_ref_mapping): Likewise.
21978 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 21979 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 21980 (record_indep_loop): Remove.
21981 (record_dep_loop): New function.
21982 (ref_indep_loop_p_1): Adjust to only walk over references
21983 in the loop, not its subloops.
21984 (ref_indep_loop_p): Rename to ...
21985 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21986 maintaining a more fine-grained cache.
21987 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21988 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21989
15d19bf8
RB		 219902013-03-22 Richard Biener <rguenther@suse.de>
21991
21992 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21993 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21994 aggregate mem_ref_loc.
21995 (free_mem_ref_locs): Inline into ...
21996 (memref_free): ... this and adjust.
21997 (mem_ref_alloc): Adjust.
21998 (mem_ref_locs_alloc): Remove.
21999 (record_mem_ref_loc): Adjust.
22000 (get_all_locs_in_loop): Rewrite into ...
22001 (for_all_locs_in_loop): ... this iterator.
22002 (rewrite_mem_ref_loc): New functor.
22003 (rewrite_mem_refs): Use for_all_locs_in_loop.
22004 (sm_set_flag_if_changed): New functor.
22005 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22006 (ref_always_accessed): New functor.
22007 (ref_always_accessed_p): Use for_all_locs_in_loop.
22008
6f37411d
MG		 220092013-03-21 Marc Glisse <marc.glisse@inria.fr>
22010
22011 * tree-pass.h (PROP_gimple_lvec): New.
22012 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22013 (init_optimization_passes): Move pass_lower_vector.
22014 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22015 PROP_gimple_lvec.
22016 (pass_lower_vector): Provide PROP_gimple_lvec.
22017 (pass_lower_vector_ssa): Likewise.
22018 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22019
511ef689
MW		 220202013-03-21 Mark Wielaard <mjw@redhat.com>
22021
22022 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22023
cc1df30b
UB		 220242013-03-21 Uros Bizjak <ubizjak@gmail.com>
22025
22026 * config/i386/i386.md (*movdi_internal): Disparage slightly
22027 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22028 conversion alternatives.
22029
c845cfe1
JJ		 220302013-03-21 Jakub Jelinek <jakub@redhat.com>
22031
37e99116
JJ		 22032 PR middle-end/48087
22033 * diagnostic.def (DK_WERROR): New kind.
22034 * diagnostic.h (werrorcount): Define.
22035 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22036 promoted to DK_ERROR, increment DK_WERROR counter instead of
22037 DK_ERROR counter.
22038 * toplev.c (toplev_main): Call print_ignored_options even if
22039 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22040 even if just werrorcount is non-zero.
22041
c845cfe1
JJ		 22042 PR debug/55608
22043 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22044 on failure.
22045 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22046 (string_cst_pool_decl): New function.
22047 (optimize_one_addr_into_implicit_ptr): New function.
22048 (resolve_addr_in_expr): Optimize DWARF location expression
22049 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22050 which doesn't live in memory, but has DW_AT_location or
22051 DW_AT_const_value, or refers to a string literal, into
22052 DW_OP_GNU_implicit_pointer.
22053 (optimize_location_into_implicit_ptr): New function.
22054 (resolve_addr): If removing DW_AT_location of a variable because
22055 it was DW_OP_addr of address of the variable, but the variable doesn't
22056 live in memory, try to emit const value attribute for the initializer.
22057
08e0cda6
MG		 220582013-03-21 Marc Glisse <marc.glisse@inria.fr>
22059
22060 * tree.h (VECTOR_TYPE_P): New macro.
22061 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22062 TYPE_MODE): Use it.
22063 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22064 VEC_COND_EXPR cannot be lvalues.
22065 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22066
d08633b4
MG		 220672013-03-21 Marc Glisse <marc.glisse@inria.fr>
22068
22069 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22070 Restrict the transformation to equal modes.
22071
e6647190
RB		 220722013-03-21 Richard Biener <rguenther@suse.de>
22073
22074 PR tree-optimization/39326
22075 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22076 (MEM_ANALYZABLE): Adjust.
22077 (record_mem_ref_loc): Move bitmap ops ...
22078 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22079 unanalyzable refs, do not record locations for it.
22080 (analyze_memory_references): Allocate ref zero as shared
22081 unanalyzable ref.
22082 (refs_independent_p): Do not test for unanalyzed mems here.
22083 (ref_indep_loop_p_1): Special-case disambiguation against
22084 the unanalyzed ref.
cc1df30b 22085 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 22086
65074f54
CL		 220872013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22088
22089 * config/arm/arm-protos.h (tune_params): Add
22090 prefer_neon_for_64bits field.
22091 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22092 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22093 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22094 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22095 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22096 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22097 (arm_option_override): Handle -mneon-for-64bits new option.
22098 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22099 (prefer_neon_for_64bits): Declare new variable.
22100 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 22101 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 22102 (arch_enabled): Handle new arch types. Remove support for onlya8
22103 and nota8.
22104 (one_cmpldi2): Use new arch names.
bc5faa5b 22105 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 22106 * config/arm/arm.opt (mneon-for-64bits): Add option.
22107 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22108 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22109 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22110 of onlya8.
22111 * doc/invoke.texi (-mneon-for-64bits): Document.
22112
5a2d2a79
RB		 221132013-03-21 Richard Biener <rguenther@suse.de>
22114
22115 PR tree-optimization/39326
22116 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22117 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 22118 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 22119
5abe1e05
RB		 221202013-03-21 Richard Biener <rguenther@suse.de>
22121
22122 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22123 (vect_insert_into_interleaving_chain): Likewise.
22124 (vect_drs_dependent_in_basic_block): Inline ...
22125 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22126 split out from ...
22127 (vect_analyze_data_ref_dependence): ... here. Simplify.
22128 (vect_check_interleaving): Simplify.
22129 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22130 (vect_slp_analyze_data_ref_dependences): ... this new function.
22131 (dr_group_sort_cmp): New function.
22132 (vect_analyze_data_ref_accesses): Compute data-reference groups
22133 here instead of in vect_analyze_data_ref_dependence. Use
22134 a more efficient algorithm.
22135 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22136 vect_slp_analyze_data_ref_dependences. Call
22137 vect_analyze_data_ref_accesses earlier.
22138 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22139 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22140 (vect_slp_analyze_data_ref_dependences): New prototype.
22141
cad1735b
RB		 221422013-03-21 Richard Biener <rguenther@suse.de>
22143
22144 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22145 ref is stored in the loop.
22146 (find_refs_for_sm): Walk only over all stores.
22147 (store_motion_loop): Allocate from lim_bitmap_obstack.
22148 (store_motion): Likewise.
22149
141310ef
RB		 221502013-03-21 Richard Biener <rguenther@suse.de>
22151
22152 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22153 Update virtual SSA form.
22154
5022315a
RO		 221552013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22156
22157 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22158 * configure: Regenerate.
22159 * config.in: Regenerate.
22160 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22161 if !HAVE_LD_EH_FRAME_CIEV3.
22162
50d4421c
RB		 221632013-03-21 Richard Biener <rguenther@suse.de>
22164
22165 * tree-cfg.c (verify_expr_no_block): New function.
22166 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22167 nor DECL_VALUE_EXPR have locations with associated blocks.
22168 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22169 (clear_unused_block_pointer): Remove code dealing with
22170 blocks in DECL_DEBUG_EXPR locations.
22171
839b422f
RB		 221722013-03-21 Richard Biener <rguenther@suse.de>
22173
22174 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22175 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22176 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22177 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22178 instead of DECL_DEBUG_EXPR_IS_FROM.
22179 * gimplify.c (gimplify_modify_expr): Likewise.
22180 * tree-cfg.c (verify_expr_location_1): Likewise.
22181 * tree-complex.c (create_one_component_var): Likewise.
22182 * tree-sra.c (create_access_replacement): Likewise.
22183 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22184 (clear_unused_block_pointer): Likewise.
22185 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22186 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22187 * var-tracking.c (var_debug_decl): Likewise.
22188 (track_expr_p): Likewise.
22189 * tree-inline.c (add_local_variables): Likewise. Set
22190 DECL_HAS_DEBUG_EXPR_P after copying it.
22191 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22192 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22193
fe04878d
UB		 221942013-03-21 Uros Bizjak <ubizjak@gmail.com>
22195
22196 PR bootstrap/56656
22197 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22198 * configure: Regenerate.
22199 * config.in: Regenerate.
22200 * config/i386/i386.md (*movdf_internal): Use
22201 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22202 movd instead of movq mnemonic for interunit moves.
22203 (*movdi_internal): Ditto.
22204
bd059b26 222052013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 22206
22207 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22208 (abd<mode>_3): New pattern.
22209 (aba<mode>_3): New pattern.
22210 (fabd<mode>_3): New pattern.
22211
bd059b26 222122013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 22213
22214 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22215 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22216 occurrence of REGISTER_PREFIX as its empty string.
22217
4f1f78b9
JL		 222182013-03-20 Jeff Law <law@redhat.com>
22219
22220 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22221 addititional equivalences for equality comparisons between an SSA_NAME
22222 and a constant where the SSA_NAME was set from a widening conversion.
22223
327a1118
WL		 222242013-03-20 Walter Lee <walt@tilera.com>
22225
22226 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22227
813e0036
UB		 222282013-03-20 Uros Bizjak <ubizjak@gmail.com>
22229
22230 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22231 depending on type attribute.
22232 (*movti_internal): Ditto.
22233 (*movtf_internal): Ditto.
22234 (*movxf_internal): Ditto.
22235 (*movdf_internal): Ditto.
22236 (*movsf_internal): Ditto.
22237
7cf34aae
UB		 222382013-03-20 Uros Bizjak <ubizjak@gmail.com>
22239
22240 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22241 maybe_vex for sselog1 and ssemov types.
22242 (*movdi_internal): Reorder operand constraints.
22243 (*movsi_internal): Ditto. Set prefix attribute to
22244 maybe_vex for sselog1 and ssemov types.
22245 (*movtf_internal): Set prefix attribute to maybe_vex
22246 for sselog1 and ssemov types.
22247 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22248 DImode ssemov types. Reorder operand constraints.
22249 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22250 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22251 attribute for SImode ssemov types. Reorder operand constraints.
22252
19321415
MJ		 222532013-03-20 Martin Jambor <mjambor@suse.cz>
22254
22255 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22256 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22257
12211b99 222582013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 22259
22260 * config/rs6000/predicates.md (indexed_address, update_address_mem
22261 update_indexed_address_mem): New predicates.
22262 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22263 attribute for load/store instructions.
22264 * config/rs6000/dfp.md (movsd_store): Likewise.
22265 (movsd_load): Likewise.
22266 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22267 (unnamed HI->DI extend define_insn): Likewise.
22268 (unnamed SI->DI extend define_insn): Likewise.
22269 (unnamed QI->SI extend define_insn): Likewise.
22270 (unnamed QI->HI extend define_insn): Likewise.
22271 (unnamed HI->SI extend define_insn): Likewise.
22272 (unnamed HI->SI extend define_insn): Likewise.
22273 (extendsfdf2_fpr): Likewise.
22274 (movsi_internal1): Likewise.
22275 (movsi_internal1_single): Likewise.
22276 (movhi_internal): Likewise.
22277 (movqi_internal): Likewise.
22278 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22279 attribute for load/store instructions.
22280 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22281 instructions.
22282 (mov<mode>_softfloat): Likewise.
22283 (mov<mode>_hardfloat32): Likewise.
22284 (mov<mode>_hardfloat64): Likewise.
22285 (mov<mode>_softfloat64): Likewise.
22286 (movdi_internal32): Likewise.
22287 (movdi_internal64): Likewise.
22288 (probe_stack_<mode>): Likewise.
22289
222902013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22291
22292 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22293 floating point, and decimal floating point to reload iterator.
22294
22295 * config/rs6000/constraints.md (wl constraint): New constraints to
22296 return FLOAT_REGS if certain options are used to reduce the number
22297 of separate patterns that exist in the file.
22298 (wx constraint): Likewise.
22299 (wz constraint): Likewise.
22300
22301 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22302 -mdebug=reg, print wg, wl, wx, and wz constraints.
22303 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22304 Initialize the reload functions for 64-bit binary/decimal floating
22305 point types.
22306 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22307 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22308 create the buffer on the stack to overcome not having a 32-bit
22309 load and store.
22310 (rs6000_emit_move): Likewise.
22311 (rs6000_secondary_memory_needed_rtx): Likewise.
22312 (rs6000_alloc_sdmode_stack_slot): Likewise.
22313 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22314 via xxlxor, just like DFmode 0.0.
22315
22316 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22317 define as 1 if we are running on a power7 or newer.
22318 (enum r6000_reg_class_enum): Add new constraints.
22319
22320 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22321 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22322 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 22323 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 22324 (movsd splitter): Likewise.
22325 (movsd_hardfloat): Likewise.
22326 (movsd_softfloat): Likewise.
22327
22328 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22329 binary and decimal floating point moves.
22330 (fmove_ok): New attributes to combine binary and decimal floating
22331 point moves, and to combine power6x (mfpgpr) moves along normal
22332 floating moves.
22333 (real_value_to_target): Likewise.
22334 (f32_lr): Likewise.
22335 (f32_lm): Likewise.
22336 (f32_li): Likewise.
22337 (f32_sr): Likewise.
22338 (f32_sm): Likewise.
22339 (f32_si): Likewise.
22340 (movsf): Combine binary and decimal floating point moves. Combine
22341 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 22342 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 22343 (mov<mode> for SFmode/SDmode); Likewise.
22344 (SFmode/SDmode splitters): Likewise.
22345 (movsf_hardfloat): Likewise.
22346 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22347 (movsf_softfloat): Likewise.
22348 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22349
22350 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22351 wx and wz constraints.
22352
22353 * config/rs6000/constraints.md (wg constraint): New constraint to
22354 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22355
22356 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22357 constraint.
22358
22359 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22360 -mdebug=reg, print wg, wl, wx, and wz constraints.
22361 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22362 Initialize the reload functions for 64-bit binary/decimal floating
22363 point types.
22364 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22365 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22366 create the buffer on the stack to overcome not having a 32-bit
22367 load and store.
22368 (rs6000_emit_move): Likewise.
22369 (rs6000_secondary_memory_needed_rtx): Likewise.
22370 (rs6000_alloc_sdmode_stack_slot): Likewise.
22371 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22372 via xxlxor, just like DFmode 0.0.
22373
c6d5ff83
MM		 22374 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22375 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22376 with other moves by using conditional constraits (wg). Use LFIWZX
22377 and STFIWX for loading SDmode on power7.
22378 (movdd splitters): Likewise.
22379 (movdd_hardfloat32): Likewise.
22380 (movdd_softfloat32): Likewise.
22381 (movdd_hardfloat64_mfpgpr): Likewise.
22382 (movdd_hardfloat64): Likewise.
22383 (movdd_softfloat64): Likewise.
22384
22385 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22386 64-bit binary and decimal floating point moves.
22387 (FMOVE64X): Likewise.
22388 (movdf): Combine 64-bit binary and decimal floating point moves.
22389 Combine power6x (mfpgpr) moves with other moves by using
22390 conditional constraits (wg).
22391 (mov<mode> for DFmode/DDmode): Likewise.
22392 (DFmode/DDmode splitters): Likewise.
22393 (movdf_hardfloat32): Likewise.
22394 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22395 (movdf_softfloat32): Likewise.
22396 (movdf_hardfloat64_mfpgpr): Likewise.
22397 (movdf_hardfloat64): Likewise.
22398 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22399 (movdf_softfloat64): Likewise.
22400 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22401 (reload_<mode>_load): Move to later in the file so they aren't in
22402 the middle of the floating point move insns.
22403 (reload_<mode>_store): Likewise.
22404
22405 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22406 constraint.
22407
22408 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22409 constraint if -mdebug=reg.
bd059b26
UB		 22410 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22411 Enable using dd reload support if needed.
c6d5ff83
MM		 22412
22413 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22414 binary and decimal floating point moves in rs6000.md.
22415 (movtd_internal): Likewise.
22416
22417 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22418 decimal floating point moves.
22419 (movtf): Likewise.
22420 (movtf_internal): Likewise.
22421 (mov<mode>_internal, TDmode/TFmode): Likewise.
22422 (movtf_softfloat): Likewise.
22423 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22424
22425 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22426 movdi_internal64, using wg constraint for move direct operations.
22427 (movdi_internal64): Likewise.
22428
22429 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22430 MODES_TIEABLE_P for selected modes. Print the numerical value of
22431 the various virtual registers. Use GPR/FPR first/last values,
22432 instead of hard coding the register numbers. Print which modes
22433 have reload functions registered.
bd059b26
UB		 22434 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22435 settings before/after setting cpu, target and subtarget settings.
22436 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22437 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 22438 (rs6000_secondary_reload_fail): Likewise.
22439 (rs6000_secondary_reload_inner): Likewise.
22440
22441 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22442 macros for first/last GPR and FPR registers.
22443 (LAST_GPR_REGNO): Likewise.
22444 (FIRST_FPR_REGNO): Likewise.
22445 (LAST_FPR_REGNO): Likewise.
22446
22447 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22448 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22449 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22450 (vcond<mode><mode>): Likewise.
22451 (vcondu<mode><mode>): Likewise.
22452 (vector_gtu<mode>): Likewise.
22453 (vector_gte<mode>): Likewise.
22454 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22455 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 22456 (ior<mode>3): Likewise.
22457 (and<mode>3): Likewise.
22458 (one_cmpl<mode>2): Likewise.
22459 (nor<mode>3): Likewise.
22460 (andc<mode>3): Likewise.
22461
22462 * config/rs6000/constraints.md (wt constraint): New constraint
22463 that returns VSX_REGS if TImode is allowed in VSX registers.
22464
22465 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22466 constant under VSX.
22467
22468 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22469 similar to TImode, but it is restricted to being in the GPRs.
22470
22471 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22472 TImode to occupy a single VSX register.
22473
22474 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22475 -mvsx-timode for power7/power8.
22476 (power7 cpu): Likewise.
22477 (power8 cpu): Likewise.
22478
22479 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22480 sure that TFmode/TDmode take up two registers if they are ever
22481 allowed in the upper VSX registers.
22482 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22483 registers.
22484 (rs6000_init_hard_regno_mode_ok): Likewise.
22485 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22486 constraint. Print if LRA is turned on.
22487 (rs6000_option_override_internal): Give an error if -mvsx-timode
22488 and VSX is not enabled.
22489 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22490 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22491 to reg+offset addressing. Use PTImode when checking offset
22492 addresses for validity.
22493 (reg_offset_addressing_ok_p): Likewise.
22494 (rs6000_legitimate_offset_address_p): Likewise.
22495 (rs6000_legitimize_address): Likewise.
22496 (rs6000_legitimize_reload_address): Likewise.
22497 (rs6000_legitimate_address_p): Likewise.
22498 (rs6000_eliminate_indexed_memrefs): Likewise.
22499 (rs6000_emit_move): Likewise.
22500 (rs6000_secondary_reload): Likewise.
22501 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22502 reloads to fpr registers to continue to use reg+offset addressing,
22503 but 64-bit reloads to altivec registers need reg+reg addressing.
22504 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22505 it. Treat LO_SUM like a PLUS operation.
22506 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22507 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 22508 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22509 registers to share a register with a smaller sized type, since VSX
22510 puts scalars in the upper 64-bits.
22511 (print_operand): Add support for PTImode.
22512 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22513 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22514 registers, but don't have arithmetic support.
22515 (rs6000_memory_move_cost): Add test for VSX.
22516 (rs6000_opt_masks): Add -mvsx-timode.
22517
22518 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22519 for TImode.
22520 (VSs): Likewise.
22521 (VSr): Use wt constraint for TImode.
22522 (VSv): Drop TImode support.
22523 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22524 (vsx_movti_64bit): Likewise.
22525 (vsx_movti_32bit): Likewise.
22526 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22527 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22528 one '?' on the appropriate output constraint. Do not allow TImode
22529 logical operations on 32-bit systems.
22530 (vsx_ior<mode>3): Likewise.
22531 (vsx_xor<mode>3): Likewise.
22532 (vsx_one_cmpl<mode>2): Likewise.
22533 (vsx_nor<mode>3): Likewise.
22534 (vsx_andc<mode>3): Likewise.
22535 (vsx_concat_<mode>): Likewise.
22536 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22537
22538 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22539 OPTION_MASK_VSX_TIMODE.
22540 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22541 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22542
22543 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22544 (TI2 iterator): New iterator for TImode, PTImode.
22545 (wd mode attribute): Add values for vector types.
bd059b26
UB		 22546 (movti_string): Replace TI move operations with operations for TImode
22547 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 22548 (mov<mode>_string, TImode/PTImode): Likewise.
22549 (movti_ppc64): Likewise.
22550 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22551 (TI mode splitters): Likewise.
22552
22553 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22554 constraint.
22555
1fc5eced
MG		 225562013-03-20 Marc Glisse <marc.glisse@inria.fr>
22557
22558 PR tree-optimization/56355
22559 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22560 Also handle integers with undefined overflow.
22561
22c4c869
CM		 225622013-03-20 Catherine Moore <clm@codesourcery.com>
22563 Maciej W. Rozycki <macro@codesourcery.com>
22564 Tom de Vries <tom@codesourcery.com>
12211b99 22565 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 22566 Iain Sandoe <iain@codesourcery.com>
22567 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22568 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 22569
22570 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22571 Document new function attributes.
22c4c869
CM		 22572 * doc/invoke.texi (minterlink-compressed, mmicromips,
22573 m14k, m14ke, m14kec): Document new options.
22574 (minterlink-mips16): Update documentation.
22575 * doc/md.texi (ZC, ZD): Document new constraints.
22576 * configure.ac (gcc_cv_as_micromips): Check if linker
22577 supports the .set micromips directive.
22578 * configure: Regenerate.
22579 * config.in: Regenerate.
22580 * config/mips/mips-tables.opt: Regenerate.
22581 * config/mips/micromips.md: New file.
22582 * constraints.md (ZC, ZD): New constraints.
22583 * config/mips/predicates.md (movep_src_register): New predicate.
22584 (movep_src_operand): New predicate.
22585 (non_volatile_mem_operand): New predicate.
22586 * config/mips/mips.md (multimem): New type.
22587 (length): Differentiate between 17-bit and 18-bit branch offsets.
22588 (MOVEP1, MOVEP2): New mode iterator.
33159866 22589 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 22590 (mov_<load>r): Likewise.
22591 (mov_<store>l): Likewise.
22592 (mov_<store>r): Likewise.
22593 (*branch_equality<mode>_inverted): Add microMIPS support.
22594 (*branch_equality<mode>): Likewise.
22595 (*jump_absolute): Likewise.
22596 (indirect_jump_<mode>): Likewise.
22597 (tablejump_<mode>): Likewise.
22598 (<optab>_internal): Likewise.
22599 (sibcall_internal): Likewise.
22600 (sibcall_value_internal): Likewise.
22601 (prefetch): Use constraint ZD.
22602 * config/mips/mips.opt (minterlink-compressed): New option.
22603 (minterlink-mips16): Now an alias for minterlink-compressed.
22604 (mmicromips): New option.
22605 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22606 (compare_and_swap_12): Likewise.
22607 (sync_add<mode>): Likewise.
22608 (sync_<optab>_12): Likewise.
22609 (sync_old_<optab>_12): Likewise.
22610 (sync_new_<optab>_12): Likewise.
22611 (sync_nand_12): Likewise.
22612 (sync_old_nand_12): Likewise.
22613 (sync_new_nand_12): Likewise.
22614 (sync_sub<mode>): Likewise.
22615 (sync_old_add<mode>): Likewise.
22616 (sync_old_sub<mode>): Likewise.
22617 (sync_new_add<mode>): Likewise.
22618 (sync_new_sub<mode>): Likewise.
22619 (sync_<optab><mode>): Likewise.
22620 (sync_old_<optab><mode>): Likewise.
22621 (sync_new_<optab><mode>): Likewise.
22622 (sync_nand<mode>): Likewise.
22623 (sync_old_nand<mode>): Likewise.
22624 (sync_new_nand<mode>): Likewise.
22625 (sync_lock_test_and_set<mode>): Likewise.
22626 (test_and_set_12): Likewise.
22627 (atomic_compare_and_swap<mode>): Likewise.
22628 (atomic_exchange<mode>_llsc): Likewise.
22629 (atomic_fetch_add<mode>_llsc): Likewise.
22630 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22631 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22632 (umips_save_restore_pattern_p): Likewise.
22633 (umips_load_store_pair_p): Likewise.
22634 (umips_output_load_store_pair): Likewise.
22635 (umips_movep_target_p): Likewise.
22636 (umips_12bit_offset_address_p): Likewise.
22637 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22638 (mips_base_mips16): Rename this...
22639 (mips_base_compression_flags): ...to this. Update all uses.
22640 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22641 (mips_mips16_decl_p): Delete.
22642 (mips_nomips16_decl_p): Delete.
33159866 22643 (mips_get_compress_on_flags): New function.
22c4c869
CM		 22644 (mips_get_compress_off_flags): New function.
22645 (mips_get_compress_mode): New function.
22646 (mips_get_compress_on_name): New function.
22647 (mips_get_compress_off_name): New function.
22648 (mips_insert_attributes): Support multiple compression types.
22649 (mips_merge_decl_attributes): Likewise.
22650 (umips_12bit_offset_address_p): New function.
22651 (mips_start_function_definition): Emit .set micromips directive.
22652 (mips_call_may_need_jalx_p): New function.
22653 (mips_function_ok_for_sibcall): Add microMIPS support.
22654 (mips_print_operand_punctuation): Support short delay slots and
22655 compact jumps.
22656 (umips_swm_mask, umips_swm_encoding): New.
22657 (umips_build_save_restore): New function.
22658 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22659 (was_mips16_p): Remove.
22660 (old_compression_mode): New.
22661 (mips_set_compression_mode): New function.
22662 (mips_set_current_function): Add microMIPS support.
22663 (mips_option_override): Likewise.
22664 (umips_save_restore_pattern_p): New function.
22665 (umips_output_save_restore): New function.
22666 (umips_load_store_pair_p_1): New function.
22667 (umips_load_store_pair_p): New function.
22668 (umips_output_load_store_pair_1): New function.
22669 (umips_output_load_store_pair): New function.
22670 (umips_movep_target_p) New function.
22671 (mips_prepare_pch_save): Add microMIPS support.
22672 * config/mips/mips.h (TARGET_COMPRESSION): New.
22673 (TARGET_CPU_CPP_BUILTINS): Update macro
22674 to use new compression flags and to support microMIPS.
22675 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22676 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22677 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22678 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22679 (ASM_SPEC): Support mmicromips and mno-micromips.
22680 (M16STORE_REG_P): New macro.
22681 (MIPS_CALL): Support TARGET_MICROMIPS.
22682 (MICROMIPS_J): New macro.
22683 (mips_base_mips16): Rename this...
22684 (mips_base_compression_flags): ...to this.
22685 (UMIPS_12BIT_OFFSET_P): New macro.
22686 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22687 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 226882013-03-20 Richard Biener <rguenther@suse.de>
22689
22690 PR tree-optimization/56661
22691 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22692 the result does not have to be distinct.
22693
54714c68
RB		 226942013-03-20 Richard Biener <rguenther@suse.de>
22695
22696 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22697 remap_gimple_op_r.
22698
cca1130d 226992013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 22700 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 22701
22702 PR rtl-optimization/56605
22703 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22704
2a930377
UB		 227052013-03-20 Uros Bizjak <ubizjak@gmail.com>
22706
22707 PR bootstrap/56656
22708 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22709 that require movd instead of movq.
22710
d6d305fe
RB		 227112013-03-20 Richard Biener <rguenther@suse.de>
22712
22713 * tree-ssa-structalias.c (struct variable_info): Add pointer
22714 to the first field of an aggregate with sub-vars. Make
22715 this and the pointer to the next subfield its ID.
22716 (vi_next): New function.
22717 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22718 storedanything_id, integer_id): Increment by one.
22719 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22720 get_call_clobber_vi): Adjust.
22721 (solution_set_expand): Simplify and speedup.
22722 (solution_set_add): Inline into ...
22723 (set_union_with_increment): ... this. Adjust accordingly.
22724 (do_sd_constraint): Likewise.
22725 (do_ds_constraint): Likewise.
22726 (do_complex_constraint): Simplify.
22727 (build_pred_graph): Adjust.
22728 (solve_graph): Likewise. Simplify and speedup.
22729 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22730 get_constraint_for_component_ref, get_constraint_for_1,
22731 first_vi_for_offset, first_or_preceding_vi_for_offset,
22732 create_function_info_for, create_variable_info_for_1,
22733 create_variable_info_for, intra_create_variable_infos): Adjust.
22734 (init_base_vars): Push NULL for ID zero.
22735 (compute_points_to_sets): Adjust.
22736
a271b42d
RB		 227372013-03-20 Richard Biener <rguenther@suse.de>
22738
22739 * cfgloop.c (verify_loop_structure): Streamline and avoid
22740 ICEing on corrupt loop tree.
22741 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22742 loop tree.
22743
ebd65954
RB		 227442013-03-20 Richard Biener <rguenther@suse.de>
22745
22746 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22747 check whether an SSA update is needed.
22748
4547b7ee
RS		 227492013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22750
22751 * config/mips/constraints.md (T): Rename to...
22752 (Yf): ...this.
22753 (U): Rename to...
22754 (Yd): ...this.
22755 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22756 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22757
0b8cdc58
IB		 227582013-03-19 Ian Bolton <ian.bolton@arm.com>
22759
22760 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22761 (*subsi3_carryin_uxtw): Likewise.
22762
5977a10d
IB		 227632013-03-19 Ian Bolton <ian.bolton@arm.com>
22764
22765 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22766 (*rorsi3_insn_uxtw): Likewise.
22767
bd83ff2c
IB		 227682013-03-19 Ian Bolton <ian.bolton@arm.com>
22769
22770 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22771 (*extrsi5_insn_uxtw): Likewise.
22772
ba65123c
RB		 227732013-03-19 Richard Biener <rguenther@suse.de>
22774
22775 PR tree-optimization/56273
22776 * passes.c (init_optimization_passes): Move second VRP after DOM.
22777
2eac0476
UB		 227782013-03-19 Uros Bizjak <ubizjak@gmail.com>
22779
22780 * config/i386/i386.md (*movti_internal): Merge from
22781 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22782 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22783 nox64 isa attributes.
22784
6e55eda7
RB		 227852013-03-18 Richard Biener <rguenther@suse.de>
22786
22787 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22788 (unite): Likewise.
22789 (merge_node_constraints): Likewise.
22790 (build_succ_graph): Likewise.
22791 (valid_graph_edge): Inline into single caller.
22792 (unify_nodes): Likewise. Use bitmap_set_bit return value
22793 and cache varinfo.
22794 (scc_visit): Fix formatting and variable use.
22795 (do_sd_constraint): Use gcc_checking_assert.
22796 (do_ds_constraint): Likewise.
22797 (do_complex_constraint): Likewise.
22798 (condense_visit): Likewise. Cleanup.
22799 (dump_pred_graph): New function.
22800 (perform_var_substitution): Dump the pred-graph before
22801 variable substitution.
22802 (find_equivalent_node): Use gcc_checking_assert.
22803 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22804
4bdd44c4
RB		 228052013-03-18 Richard Biener <rguenther@suse.de>
22806
22807 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22808 Remove cond_expr_stmt_list argument and do not gimplify the
22809 built expression.
22810 (vect_loop_versioning): Adjust.
22811 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22812 Cleanup to use less temporaries.
22813 (vect_create_data_ref_ptr): Cleanup.
22814
38c56a5b
JJ		 228152013-03-18 Jakub Jelinek <jakub@redhat.com>
22816
22817 PR tree-optimization/56635
22818 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22819 require types_compatible_p types.
22820
20b2e6a0
NC		 228212013-03-18 Nick Clifton <nickc@redhat.com>
22822
a6178a25
NC		 22823 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22824 spurious backslash.
22825
20b2e6a0
NC		 22826 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22827 Add missing line to comment describing function.
22828
92608d0e
RB		 228292013-03-18 Richard Biener <rguenther@suse.de>
22830
22831 PR tree-optimization/56210
22832 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22833 Handle string / character search functions.
22834 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22835
31348d52
RB		 228362013-03-18 Richard Biener <rguenther@suse.de>
22837
22838 PR middle-end/56483
22839 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22840 and implement properly.
22841 * gimple.h (gimple_cond_single_var_p): Remove.
22842
fcac74a1
RB		 228432013-03-18 Richard Biener <rguenther@suse.de>
22844
22845 * tree-data-ref.h (find_data_references_in_loop): Declare.
22846 * tree-data-ref.c (get_references_in_stmt): Use a stack
22847 vector pre-allocated in the callers.
22848 (find_data_references_in_stmt): Adjust.
22849 (graphite_find_data_references_in_stmt): Likewise.
22850 (create_rdg_vertices): Likewise.
22851 (find_data_references_in_loop): Export.
22852 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22853 Compute dependences here...
22854 (vect_analyze_data_refs): ...not here. When we encounter
22855 a non-vectorizable data reference in basic-block vectorization
22856 truncate the data reference vector. Do not bother to
22857 fixup data-dependence information for gather loads.
22858 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22859 of data references, as reported.
22860
0d5a1b56
RB		 228612013-03-18 Richard Biener <rguenther@suse.de>
22862
22863 PR tree-optimization/3713
22864 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22865 has_constants and expr.
22866 (stmt_has_constants): Properly valueize SSA names when deciding
22867 whether the stmt has constants.
22868
789c34e3
RB		 228692013-03-18 Richard Biener <rguenther@suse.de>
22870
22871 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22872 whole function when there is nothing to do.
22873 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22874 * tree-vectorizer.c (vectorize_loops): Update virtual and
22875 loop-closed SSA once.
22876 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22877
076b4605
RB		 228782013-03-18 Richard Biener <rguenther@suse.de>
22879
22880 PR middle-end/56113
22881 * domwalk.c (bb_postorder): New global static.
22882 (cmp_bb_postorder): New function.
22883 (walk_dominator_tree): Replace scheme imposing an order for
22884 visiting dominator sons by one sorting them at the time they
22885 are pushed on the stack.
22886
bdb01696
RB		 228872013-03-18 Richard Biener <rguenther@suse.de>
22888
22889 PR tree-optimization/39326
22890 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22891 (struct mem_ref): Replace mem member with ao_ref typed member.
22892 (MEM_ANALYZABLE): Adjust.
22893 (memref_eq): Likewise.
22894 (mem_ref_alloc): Likewise.
22895 (gather_mem_refs_stmt): Likewise.
22896 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22897 (execute_sm_if_changed_flag_set): Adjust.
22898 (execute_sm): Likewise.
22899 (ref_always_accessed_p): Likewise.
22900 (refs_independent_p): Likewise.
22901 (can_sm_ref_p): Likewise.
22902
12d2dc5e
JJ		 229032013-03-18 Jakub Jelinek <jakub@redhat.com>
22904
22905 PR c/56566
22906 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22907 return 1 even for !unsignedp.
22908
286fb677
UB		 229092013-03-17 Uros Bizjak <ubizjak@gmail.com>
22910
22911 * config/i386/i386.md (isa): Add x64 and nox64.
22912 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22913 (*pushtf): Enable *roF alternative for x64 isa only.
22914 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22915 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22916 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22917 (*movtf_internal): Merge from *movtf_internal_rex64 and
22918 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22919 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22920 nox64 isa attributes.
22921 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22922 nox64 isa attributes.
22923 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22924
88b97037
UB		 229252013-03-17 Uros Bizjak <ubizjak@gmail.com>
22926
22927 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22928
9f4f1735
JJ		 229292013-03-17 Jakub Jelinek <jakub@redhat.com>
22930
22931 PR target/56640
22932 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22933 class names. Remove trailing comma after "ALL_REGS".
22934
48f4a6fa
JH		 229352013-03-16 Jan Hubicka <jh@suse.cz>
22936
22937 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22938 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22939 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22940 of cgraph_get_create_node.
22941 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22942
98e81123
JM		 229432013-03-16 Jason Merrill <jason@redhat.com>
22944
22945 PR debug/49090
22946 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22947 with DW_AT_default_value.
22948
8a298c04
JJ		 229492013-03-16 Jakub Jelinek <jakub@redhat.com>
22950
22951 * BASE-VER: Set to 4.9.0.
22952
4323afa0
AK		 229532013-03-14 Andi Kleen <ak@linux.intel.com>
22954
22955 PR target/56619
22956 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22957 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22958 Document _x* TSX intrinsics.
22959
b3c0d469
JJ		 229602013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22961 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 22962
22963 * configure.ac: Add MicroBlaze TLS support detection.
22964 * configure: Regenerate.
88b97037
UB		 22965 * config/microblaze/microblaze-protos.h
22966 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22967 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22968 * config/microblaze/microblaze.c (microblaze_address_type): Add
22969 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 22970 (microblaze_address_info): Add tls_reloc.
22971 (TARGET_HAVE_TLS): Define.
22972 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22973 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 22974 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22975 load_tls_operand, microblaze_call_tls_get_addr,
22976 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 22977 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22978 (get_base_reg): Use microblaze_tls_symbol_p.
22979 (microblaze_classify_address): Handle TLS.
88b97037
UB		 22980 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22981 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 22982 (microblaze_legitimize_address): Handle TLS.
22983 (microblaze_address_insns): Handle ADDRESS_TLS.
22984 (pic_address_needs_scratch): Handle TLS.
22985 (print_operand_address): Handle TLS.
22986 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22987 (microblaze_expand_move): Handle TLS.
88b97037
UB		 22988 (microblaze_legitimate_constant_p): Check
22989 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 22990 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22991 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22992 (PIC_OFFSET_TABLE_REGNUM): Set.
22993 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22994 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22995 (addsi3, movsi_internal2, movdf_internal): Update constraints
22996 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 22997 (move_operand): Redefine as move_src_operand,
22998 check microblaze_tls_referenced_p.
8cc9a5a5 22999
d803a491
IB		 230002013-03-14 Ian Bolton <ian.bolton@arm.com>
23001
23002 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23003 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23004
a8504f22
IB		 230052013-03-14 Ian Bolton <ian.bolton@arm.com>
23006
23007 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23008 CC mode for AND.
23009
df2dfaea
JJ		 230102013-03-14 Jakub Jelinek <jakub@redhat.com>
23011
fbd28bc3
JJ		 23012 PR tree-optimization/53265
23013 * common.opt (Waggressive-loop-optimizations): New option.
23014 * tree-ssa-loop-niter.c: Include tree-pass.h.
23015 (do_warn_aggressive_loop_optimizations): New function.
23016 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23017 if number_of_latch_executions returned constant.
23018 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23019 early. If number_of_latch_executions returned constant, set
23020 nb_iterations_upper_bound back to it.
23021 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23022 field.
23023 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23024 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23025
df2dfaea
JJ		 23026 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23027 (MULTILIB_OSDIRNAMES): Set.
23028 * genmultilib: If defaultosdirname doesn't start with :: , set
23029 defaultosdirname2 instead, clear it and emit two . multilib_raw
23030 entries instead of just one.
23031
ee0d2b68
KK		 230322013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23033
23034 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23035 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23036 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23037 (SUBTARGET_OVERRIDE_OPTIONS): New.
23038
decc676e
OE		 230392013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23040
23041 PR target/49880
23042 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23043 (musermode): Convert to Var(TARGET_USERMODE).
23044 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23045 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23046 * config/sh/sh.c (sh_option_override): Use
23047 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23048 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23049 condition.
23050 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23051 TARGET_SH4.
23052 (udivsi3_i4_single, divsi3_i4_single): Use
23053 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23054
f4b719c7
DK		 230552013-03-13 Dave Korn <dave.korn.cygwin@....>
23056
23057 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23058 default setting.
23059
c40eced0
RB		 230602013-03-13 Richard Biener <rguenther@suse.de>
23061
23062 PR tree-optimization/56608
23063 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23064 calls when vectorizing basic-blocks.
23065
1bfa5973
JJ		 230662013-03-13 Jakub Jelinek <jakub@redhat.com>
23067
23068 PR plugins/45078
23069 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23070 tm_file.
23071
6fcf5434
JJ		 230722013-03-12 Jakub Jelinek <jakub@redhat.com>
23073
23074 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23075
4f38fa8c
JH		 230762013-03-11 Jan Hubicka <jh@suse.cz>
23077
23078 PR lto/56557
23079 * lto-streamer-out.c (output_symbol_p): Skip references from
23080 constructors of external variables.
23081
c5c5ba89
JH		 230822013-03-11 Jan Hubicka <jh@suse.cz>
23083
23084 PR middle-end/56571
23085 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23086 from pseudos.
23087 * emit-rtl.c (verify_rtx_sharing): Likewise.
23088 (copy_insn_1): Likewise.
23089 * rtl.c (copy_rtx): Likewise.
23090
c2a939b1
GJL		 230912013-03-11 Georg-Johann Lay <avr@gjlay.de>
23092
23093 PR target/56591
23094 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23095 output_operand_lossage message.
88b97037 23096
3c3279fb
RE		 230972013-03-11 Richard Earnshaw <rearnsha@arm.com>
23098
23099 PR target/56470
23100 * arm.c (shift_op): Validate RTL pattern on the fly.
23101 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23102 the RTL.
23103
aef5ef9d
JDA		 231042013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23105
23106 PR target/56347
23107 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23108 new call patterns that clobber %fr12.
23109 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23110 split and postreload patterns.
23111 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23112 registers %fr12 and %fr12R as call used.
23113
456610d3
SB		 231142013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23115
23116 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23117 (canon_address, record_store, replace_read, check_mem_read_rtx,
23118 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23119 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23120 rest_of_handle_dse): Likewise.
23121
4b1baac8
RS		 231222013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23123
23124 PR middle-end/56524
23125 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23126 Add base_optabs.
23127 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23128 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23129 (save_optabs_if_changed): Replace with...
23130 (init_tree_optimization_optabs): ...this.
23131 * optabs.c (save_optabs_if_changed): Rename to...
23132 (init_tree_optimization_optabs): ...this. Take the optimization node
23133 as argument. Do nothing if the base optabs are already correct.
23134 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23135 to recompute optabs.
23136 * function.h (function): Remove optabs field.
23137 * function.c (invoke_set_current_function_hook): Call
23138 init_tree_optimization_optabs. Use the result to initialize
23139 this_fn_optabs.
23140
b7a78683
AH		 231412013-02-27 Aldy Hernandez <aldyh@redhat.com>
23142
23143 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23144 if GTMA_HAS_NO_INSTRUMENTATION.
23145 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23146 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23147 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23148 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23149 GTMA_HAS_NO_INSTRUMENTATION.
23150
6384c29b
JJ		 231512013-03-08 Jakub Jelinek <jakub@redhat.com>
23152
23153 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23154 libasan_preinit.o.
23155
ecd4f20a
MP		 231562013-03-08 Marek Polacek <polacek@redhat.com>
23157 Jakub Jelinek <jakub@redhat.com>
23158
23159 PR tree-optimization/56478
23160 * predict.c (is_comparison_with_loop_invariant_p): Change the
23161 type of loop_step to tree.
23162 (predict_loops): Adjust.
23163 (predict_iv_comparison): Perform the computations on double_ints.
23164
64366d35
RB		 231652013-03-08 Richard Biener <rguenther@suse.de>
23166
23167 PR tree-optimization/56570
23168 * tree-cfg.c (verify_expr_location_1): Verify locations for
23169 DECL_DEBUG_EXPR.
23170 * tree-sra.c (create_access_replacement): Strip locations
23171 from DECL_DEBUG_EXPRs.
23172
a9d5a059
RB		 231732013-03-08 Richard Biener <rguenther@suse.de>
23174
23175 * tree-inline.c (expand_call_inline): Do not associate
23176 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23177 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23178
b7aa4e9a
RB		 231792013-03-08 Richard Biener <rguenther@suse.de>
23180
23181 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23182 or block changes with -Og. Fix for location / block encoding
23183 changes and PHI arguments with locations.
23184
c4c2f9fa
SB		 231852013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23186
23187 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23188 for all counters.
23189 (struct output_info): Likewise.
23190 (register_overhead): Remove bad gcc_assert.
23191 (bitmap_find_bit): If there is only a single bitmap element, do not
23192 count a miss as a search.
23193 (print_statistics): Update for counter type changes.
23194 (dump_bitmap_statistics): Likewise. Print headers such that they
23195 are properly lined up with the printed counters.
23196
5bf6606a
JJ		 231972013-03-07 Jakub Jelinek <jakub@redhat.com>
23198
23199 PR tree-optimization/56559
23200 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23201 check that it has only a single use.
23202
2c653d46
RB		 232032013-03-07 Richard Biener <rguenther@suse.de>
23204
23205 * doc/invoke.texi (fwhole-program): Discourage use in combination
23206 with -flto.
23207
a72d8780
JJ		 232082013-03-06 Jakub Jelinek <jakub@redhat.com>
23209
01a454df
JJ		 23210 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23211
c1781047
JJ		 23212 PR tree-optimization/56539
23213 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23214 instead of GSI_CONTINUE_LINKING as last argument to
23215 force_gimple_operand_gsi. Adjust function comment.
23216
9772c47a
JJ		 23217 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23218 aarch64-cores.def.
23219
a72d8780
JJ		 23220 PR middle-end/56548
23221 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23222 promoted mode, convert the result back to the original mode.
23223
fa5556de
RB		 232242013-03-06 Richard Biener <rguenther@suse.de>
23225
23226 PR middle-end/56294
23227 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23228 (insert_updated_phi_nodes_compare_uids): New function.
23229 (update_ssa): Sort symbols_to_rename after UID before
23230 traversing it to insert PHI nodes.
23231
010403d1
RB		 232322013-03-06 Richard Biener <rguenther@suse.de>
23233
23234 PR middle-end/50494
23235 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23236 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23237
23238 Revert
23239 2013-02-13 Richard Biener <rguenther@suse.de>
23240
23241 PR lto/50494
23242 * varasm.c (output_constant_def_1): Get the decl representing
23243 the constant as argument.
23244 (output_constant_def): Wrap output_constant_def_1.
23245 (make_decl_rtl): Use output_constant_def_1 with the decl
23246 representing the constant.
23247 (build_constant_desc): Optionally re-use a decl already
23248 representing the constant.
23249 (tree_output_constant_def): Adjust.
23250
3c27ce4c
JY		 232512013-03-06 Joey Ye <joey.ye@arm.com>
23252
23253 PR lto/50293
23254 * gcc.c (convert_white_space): New function.
23255 (main): Handles white space in function name.
23256
8f6d1c86
OE		 232572013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23258
23259 PR target/56529
23260 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23261 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23262 to SH_DIV_CALL_TABLE for TARGET_SH2.
23263 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23264 list.
23265 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23266 call-table options.
23267
798a209f
SA		 232682013-03-05 Sterling Augustine <saugustine@google.com>
23269 Cary Coutant <ccoutant@google.com>
23270
23271 PR debug/55364
23272 * dwarf2out.c (resolve_addr): Don't call
23273 remove_loc_list_addr_table_entries a second time for the same
23274 expression.
23275
6cfa417f 232762013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 23277
6cfa417f
JJ		 23278 PR debug/56510
23279 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23280 (avoid_complex_debug_insns): New function.
23281 (expand_debug_locations): Call it.
23282
23283 PR rtl-optimization/56484
23284 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23285 lifetimes of hard registers on small register class machines.
23286
232872013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23288
23289 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 23290 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 23291 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 23292 fast_interrupt.
23293 (microblaze_fast_interrupt_function_p): New function.
23294 (microblaze_is_interrupt_handler): Rename to
23295 microblaze_is_interrupt_variant and add fast_interrupt check.
23296 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23297 (save_restore_insns): Likewise.
23298 (compute_frame_size): Likewise.
23299 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23300 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 23301 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23302 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 23303 microblaze_is_interrupt_variant.
23304
731aefac
KT		 233052013-03-05 Kai Tietz <ktietz@redhat.com>
23306
23307 * sdbout.c (sdbout_one_type): Switch to current function's section
23308 supporting cold/hot.
23309
a72d8780 233102013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 23311
23312 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23313 -mxl-reorder.
23314
334e71e8
JJ		 233152013-03-05 Jakub Jelinek <jakub@redhat.com>
23316
0b50e654
JJ		 23317 PR middle-end/56461
23318 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23319 if VALGRIND_GET_VBITS is defined, temporarily make object
23320 memory all defined, and restore previous valgrind addressability
23321 and definability afterwards. Free this_object at the end.
23322
4ccf8f43
JJ		 23323 PR middle-end/56461
23324 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23325 right before calling lra_create_live_ranges, also call it
23326 when clearing live_p. Only call lra_clear_live_ranges
23327 at the end if live_p.
23328
334e71e8
JJ		 23329 PR middle-end/56461
23330 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23331
9ca966ca
RB		 233322013-03-05 Richard Biener <rguenther@suse.de>
23333
23334 PR tree-optimization/56521
23335 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23336 value-id.
23337
d273b176
SB		 233382013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23339
23340 PR c++/55135
23341 * except.h (remove_unreachable_eh_regions): New prototype.
23342 * except.c (remove_eh_handler_splicer): New function, split out
23343 of remove_eh_handler.
23344 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23345 warning about running it on many EH regions one at a time.
23346 (remove_unreachable_eh_regions_worker): New function, walk the
23347 EH tree in depth-first order and remove non-marked regions.
23348 (remove_unreachable_eh_regions): New function.
23349 * tree-eh.c (mark_reachable_handlers): New function, split out
23350 from remove_unreachable_handlers.
23351 (remove_unreachable_handlers): Use mark_reachable_handlers and
23352 remove_unreachable_eh_regions.
23353 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23354 and remove_unreachable_eh_regions.
23355
5e7f1aef
RB		 233562013-03-05 Richard Biener <rguenther@suse.de>
23357
23358 PR middle-end/56525
23359 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23360 not freeing them until the end.
23361
f276b762
AK		 233622013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23363
23364 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23365
12a43ab8
RB		 233662013-03-05 Richard Biener <rguenther@suse.de>
23367
23368 PR tree-optimization/56270
23369 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23370 of loads after scheduling an SLP instance.
23371
93675444
JJ		 233722013-03-05 Jakub Jelinek <jakub@redhat.com>
23373
db4138e3
JJ		 23374 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23375 tic6x.exp.
23376 (check_gcc_parallelize): Run guality.exp as a separate job from
23377 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23378 struct-layout-1.exp with stackalign.exp.
23379
dd3d1ec0
JJ		 23380 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23381
b8d381a3
JJ		 23382 PR middle-end/56461
23383 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23384 load_index sbitmap even if some bit in it isn't set.
23385
b4f9786b
JJ		 23386 PR middle-end/56461
23387 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23388 (discover_iteration_bound_by_body_walk): Change queues to
23389 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23390 spelling in comment. Call safe_push on queues[bound_index] directly.
23391 Release queues[queue_index] in every iteration unconditionally.
23392 Release bounds vector.
23393
93675444
JJ		 23394 PR middle-end/56461
23395 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23396 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23397 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23398 inner_phis vector.
23399
3e492e9c
RB		 234002013-03-05 Richard Biener <rguenther@suse.de>
23401
23402 PR lto/56515
23403 * tree-inline.c (remap_blocks_to_null): New function.
23404 (expand_call_inline): When expanding a call stmt without
23405 an associated block inline remap all callee blocks to NULL.
23406
a5d1569a
JJ		 234072013-03-05 Jakub Jelinek <jakub@redhat.com>
23408
86efb5cd
JJ		 23409 PR rtl-optimization/56494
23410 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23411 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23412 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23413
85f5dbea
JJ		 23414 PR middle-end/56461
23415 * sel-sched-ir.c (free_sched_pools): Release
23416 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23417 if succs_info_pool.max_top isn't -1.
23418
a5d1569a
JJ		 23419 PR bootstrap/56509
23420 * opts.c (opts_obstack, opts_concat): Moved to...
23421 * opts-common.c (opts_obstack, opts_concat): ... here.
23422
4432aa6c
JJ		 234232013-03-04 Jakub Jelinek <jakub@redhat.com>
23424
23425 PR middle-end/56461
23426 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23427
f3a1fb91
MJ		 234282013-03-04 Martin Jambor <mjambor@suse.cz>
23429
23430 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23431 all appropriate places.
23432
22110e6c
EB		 234332013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23434
23435 PR tree-optimization/56424
23436 * ipa-split.c (split_function): Do not set the RSO flag if result is
23437 not by reference and its type is a register type.
23438
a72d8780 234392013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23440
88b97037 23441 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23442 (microblaze_legitimate_pic_operand): Likewise
88b97037 23443 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 23444 new function microblaze_legitimate_pic_operand
23445 * config/microblaze/microblaze-protos.h
23446 (microblaze_legitimate_pic_operand): Declare.
23447
a72d8780 234482013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23449
a72d8780 23450 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23451 New predicate for supported rtx code types.
a72d8780 23452 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 23453 call_insn_simple_operand predicate.
23454
541d9ac8
JJ		 234552013-03-04 Jakub Jelinek <jakub@redhat.com>
23456
5eb010bc
JJ		 23457 PR middle-end/56461
23458 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23459 partitions.ordered_remove.
23460
30862efc
JJ		 23461 PR middle-end/56461
23462 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23463 vec_oprnds0.create (1) for modifier == NONE.
23464
8930f723
JJ		 23465 PR middle-end/56461
23466 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23467 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23468 vec_oprnds1 right before pushing anything to it for
23469 scalar_shift_arg.
23470
541d9ac8
JJ		 23471 PR middle-end/56461
23472 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23473 set nbbs to 0 instead of having separate code path.
23474 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23475 instead of false as last argument if returning NULL.
23476
7aa7f2e3
SL		 234772013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23478
23479 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23480 the attribute is now called "target" instead of "option".
23481 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23482 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23483 attribute/pragma name for TARGET_OPTION_VALID_P and
23484 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23485 * doc/tm.texi: Regenerated.
23486
8930f723 234872013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23488
8930f723 23489 * config/microblaze/microblaze.c:
cb8a1637 23490 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 23491 * config/microblaze/microblaze.h: Add -mxl-reorder to
23492 DRIVER_SELF_SPECS.
23493 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23494 instructions emitted if TARGET_REORDER.
88b97037
UB		 23495 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23496 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23497 separately.
cb8a1637 23498
4dc1d68c
XDL		 234992013-03-01 Xinliang David Li <davidxl@google.com>
23500
23501 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23502 walk length.
23503
689eaba3
JJ		 235042013-03-01 Jakub Jelinek <jakub@redhat.com>
23505
9d676bf6
JJ		 23506 PR middle-end/56461
23507 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23508 vector even when returning true. Fix up function comment formatting.
23509
caff7edf
JJ		 23510 PR middle-end/56461
23511 * ira-build.c (ira_loop_nodes_count): New variable.
23512 (create_loop_tree_nodes): Initialize it.
23513 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23514
b6b9227d
JJ		 23515 PR middle-end/56461
23516 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23517 method on dr_chain and result_chain.
23518 * tree-vect-stmts.c (vectorizable_store): Only call
23519 result_chain.create if j == 0.
23520
689eaba3
JJ		 23521 PR middle-end/56461
23522 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23523 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23524 before overwriting it.
23525
40bf31ed
TB		 235262013-03-01 Tobias Burnus <burnus@net-b.de>
23527
23528 * doc/extended.texi (C Extensions): Change order in @menu
23529 to match @node.
23530 (Other MIPS Built-in Functions): Move last MIPS entry before
23531 "picoChip Built-in Functions".
23532 (SH Built-in Functions): Move after RX Built-in Functions.
23533 * doc/gcc.texi (Introduction): Change order in @menu
23534 to match @node.
23535 * doc/md.texi (Constraints): Ditto.
23536 * gty.texi (Type Information): Ditto.
23537 (User-provided marking routines for template types): Make
23538 subsection.
23539 * doc/invoke.texi (AArch64 Options): Move before
23540 "Adapteva Epiphany Options".
23541
e664c61c
KS		 235422013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23543 Jakub Jelinek <jakub@redhat.com>
23544
23545 PR sanitizer/56454
23546 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23547 no_address_safety_analysis attribute.
23548 * doc/extend.texi (no_address_safety_attribute): Rename to
23549 no_sanitize_address attribute, mention no_address_safety_analysis
23550 attribute as deprecated alias.
23551
37b5ec8f
JJ		 235522013-02-28 Jakub Jelinek <jakub@redhat.com>
23553
23554 PR middle-end/56461
23555 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23556 type to vec<vec<tree> > *.
23557 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23558 to be vec<tree> instead of vec<tree> *, set vec_defs
23559 to vNULL and call vec_defs.create (number_of_vects), adjust other
23560 uses of vec_defs.
23561 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23562 vectorizable_condition): Adjust vect_get_slp_defs callers.
23563
ba96cdfb
JG		 235642013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23565
23566 * config/aarch64/aarch64.c
23567 (aarch64_float_const_representable): Remove unused variable.
23568
6f549691
JG		 235692013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23570
23571 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23572
af55e82d
JG		 235732013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23574
23575 * config/aarch64/aarch64-builtins.c
23576 (aarch64_init_simd_builtins): Make static.
23577
1df3f464
JG		 235782013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23579
23580 * config/aarch64/aarch64.c
23581 (aarch64_simd_make_constant): Make static.
23582
f8f42513
MJ		 235832013-02-28 Martin Jambor <mjambor@suse.cz>
23584
23585 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23586 with no initialization to the RHS of debug statements.
23587
b48b3fc4
MJ		 235882013-02-28 Martin Jambor <mjambor@suse.cz>
23589
23590 PR tree-optimization/56294
23591 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23592 Adjust dumping.
23593 (get_access_replacement): Do not call create_access_replacement.
23594 Assert a replacement exists.
23595 (get_repl_default_def_ssa_name): Create the replacement declaration
23596 itself.
23597
c3ae224c
RR		 235982013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23599
23600 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23601 final_end_function.
23602
45fa0eef
MP		 236032013-02-28 Marek Polacek <polacek@redhat.com>
23604
23605 PR rtl-optimization/56466
23606 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23607 if we're changing a loop.
23608 (peel_loops_completely): Likewise.
23609
502c067d
PC		 236102013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23611
23612 PR c++/55813
23613 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23614
ff24afc8
GJL		 236152013-02-28 Georg-Johann Lay <avr@gjlay.de>
23616
23617 PR target/56445
23618 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23619 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23620 INTX_FTYPE_FX, FX_FTYPE_INTX.
23621 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23622
c7b1fc1b
GJL		 236232013-02-28 Georg-Johann Lay <avr@gjlay.de>
23624
23625 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23626 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23627 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23628 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23629 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23630 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23631 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23632 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23633 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23634 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23635 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23636 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23637 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23638 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23639 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23640 (avrxmega6): Increase max flash segments from 5 to 6.
23641 * config/avr/t-multilib: Regenerate.
23642 * config/avr/avr-tables.opt: Regenerate.
23643 * doc/avr-mmcu.texi: Regenerate.
23644
4a0e3cfe
GJL		 236452013-02-28 Georg-Johann Lay <avr@gjlay.de>
23646
23647 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23648 (avr_device_to_arch): Rename to avr_device_to_ld.
23649 (avr_device_to_as): New prototype.
23650 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23651 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23652 * config/avr/driver-avr.c (avr_device_to_as): New.
23653 (avr_device_to_arch): Rename to avr_device_to_ld.
23654
97785e52
JJ		 236552013-02-27 Jakub Jelinek <jakub@redhat.com>
23656
3f292312
JJ		 23657 PR middle-end/56461
23658 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23659 method on dr_chain and result_chain.
23660
a344216b
JJ		 23661 PR middle-end/56461
23662 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23663 pointer_set_destroy on not_executed_last_iteration.
23664
f121ad02 23665 PR middle-end/56461
88b97037 23666 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 23667
307f83a3
JJ		 23668 PR middle-end/56461
23669 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23670 FOR_EACH_DEFINED_FUNCTION when freeing state.
23671
e19624ee
JJ		 23672 PR middle-end/56461
23673 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23674 pool_free.
23675 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23676 overwriting it.
23677
90e709fd
JJ		 23678 PR middle-end/56461
23679 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23680 known_aggs[i].items and release known_aggs vector.
23681
97785e52
JJ		 23682 PR middle-end/56461
23683 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23684
236852013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 23686
97785e52
JJ		 23687 * config/microblaze/microblaze.c (microblaze_emit_compare):
23688 Use xor for EQ/NE comparisions.
23689 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23690 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 23691
6fa5e0ed
JJ		 236922013-02-27 Jakub Jelinek <jakub@redhat.com>
23693
23694 PR middle-end/56461
23695 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23696 vector.
23697 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23698 vec_safe_push, always update *slot.
23699 (redirect_edge_var_map_clear): Use vec_free.
23700 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23701 (free_var_map_entry): Use vec_free.
23702 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23703 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23704
436a956a
AB		 237052013-02-27 Andrey Belevantsev <abel@ispras.ru>
23706
23707 PR middle-end/45472
436a956a
AB		 23708 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23709 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 23710 Reorder tests for speculativeness in the logical and operator.
23711
f5c2caca
JJ		 237122013-02-27 Jakub Jelinek <jakub@redhat.com>
23713
0fcb564b
JJ		 23714 * incpath.c (add_standard_paths): Use reconcat instead of concat
23715 where appropriate and avoid leaking memory.
23716
dc357798
JJ		 23717 * opts.h: Include obstack.h.
23718 (opts_concat): New prototype.
23719 (opts_obstack): New declaration.
23720 * opts.c (opts_concat): New function.
23721 (opts_obstack): New variable.
23722 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23723 (finish_options): Use opts_concat instead of concat
23724 and XOBNEWVEC instead of XNEWVEC.
23725 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23726 generate_option): Likewise.
23727 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23728 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23729
f5c2caca
JJ		 23730 PR target/56455
23731 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23732 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23733
d0163673
JJ		 237342013-02-26 Jakub Jelinek <jakub@redhat.com>
23735
23736 PR middle-end/56461
23737 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23738
b5ad2b8e
JR		 237392013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23740
23741 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23742 (arm_block_move_unaligned_straight): Likewise.
23743 (arm_adjust_block_mem): Likewise.
23744
9b639e2c
JR		 237452013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23746
23747 PR target/48901
23748 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23749 temp, cond and label.
23750 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23751
e13a0ccb
JR		 23752 PR target/52500
23753 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23754 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23755
fbe4f171
JR		 23756 PR target/52501
23757 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23758 of prologue/epilogue functions.
23759
ae006543
JR		 23760 PR target/52550
23761 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23762 Remove unused variable cfa_offset.
23763 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23764
c81369fa
JR		 23765 PR target/54639
23766 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23767 type promotion to unsigned.
23768
f8a8fea7
JR		 23769 PR target/54640
23770 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23771 for HOST_WIDE_INT of 32 bit / same size as int.
23772 (arm_block_move_unaligned_straight): Likewise.
23773 (arm_adjust_block_mem): Likewise.
23774
f8be5169
JR		 23775 PR target/54662
23776 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23777 ALL_CFLAGS.
23778
f1ad3354
MP		 237792013-02-26 Marek Polacek <polacek@redhat.com>
23780
23781 PR tree-optimization/56426
88b97037 23782 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 23783
a6af0f42
RB		 237842013-02-26 Richard Biener <rguenther@suse.de>
23785
23786 PR target/56444
23787 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23788 unused variable loops.
23789
aca43c6c
JJ		 237902013-02-26 Jakub Jelinek <jakub@redhat.com>
23791
cecbe5d9
JJ		 23792 PR tree-optimization/56448
23793 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23794 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23795 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23796 later operands of the references, or even first operand for
23797 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23798
aca43c6c
JJ		 23799 PR tree-optimization/56443
23800 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23801 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23802 to type_for_mode langhook.
23803
53e2e141
MT		 238042013-02-25 Matt Turner <mattst88@gmail.com>
23805
23806 * doc/invoke.texi: Document r4700.
23807
259ee451
RB		 238082013-02-25 Richard Biener <rguenther@suse.de>
23809
23810 PR tree-optimization/56175
23811 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23812 split out from ...
23813 (simplify_bitwise_binary): ... here. Also guard the conversion
23814 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23815
6d65e8f1
CM		 238162013-02-25 Catherine Moore <clm@codesourcery.com>
23817
23818 Revert:
23819 2013-02-24 Catherine Moore <clm@codesourcery.com>
23820 Maciej W. Rozycki <macro@codesourcery.com>
23821 Tom de Vries <tom@codesourcery.com>
a72d8780 23822 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 23823 Iain Sandoe <iain@codesourcery.com>
23824 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23825 Chao-ying Fu <fu@mips.com>
88b97037 23826
6d65e8f1 23827 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23828 Document new function attributes.
6d65e8f1
CM		 23829 * doc/invoke.texi (minterlink-compressed, mmicromips,
23830 m14k, m14ke, m14kec): Document new options.
23831 (minterlink-mips16): Update documentation.
23832 * doc/md.texi (ZC, ZD): Document new constraints.
23833 * configure.ac (gcc_cv_as_micromips): Check if linker
23834 supports the .set micromips directive.
23835 * configure: Regenerate.
23836 * config.in: Regenerate.
23837 * config/mips/mips-tables.opt: Regenerate.
23838 * config/mips/micromips.md: New file.
23839 * constraints.md (ZC, AD): New constraints.
23840 * config/mips/predicates.md (movep_src_register): New predicate.
23841 (movep_src_operand): New predicate.
23842 (non_volatile_mem_operand): New predicate.
23843 * config/mips/mips.md (multimem): New type.
23844 (length): Differentiate between 17-bit and 18-bit branch offsets.
23845 (MOVEP1, MOVEP2): New mode iterator.
23846 (mov_<load>l): Use ZC constraint.
23847 (mov_<load>r): Likewise.
23848 (mov_<store>l): Likewise.
23849 (mov_<store>r): Likewise.
23850 (*branch_equality<mode>_inverted): Add microMIPS support.
23851 (*branch_equality<mode>): Likewise.
23852 (*jump_absolute): Likewise.
23853 (indirect_jump_<mode>): Likewise.
23854 (tablejump_<mode>): Likewise.
23855 (<optab>_internal): Likewise.
23856 (sibcall_internal): Likewise.
23857 (sibcall_value_internal): Likewise.
23858 (prefetch): Use constraint ZD.
23859 * config/mips/mips.opt (minterlink-compressed): New option.
23860 (minterlink-mips16): Now an alias for minterlink-compressed.
23861 (mmicromips): New option.
23862 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23863 (compare_and_swap_12): Likewise.
23864 (sync_add<mode>): Likewise.
23865 (sync_<optab>_12): Likewise.
23866 (sync_old_<optab>_12): Likewise.
23867 (sync_new_<optab>_12): Likewise.
23868 (sync_nand_12): Likewise.
23869 (sync_old_nand_12): Likewise.
23870 (sync_new_nand_12): Likewise.
23871 (sync_sub<mode>): Likewise.
23872 (sync_old_add<mode>): Likewise.
23873 (sync_old_sub<mode>): Likewise.
23874 (sync_new_add<mode>): Likewise.
23875 (sync_new_sub<mode>): Likewise.
23876 (sync_<optab><mode>): Likewise.
23877 (sync_old_<optab><mode>): Likewise.
23878 (sync_new_<optab><mode>): Likewise.
23879 (sync_nand<mode>): Likewise.
23880 (sync_old_nand<mode>): Likewise.
23881 (sync_new_nand<mode>): Likewise.
23882 (sync_lock_test_and_set<mode>): Likewise.
23883 (test_and_set_12): Likewise.
23884 (atomic_compare_and_swap<mode>): Likewise.
23885 (atomic_exchange<mode>_llsc): Likewise.
23886 (atomic_fetch_add<mode>_llsc): Likewise.
23887 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23888 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23889 (umips_save_restore_pattern_p): Likewise.
23890 (umips_load_store_pair_p): Likewise.
23891 (umips_output_load_store_pair): Likewise.
23892 (umips_movep_target_p): Likewise.
23893 (umips_12bit_offset_address_p): Likewise.
23894 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23895 (mips_base_mips16): Rename this...
23896 (mips_base_compression_flags): ...to this. Update all uses.
23897 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23898 (mips_mips16_decl_p): Delete.
23899 (mips_nomips16_decl_p): Delete.
23900 (mips_get_compress_on_flags): New function.
23901 (mips_get_compress_off_flags): New function.
23902 (mips_get_compress_mode): New function.
23903 (mips_get_compress_on_name): New function.
23904 (mips_get_compress_off_name): New function.
23905 (mips_insert_attributes): Support multiple compression types.
23906 (mips_merge_decl_attributes): Likewise.
23907 (umips_12bit_offset_address_p): New function.
23908 (mips_start_function_definition): Emit .set micromips directive.
23909 (mips_call_may_need_jalx_p): New function.
23910 (mips_function_ok_for_sibcall): Add microMIPS support.
23911 (mips_print_operand_punctuation): Support short delay slots and
23912 compact jumps.
23913 (umips_swm_mask, umips_swm_encoding): New.
23914 (umips_build_save_restore): New function.
23915 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23916 (was_mips16_p): Remove.
23917 (old_compression_mode): New.
23918 (mips_set_compression_mode): New function.
23919 (mips_set_current_function): Add microMIPS support.
23920 (mips_option_override): Likewise.
23921 (umips_save_restore_pattern_p): New function.
23922 (umips_output_save_restore): New function.
23923 (umips_load_store_pair_p_1): New function.
23924 (umips_load_store_pair_p): New function.
23925 (umips_output_load_store_pair_1): New function.
23926 (umips_output_load_store_pair): New function.
23927 (umips_movep_target_p) New function.
23928 (mips_prepare_pch_save): Add microMIPS support.
23929 * config/mips/mips.h (TARGET_COMPRESSION): New.
23930 (TARGET_CPU_CPP_BUILTINS): Update macro
23931 to use new compression flags and to support microMIPS.
23932 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23933 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23934 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23935 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23936 (ASM_SPEC): Support mmicromips and mno-micromips.
23937 (M16STORE_REG_P): New macro.
23938 (MIPS_CALL): Support TARGET_MICROMIPS.
23939 (MICROMIPS_J): New macro.
23940 (mips_base_mips16): Rename this...
23941 (mips_base_compression_flags): ...to this.
23942 (UMIPS_12BIT_OFFSET_P): New macro.
23943 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23944 (MULTILIB_DIRNAMES): Likewise.
23945
5e5df392
TV		 239462013-02-25 Tom de Vries <tom@codesourcery.com>
23947
23948 PR rtl-optimization/56131
23949 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23950 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23951 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23952
3fdb53c1
TB		 239532013-02-25 Tobias Burnus <burnus@net-b.de>
23954
23955 * doc/invoke.texi (-fsanitize=): Move from optimization
23956 to debugging options.
23957
ed358aea
AB		 239582013-02-25 Andrey Belevantsev <abel@ispras.ru>
23959
23960 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23961
f45e9053
AB		 239622013-02-25 Andrey Belevantsev <abel@ispras.ru>
23963 Alexander Monakov <amonakov@ispras.ru>
23964
23965 PR middle-end/56077
23966 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23967 flush pending lists also on non-jumps. Adjust comment.
23968
6941b508
CM		 239692013-02-24 Catherine Moore <clm@codesourcery.com>
23970 Maciej W. Rozycki <macro@codesourcery.com>
23971 Tom de Vries <tom@codesourcery.com>
a72d8780 23972 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 23973 Iain Sandoe <iain@codesourcery.com>
23974 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23975 Chao-ying Fu <fu@mips.com>
6941b508
CM		 23976
23977 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23978 Document new function attributes.
6941b508
CM		 23979 * doc/invoke.texi (minterlink-compressed, mmicromips,
23980 m14k, m14ke, m14kec): Document new options.
23981 (minterlink-mips16): Update documentation.
23982 * doc/md.texi (ZC, ZD): Document new constraints.
23983 * configure.ac (gcc_cv_as_micromips): Check if linker
23984 supports the .set micromips directive.
23985 * configure: Regenerate.
23986 * config.in: Regenerate.
23987 * config/mips/mips-tables.opt: Regenerate.
23988 * config/mips/micromips.md: New file.
23989 * constraints.md (ZC, AD): New constraints.
23990 * config/mips/predicates.md (movep_src_register): New predicate.
23991 (movep_src_operand): New predicate.
23992 (non_volatile_mem_operand): New predicate.
23993 * config/mips/mips.md (multimem): New type.
23994 (length): Differentiate between 17-bit and 18-bit branch offsets.
23995 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 23996 (mov_<load>l): Use ZC constraint.
6941b508
CM		 23997 (mov_<load>r): Likewise.
23998 (mov_<store>l): Likewise.
23999 (mov_<store>r): Likewise.
24000 (*branch_equality<mode>_inverted): Add microMIPS support.
24001 (*branch_equality<mode>): Likewise.
24002 (*jump_absolute): Likewise.
24003 (indirect_jump_<mode>): Likewise.
24004 (tablejump_<mode>): Likewise.
24005 (<optab>_internal): Likewise.
24006 (sibcall_internal): Likewise.
24007 (sibcall_value_internal): Likewise.
24008 (prefetch): Use constraint ZD.
24009 * config/mips/mips.opt (minterlink-compressed): New option.
24010 (minterlink-mips16): Now an alias for minterlink-compressed.
24011 (mmicromips): New option.
24012 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24013 (compare_and_swap_12): Likewise.
24014 (sync_add<mode>): Likewise.
24015 (sync_<optab>_12): Likewise.
24016 (sync_old_<optab>_12): Likewise.
24017 (sync_new_<optab>_12): Likewise.
24018 (sync_nand_12): Likewise.
24019 (sync_old_nand_12): Likewise.
24020 (sync_new_nand_12): Likewise.
24021 (sync_sub<mode>): Likewise.
24022 (sync_old_add<mode>): Likewise.
24023 (sync_old_sub<mode>): Likewise.
24024 (sync_new_add<mode>): Likewise.
24025 (sync_new_sub<mode>): Likewise.
24026 (sync_<optab><mode>): Likewise.
24027 (sync_old_<optab><mode>): Likewise.
24028 (sync_new_<optab><mode>): Likewise.
24029 (sync_nand<mode>): Likewise.
24030 (sync_old_nand<mode>): Likewise.
24031 (sync_new_nand<mode>): Likewise.
24032 (sync_lock_test_and_set<mode>): Likewise.
24033 (test_and_set_12): Likewise.
24034 (atomic_compare_and_swap<mode>): Likewise.
24035 (atomic_exchange<mode>_llsc): Likewise.
24036 (atomic_fetch_add<mode>_llsc): Likewise.
24037 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24038 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24039 (umips_save_restore_pattern_p): Likewise.
24040 (umips_load_store_pair_p): Likewise.
24041 (umips_output_load_store_pair): Likewise.
24042 (umips_movep_target_p): Likewise.
24043 (umips_12bit_offset_address_p): Likewise.
24044 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24045 (mips_base_mips16): Rename this...
24046 (mips_base_compression_flags): ...to this. Update all uses.
24047 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24048 (mips_mips16_decl_p): Delete.
24049 (mips_nomips16_decl_p): Delete.
3fdb53c1 24050 (mips_get_compress_on_flags): New function.
6941b508
CM		 24051 (mips_get_compress_off_flags): New function.
24052 (mips_get_compress_mode): New function.
24053 (mips_get_compress_on_name): New function.
24054 (mips_get_compress_off_name): New function.
24055 (mips_insert_attributes): Support multiple compression types.
24056 (mips_merge_decl_attributes): Likewise.
24057 (umips_12bit_offset_address_p): New function.
24058 (mips_start_function_definition): Emit .set micromips directive.
24059 (mips_call_may_need_jalx_p): New function.
24060 (mips_function_ok_for_sibcall): Add microMIPS support.
24061 (mips_print_operand_punctuation): Support short delay slots and
24062 compact jumps.
24063 (umips_swm_mask, umips_swm_encoding): New.
24064 (umips_build_save_restore): New function.
24065 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24066 (was_mips16_p): Remove.
24067 (old_compression_mode): New.
24068 (mips_set_compression_mode): New function.
24069 (mips_set_current_function): Add microMIPS support.
24070 (mips_option_override): Likewise.
24071 (umips_save_restore_pattern_p): New function.
24072 (umips_output_save_restore): New function.
24073 (umips_load_store_pair_p_1): New function.
24074 (umips_load_store_pair_p): New function.
24075 (umips_output_load_store_pair_1): New function.
24076 (umips_output_load_store_pair): New function.
24077 (umips_movep_target_p) New function.
24078 (mips_prepare_pch_save): Add microMIPS support.
24079 * config/mips/mips.h (TARGET_COMPRESSION): New.
24080 (TARGET_CPU_CPP_BUILTINS): Update macro
24081 to use new compression flags and to support microMIPS.
24082 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24083 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24084 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24085 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24086 (ASM_SPEC): Support mmicromips and mno-micromips.
24087 (M16STORE_REG_P): New macro.
24088 (MIPS_CALL): Support TARGET_MICROMIPS.
24089 (MICROMIPS_J): New macro.
24090 (mips_base_mips16): Rename this...
24091 (mips_base_compression_flags): ...to this.
24092 (UMIPS_12BIT_OFFSET_P): New macro.
24093 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24094 (MULTILIB_DIRNAMES): Likewise.
24095
cdebbc6d
JJ		 240962013-02-24 Jakub Jelinek <jakub@redhat.com>
24097
24098 PR target/52555
24099 * target-globals.c (save_target_globals): For init_reg_sets and
24100 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24101
18c63565
JG		 241022013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24103
24104 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24105 * config/aarch64/t-aarch64
24106 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24107
2194f7a2
VM		 241082013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24109
24110 PR inline-asm/56148
24111 * lra-constraints.c (process_alt_operands): Reload operand
24112 conflicting with earlier clobber only if no more other conflicting
24113 operands.
24114
7d613735
JJ		 241152013-02-22 Jakub Jelinek <jakub@redhat.com>
24116
24117 PR sanitizer/56393
24118 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24119 if not linking a shared library.
24120
ac8d93a7
SL		 241212013-02-22 Seth LaForge <sethml@google.com>
24122
24123 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24124
e0237780
GY		 241252013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24126
3fdb53c1
TB		 24127 * config/arm/arm.md (split for extendsidi): Update condition.
24128 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24129 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24130 (qhs_zextenddi_cstr): Likewise.
e0237780 24131
d7fde18c
JJ		 241322013-02-21 Jakub Jelinek <jakub@redhat.com>
24133
be63b77d
JJ		 24134 PR middle-end/56420
24135 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24136 avoid signed wrapping.
24137 (expand_mult): Handle properly multiplication by
24138 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24139 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24140 in the compiler if coeff is HOST_WIDE_INT_MIN.
24141 (expand_divmod): Don't make ext_op1 static, change it's type to
24142 uhwi. Avoid undefined behavior in -INTVAL (op1).
24143
d7fde18c
JJ		 24144 PR rtl-optimization/50339
24145 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24146 field.
24147 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24148 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24149 into splitting_ashiftrt field.
24150 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24151 ASHIFTRT.
24152 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24153 choices.
24154
6aad4455
AH		 241552013-02-20 Aldy Hernandez <aldyh@redhat.com>
24156
24157 PR middle-end/56108
24158 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24159 are sure to go irrevocable.
24160
38fe784d
HPN		 241612013-02-21 Hans-Peter Nilsson <hp@axis.com>
24162
24163 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24164 scalars are valid operands.
24165
0fd44da3
MJ		 241662013-02-21 Martin Jambor <mjambor@suse.cz>
24167
24168 PR tree-optimization/56310
24169 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24170 only matching indices and non-negative final offsets.
24171 (intersect_aggregates_with_edge): Pass src_idx to
24172 agg_replacements_to_vector. Pass src_idx insstead of index to
24173 intersect_with_agg_replacements.
24174
7a92038b
MJ		 241752013-02-21 Martin Jambor <mjambor@suse.cz>
24176
24177 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24178 instead of hard-wired defaults.
24179
c0da9c37
MR		 241802013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24181
24182 * doc/invoke.texi (MIPS Options): Update documentation of the
24183 floating-point multiply-accumulate instruction restrictions.
24184
d247ea0c 241852013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 24186
24187 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24188 asan_shadow_offset on x86_64 linux.
24189
22deefcb
RB		 241902013-02-21 Richard Biener <rguenther@suse.de>
24191
24192 PR tree-optimization/56415
24193 Revert
24194 2013-02-11 Richard Biener <rguenther@suse.de>
24195
24196 PR tree-optimization/56273
24197 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24198 first VRP run.
24199
7bcc6e75
JJ		 242002013-02-21 Jakub Jelinek <jakub@redhat.com>
24201
a0ad148f
JJ		 24202 PR bootstrap/56258
24203 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24204 instead of @itemx.
24205
7bcc6e75
JJ		 24206 PR inline-asm/56405
24207 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24208 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24209
a0a7b611
JH		 242102013-02-20 Jan Hubicka <jh@suse.cz>
24211
24212 PR tree-optimization/56265
88b97037
UB		 24213 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24214 when target is referenced for first time.
a0a7b611 24215
c0e50f72
RB		 242162013-02-20 Richard Biener <rguenther@suse.de>
24217
24218 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24219 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24220 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24221 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24222 not return anything.
24223 (rename_ssa_copies): Do not remove unused locals.
24224 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 24225 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 24226 * passes.c (execute_function_todo): Do not schedule unused locals
24227 removal if cleanup_tree_cfg did something.
24228 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24229 about the number of removed locals.
24230
a52ca739
RB		 242312013-02-20 Richard Biener <rguenther@suse.de>
24232
24233 PR tree-optimization/56398
88b97037 24234 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 24235
ec9c9d1b
MJ		 242362013-02-20 Martin Jambor <mjambor@suse.cz>
24237
24238 PR tree-optimization/55334
24239 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24240 restricted pointers to arrays.
24241
e91c8ed6 242422013-02-20 Richard Biener <rguenther@suse.de>
88b97037 24243 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 24244
24245 PR tree-optimization/56396
24246 * tree-ssa-ccp.c (n_const_val): New static variable.
24247 (get_value): Return NULL for SSA names we don't have a lattice
24248 entry for.
24249 (ccp_initialize): Initialize n_const_val.
24250 * tree-ssa-copy.c (n_copy_of): New static variable.
24251 (init_copy_prop): Initialize n_copy_of.
24252 (get_value): Return NULL_TREE for SSA names we don't have a
24253 lattice entry for.
24254
3d916479
MJ		 242552013-02-20 Martin Jambor <mjambor@suse.cz>
24256
24257 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24258
71a86758
RB		 242592013-02-20 Richard Biener <rguenther@suse.de>
24260
24261 * genpreds.c (write_lookup_constraint): Do not compare first
24262 letter of the constraint again.
24263
79836a12
RB		 242642013-02-20 Richard Biener <rguenther@suse.de>
24265
24266 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24267 and ceil_log2.
24268 (get_use_iv_cost): Terminate hashtable walk when coming across
24269 an empty entry.
24270
bbe4fb2c
IZ		 242712013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24272
24273 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24274 reassociation for avx2 targets.
24275
a72d8780 242762012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 24277
c77f83d5
EI		 24278 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24279 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 24280 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 24281 version and TARGET_PATTERN_COMPARE check
24282 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 24283
a72d8780 242842012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 24285
a72d8780 24286 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 24287 function before branching.
24288
242387fa
AB		 242892012-02-19 Andrey Belevantsev <abel@ispras.ru>
24290
24291 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24292 DUMP_INSN_RTX_UID.
24293 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24294
b4979ab9
AB		 242952012-02-19 Andrey Belevantsev <abel@ispras.ru>
24296
24297 PR middle-end/55889
b4979ab9
AB		 24298 * sel-sched.c: Include ira.h.
24299 (implicit_clobber_conflict_p): New function.
24300 (moveup_expr): Use it.
88b97037 24301 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 24302
24d63016
RB		 243032013-02-19 Richard Biener <rguenther@suse.de>
24304
24305 PR tree-optimization/56384
24306 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24307 (vn_hash_type): Split out from ...
24308 (vn_hash_constant_with_type): ... here.
24309 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24310 (vn_phi_eq): Compare types from vn_phi_s structure.
24311 (vn_phi_lookup): Populate vn_phi_s type.
24312 (vn_phi_insert): Likewise.
24313
a475fd3d
JJ		 243142013-02-19 Jakub Jelinek <jakub@redhat.com>
24315
47cc28f5
JJ		 24316 PR tree-optimization/56350
24317 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24318 if haven't found reduction or nested cycle operand, rather than
24319 asserting we must find it.
24320
a475fd3d
JJ		 24321 PR tree-optimization/56381
24322 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24323 to fold_build3.
24324
243252013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 24326 Jakub Jelinek <jakub@redhat.com>
24327
24328 PR target/52555
24329 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24330 (swap_optab_enable): Same.
24331 (init_all_optabs): Use argument instead of global.
88b97037 24332 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 24333 * expr.h (init_all_optabs): Add argument to prototype.
24334 (TREE_OPTIMIZATION_OPTABS): New.
24335 (save_optabs_if_changed): Protoize.
24336 * optabs.h: Declare this_fn_optabs.
24337 * optabs.c (save_optabs_if_changed): New.
24338 Declare this_fn_optabs.
24339 (init_optabs): Add argument to init_all_optabs() call.
24340 * function.c (invoke_set_current_function_hook): Handle per
24341 function optabs.
24342 * function.h (struct function): New field optabs.
24343 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24344 optimization_current_node has changed.
24345 * target-globals.h (save_target_globals_default_opts): Protoize.
24346 * target-globals.c (save_target_globals_default_opts): New.
24347
3f587ca3
JDA		 243482013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24349
24350 PR target/56347
24351 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24352 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 24353
24354 PR target/56214
24355 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 24356 and HImode, require all displacements to be an integer multiple of
24357 their mode size.
ceaca33e
JDA		 24358 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24359 only allow QImode and HImode when reload is in progress and strict is
24360 true. Likewise for symbolic addresses. Use base14_operand to check
24361 displacements in REG+BASE addresses.
24362
fe0b4796
RB		 243632013-02-18 Richard Biener <rguenther@suse.de>
24364
24365 PR tree-optimization/56366
24366 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24367 handle sign-conversion of outer-loop initial induction value.
24368
6aaf596b
RB		 243692013-02-18 Richard Biener <rguenther@suse.de>
24370
73db8ff1 24371 PR middle-end/56349
6aaf596b
RB		 24372 * cfghooks.c (merge_blocks): If we merge a latch into another
24373 block adjust references to it.
24374 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24375 (verify_loop_structure): Verify that a recorded latch is in fact
24376 a latch.
24377
5e97dfb6
RB		 243782013-02-18 Richard Biener <rguenther@suse.de>
24379
24380 PR tree-optimization/56321
24381 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24382 order SSA name release and virtual operand unlinking.
24383
825527e8
EI		 243842013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24385
24386 * config/microblaze/microblaze.md (save_stack_block): Define.
24387 (restore_stack_block): Likewise.
24388
debd11d9
EI		 243892013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24390
24391 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24392 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24393 * config/microblaze/microblaze.c (microblaze_option_override):
24394 Bail out early for PIC modes when target does not support PIC.
24395
8ec77be0
EI		 243962013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24397
88b97037 24398 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 24399 Replace with a microblaze version.
24400 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 24401 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 24402 microblaze.
24403
b41288b3
JJ		 244042013-02-16 Jakub Jelinek <jakub@redhat.com>
24405 Dodji Seketeli <dodji@redhat.com>
24406
24407 PR asan/56330
88b97037 24408 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 24409 (instrument_mem_region_access): Do not forget to always put
24410 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24411 0) statement, even for cases where either 'base' or 'base + len'
24412 are not instrumented -- because they have been previously
24413 instrumented. Simplify the logic by putting all the statements
24414 instrument 'base + len' inside a sequence, and then insert that
24415 sequence right before the current insertion point. Then, to
24416 instrument 'base + len', just get an iterator on that statement.
24417 And do not forget to update the pointer to iterator the function
24418 received as argument.
24419
47918951
VM		 244202013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24421
24422 PR rtl-optimization/56348
24423 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24424
3fb7c699
SB		 244252013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24426
24427 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24428 (clean_graph_dump_file): Pass base to start_graph_dump.
24429
576fe41a
RH		 244302013-02-14 Richard Henderson <rth@redhat.com>
24431
24432 PR target/55941
24433 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24434
bb3accfa
SB		 244352013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24436
24437 * collect2-aix.h: Define F_LOADONLY.
24438
03143140
RB		 244392013-02-14 Richard Biener <rguenther@suse.de>
24440
24441 PR lto/50494
24442 * varasm.c (output_constant_def_1): Get the decl representing
24443 the constant as argument.
24444 (output_constant_def): Wrap output_constant_def_1.
24445 (make_decl_rtl): Use output_constant_def_1 with the decl
24446 representing the constant.
24447 (build_constant_desc): Optionally re-use a decl already
24448 representing the constant.
24449 (tree_output_constant_def): Adjust.
24450
8fb06726
DS		 244512013-02-14 Dodji Seketeli <dodji@redhat.com>
24452
24453 Fix an asan crash
24454 * asan.c (instrument_builtin_call): Really put the length of the
24455 second source argument into src1_len.
24456
c4bfe8bf
JJ		 244572013-02-13 Jakub Jelinek <jakub@redhat.com>
24458
24459 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24460 argument. If it is false, don't create edge from then_bb to
24461 fallthru_bb.
24462 (insert_if_then_before_iter): Pass true to it.
24463 (build_check_stmt): Pass false to it.
24464 (transform_statements): Flush hash table only on extended basic
24465 block boundaries, rather than at the beginning of every bb.
24466 Don't flush hash table on nonfreeing_call_p calls.
24467 * tree-flow.h (nonfreeing_call_p): New prototype.
24468 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24469
7afe2801
DM		 244702013-02-13 David S. Miller <davem@davemloft.net>
24471
24472 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24473
70cc3288
VM		 244742013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24475
24476 PR target/56184
24477 * ira.c (max_regno_before_ira): Move from ...
24478 (ira): ... here.
24479 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24480 vec_safe_length.
24481
6422242b
JJ		 244822013-02-13 Jakub Jelinek <jakub@redhat.com>
24483
24484 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24485
976a81ee
RB		 244862013-02-13 Richard Biener <rguenther@suse.de>
24487
24488 PR lto/56295
24489 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24490 globals in MEM_REFs.
24491
c1874a87
RB		 244922013-02-13 Richard Biener <rguenther@suse.de>
24493
24494 * loop-init.c (loop_optimizer_init): Clear loop state when
24495 re-initializing preserved loops.
24496 * loop-unswitch.c (unswitch_single_loop): Return whether
24497 we unswitched the loop. Do not verify loop state here.
88b97037 24498 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24499
b4ab7d34
KS		 245002013-02-13 Kostya Serebryany <kcc@google.com>
24501
88b97037
UB		 24502 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24503 on x86_64 linux.
b4ab7d34
KS		 24504 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24505
bdcbe80c
DS		 245062013-02-12 Dodji Seketeli <dodji@redhat.com>
24507
24508 Avoid instrumenting duplicated memory access in the same basic block
24509 * Makefile.in (asan.o): Add new dependency on hash-table.h
24510 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24511 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24512 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24513 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24514 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24515 (get_mem_ref_of_assignment): New functions.
24516 (get_mem_refs_of_builtin_call): Extract from
24517 instrument_builtin_call and tweak a little bit to make it fit with
24518 the new signature.
24519 (instrument_builtin_call): Use the new
24520 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24521 of is_gimple_builtin_call.
24522 (instrument_derefs, instrument_mem_region_access): Insert the
24523 instrumented memory reference into the hash table.
24524 (maybe_instrument_assignment): Renamed instrument_assignment into
24525 this, and change it to advance the iterator when instrumentation
24526 actually happened and return true in that case. This makes it
24527 homogeneous with maybe_instrument_assignment, and thus give a
24528 chance to callers to be more 'regular'.
24529 (transform_statements): Clear the memory reference hash table
24530 whenever we enter a new BB, when we cross a function call, or when
24531 we are done transforming statements. Use
24532 maybe_instrument_assignment instead of instrumentation. No more
24533 need to special case maybe_instrument_assignment and advance the
24534 iterator after calling it; it's now handled just like
24535 maybe_instrument_call. Update comment.
24536
4861a1f7
RB		 245372013-02-13 Richard Biener <rguenther@suse.de>
24538
24539 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24540 Fix loop discovery code.
24541
1c86bd80
VM		 245422013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24543
24544 PR inline-asm/56148
24545 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24546 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 24547 the operand is not reloaded. Prefer to reload conflicting operand
24548 if earlyclobber and matching operands are the same.
24549
7cbda518
RB		 245502013-02-12 Richard Biener <rguenther@suse.de>
24551
24552 PR lto/56297
24553 * lto-streamer-out.c (write_symbol): Do not output symbols
24554 for hard register variables.
24555
e68a4ef6
GJL		 245562013-02-12 Georg-Johann Lay <avr@gjlay.de>
24557
24558 PR target/54222
24559 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24560 (umulsidi3_insn, mulsidi3_insn): New insns.
24561
a72d8780 245622013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 24563
24564 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24565 (struct tune_params): Add vec_costs field.
24566 * config/arm/arm.c (arm_builtin_vectorization_cost)
24567 (arm_add_stmt_cost): New functions.
24568 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24569 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24570 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24571 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24572 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24573 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24574 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24575
43320568
RB		 245762013-02-12 Richard Biener <rguenther@suse.de>
24577
24578 PR lto/56295
24579 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24580 decls again if possible.
24581
a011aa39
RB		 245822013-02-12 Richard Biener <rguenther@suse.de>
24583
24584 PR middle-end/56288
24585 * tree-ssa.c (verify_ssa_name): Fix check, move
24586 SSA_NAME_IN_FREE_LIST check up.
24587
6da26889
JJ		 245882013-02-12 Jakub Jelinek <jakub@redhat.com>
24589 Steven Bosscher <steven@gcc.gnu.org>
24590
24591 PR rtl-optimization/56151
24592 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24593 equal to op0 or op1, and last_insn pattern is CODE operation
24594 with MEM dest and one of the operands matches that MEM.
24595
f80e0faf
ST		 245962013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24597
24598 * doc/extend.texi: Document Function Multiversioning and "default"
24599 parameter string to target attribute.
24600 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24601 target attribute parameter is "default".
24602 (ix86_compare_version_priority): Remove checks for target attribute.
24603 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24604 Remove check for target attribute equal to NULL. Add assert.
24605 (ix86_generate_version_dispatcher_body): Change error to sorry.
24606
6c59ffd1
IS		 246072013-02-11 Iain Sandoe <iain@codesourcery.com>
24608 Jack Howarth <howarth@bromo.med.uc.edu>
24609 Patrick Marlier <patrick.marlier@gmail.com>
24610
24611 PR libitm/55693
24612 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24613 define ENDFILE_SPEC as TM_DESTRUCTOR.
24614 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24615
a72d8780 246162013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 24617 Jack Howarth <howarth@bromo.med.uc.edu>
24618 Jakub Jelinek <jakub@redhat.com>
24619
24620 PR sanitizer/55617
24621 * config/darwin.c (cdtor_record): Rename ctor_record.
24622 (sort_cdtor_records): Rename sort_ctor_records.
24623 (finalize_dtors): New routine to sort destructors by
24624 priority before use in assemble_integer.
24625 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24626
b63fe007
UB		 246272013-02-11 Uros Bizjak <ubizjak@gmail.com>
24628
24629 PR rtl-optimization/56275
24630 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24631 offset is non-negative and less than cmode size before
24632 calling simplify_subreg.
24633
8e89b5b5
RB		 246342013-02-11 Richard Biener <rguenther@suse.de>
24635
24636 PR tree-optimization/56264
24637 * cfgloop.h (fix_loop_structure): Adjust prototype.
24638 * loop-init.c (fix_loop_structure): Return the number of
24639 newly discovered loops.
24640 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24641 are discovered, do a full loop-closed SSA rewrite.
24642
b4a4b56d
RB		 246432013-02-11 Richard Biener <rguenther@suse.de>
24644
24645 PR tree-optimization/56273
24646 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24647 first VRP run.
24648 (check_array_ref): Fix missing newline in dumps.
24649 (search_for_addr_array): Likewise.
24650
0c885229
DE		 246512013-02-09 David Edelsohn <dje.gcc@gmail.com>
24652
24653 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24654
59ac9a55
JJ		 246552013-02-09 Jakub Jelinek <jakub@redhat.com>
24656
24657 PR target/56256
24658 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24659
25bb0bb5
VM		 246602013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24661
24662 PR rtl-optimization/56246
0c885229 24663 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 24664 reload pseudo.
24665 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24666 constraints are satisfied.
24667
a698cc03
JL		 246682013-02-08 Jeff Law <law@redhat.com>
24669
24670 PR debug/53948
24671 * emit-rtl.c (reg_is_parm_p): New function.
24672 * regs.h (reg_is_parm_p): New prototype.
24673 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24674 callee-clobbered registers.
24675
e1122ddd
MM		 246762013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24677
24678 PR target/56043
24679 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24680 If there is no implicit builtin declaration, just return NULL.
24681
19c5f6e6
UB		 246822013-02-08 Uros Bizjak <ubizjak@gmail.com>
24683
24684 * config/i386/sse.md (FMAMODEM): New mode iterator.
24685 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24686 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24687
2480f2ca 246882013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 24689
2480f2ca
UB		 24690 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24691 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24692 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24693
246942013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24695
24696 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24697 (microblaze*-*-elf): Likewise.
24698 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24699 LINK_SPEC.
24700 * config/microblaze/microblaze-c.c: Add builtin defines for
24701 _LITTLE_ENDIAN and _BIG_ENDIAN.
24702 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24703 add to TARGET_DEFAULT flags.
76ef61fb 24704 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 24705 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24706 * config/microblaze/microblaze.md: Update extendsidi2 and
24707 movdi_internal instructions to use low-order / high-order reg
24708 print_operands.
24709 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24710 options and inversemask / mask of LITTLE_ENDIAN.
24711 * config/microblaze/t-microblaze: Expand multilib options to
24712 include mlittle-endian (le) and update exceptions patterns.
24713
600a5961
JJ		 247142013-02-08 Jakub Jelinek <jakub@redhat.com>
24715
5df81313
JJ		 24716 PR rtl-optimization/56195
24717 * lra-constraints.c (get_reload_reg): Don't reuse regs
24718 if they have smaller mode than requested, if they have
24719 wider mode than requested, try to return a SUBREG.
24720
600a5961
JJ		 24721 PR tree-optimization/56250
24722 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24723 if type is unsigned and code isn't MULT_EXPR.
24724
ff544649
GJL		 247252013-02-08 Georg-Johann Lay <avr@gjlay.de>
24726
24727 PR tree-optimization/56064
24728 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24729 bits according to mode.
24730 * fixed-value.h (fixed_from_double_int)
24731 (const_fixed_from_double_int): Adjust comments.
24732
e45cde98
RB		 247332013-02-08 Richard Biener <rguenther@suse.de>
24734
24735 PR lto/56231
24736 * lto-streamer.h (struct data_in): Remove current_file, current_line
24737 and current_col members.
24738 * lto-streamer-out.c (lto_output_location): Stream changed bits
24739 en-block for efficiency.
24740 * lto-streamer-in.c (clear_line_info): Remove.
24741 (lto_input_location): Cache current file, line and column
24742 globally via local statics. Read changed bits en-block.
24743 (input_function): Do not call clear_line_info.
24744 (lto_read_body): Likewise.
24745 (lto_input_toplevel_asms): Likewise.
24746
c1ca73d8
MM		 247472013-02-08 Michael Matz <matz@suse.de>
24748
24749 PR tree-optimization/52448
24750 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24751 (nt_call_phase): New static.
24752 (add_or_mark_expr): Only mark accesses with newer phase than any
24753 call seen.
24754 (nonfreeing_call_p): New.
24755 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24756 (nt_fini_block): Keep blocks marked as visited.
24757 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24758
57548aa2
RB		 247592013-02-08 Richard Biener <rguenther@suse.de>
24760
24761 * ira.c (ira): Free broken dominator information.
24762
8e10366f
UB		 247632013-02-08 Uros Bizjak <ubizjak@gmail.com>
24764
24765 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24766
f64fb0fa
MP		 247672013-02-08 Marek Polacek <polacek@redhat.com>
24768
8e10366f 24769 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 24770
0375167b
RB		 247712013-02-08 Richard Biener <rguenther@suse.de>
24772
24773 PR middle-end/56181
24774 * cfgloop.h (flow_loops_find): Adjust.
24775 (bb_loop_header_p): Declare.
24776 * cfgloop.c (bb_loop_header_p): New function split out from ...
24777 (flow_loops_find): ... here. Adjust function signature,
24778 support incremental loop structure update.
24779 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24780 * cfgloopmanip.c (fix_loop_structure): Move ...
24781 * loop-init.c (fix_loop_structure): ... here.
24782 (apply_loop_flags): Split out from ...
24783 (loop_optimizer_init): ... here.
24784 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24785 in incremental mode, only remove dead loops here.
24786
85d768f3
GJL		 247872013-02-08 Georg-Johann Lay <avr@gjlay.de>
24788
24789 PR target/54222
24790 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24791 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24792 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24793 (*round<mode>3.libgcc): New insns for fixed-modes.
24794 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24795 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24796 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24797 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24798 implementations. Define to __builtin_avr_absFX,
24799 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24800 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24801 __builtin_avr_countlsFX, respectively.
24802 * config/avr/avr-c.c (target.h): Include it.
24803 (enum avr_builtin_id): New enum.
24804 (avr_resolve_overloaded_builtin): New static function.
24805 (avr_register_target_pragmas): Use it to set
24806 targetm.resolve_overloaded_builtin.
24807 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24808 tree nodes used by DEF_BUILTIN.
24809 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24810 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24811 <AVR_BUILTIN_xxBITS>: Same.
24812
661bc682
RB		 248132013-02-08 Richard Biener <rguenther@suse.de>
24814
24815 * cfgloop.c (verify_loop_structure): Properly handle
24816 a loop exiting to another loop header.
24817 * ira-int.h (ira_loops): Remove.
24818 * ira.c (ira_loops): Remove.
24819 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24820 (do_reload): Use loop_optimizer_finalize.
24821 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24822 number_of_loops to access the loop tree.
24823 (more_one_region_p): Likewise.
24824 (finish_loop_tree_nodes): Likewise.
24825 (rebuild_regno_allocno_maps): Likewise.
24826 (mark_loops_for_removal): Likewise.
24827 (mark_all_loops_for_removal): Likewise.
24828 (remove_unnecessary_regions): Likewise.
24829 (ira_build): Likewise.
24830 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24831
0d5049b2
RB		 248322013-02-08 Richard Biener <rguenther@suse.de>
24833
24834 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24835 * ipa-pure-const.c (analyze_function): Avoid calling
24836 mark_irreducible_loops twice.
8e10366f 24837 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 24838
499f32e8
DM		 248392013-02-07 David S. Miller <davem@davemloft.net>
24840
24841 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24842 on 'reg'.
24843 * var-tracking.c (vt_add_function_parameter): Test the presence of
24844 HAVE_window_save properly and do not remap argument registers when
24845 we have a leaf function.
24846
6edc3e32
UB		 248472013-02-07 Uros Bizjak <ubizjak@gmail.com>
24848
24849 PR bootstrap/56227
24850 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24851 instead of "ll".
24852 * config/i386/i386.c (ix86_print_operand): Ditto.
24853
5306401f
VM		 248542013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24855
6edc3e32 24856 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 24857
027ece11
VM		 248582013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24859
24860 PR rtl-optimization/56225
24861 * lra-constraints.c (process_alt_operands): Check that reload hard
24862 reg can hold value for strict_low_part.
24863
f980dfdb
JJ		 248642013-02-07 Jakub Jelinek <jakub@redhat.com>
24865
24866 PR debug/56154
24867 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24868 dwarf2out_end_function.
24869 (in_first_function_p, maybe_at_text_label_p,
24870 first_loclabel_num_not_at_text_label): New variables.
24871 (dwarf2out_var_location): In the first function find out
6edc3e32 24872 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 24873 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24874 functions.
24875
73dd3123
EB		 248762013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24877
24878 PR rtl-optimization/56178
24879 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24880 SUBREG of a register. Tidy up related block of code.
24881 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24882 note if the source is a register or a SUBREG of a register.
24883
e3936f47
JJ		 248842013-02-07 Jakub Jelinek <jakub@redhat.com>
24885
24886 PR target/56228
24887 * config/rs6000/rs6000.md (ptrm): New mode attr.
24888 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24889 call_value_indirect_aix<pttrsize>,
24890 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24891 m in constraints.
24892
d96d674b
MH		 248932013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24894
24895 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24896 if -bnortl. Convert to strcmp and strncmp.
24897
921f2dee
AM		 248982013-02-07 Alan Modra <amodra@gmail.com>
24899
24900 PR target/54009
24901 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24902 addresses won't wrap when offsetting.
24903 (rs6000_secondary_reload): Provide secondary reloads needed for
24904 wrapping LO_SUM addresses.
24905
d09c7dba
TS		 249062013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24907
24908 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24909 MACH, just __MACH__.
24910
a44bbd48
RB		 249112013-02-06 Richard Biener <rguenther@suse.de>
24912
24913 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24914 instead of calling fix_loop_structure.
24915
6d840d99
JJ		 249162013-02-06 Jakub Jelinek <jakub@redhat.com>
24917
24918 PR middle-end/56217
24919 * omp-low.c (use_pointer_for_field): Return false if
24920 lower_send_shared_vars doesn't generate any copy-out code.
24921
0f33baa9
TV		 249222013-02-06 Tom de Vries <tom@codesourcery.com>
24923
24924 PR rtl-optimization/56131
24925 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24926 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24927 of the label is NULL. Add comment.
24928
11f1e3ab
JJ		 249292013-02-05 Jakub Jelinek <jakub@redhat.com>
24930
31502f9f
JJ		 24931 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24932
a74db9bd
JJ		 24933 PR sanitizer/55374
24934 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24935 (STATIC_LIBTSAN_LIBS): Likewise.
24936 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24937 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24938 is defined, don't add anything else beyond that.
24939 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24940 (LINK_COMMAND_SPEC): Use them.
24941
11f1e3ab
JJ		 24942 PR tree-optimization/56205
24943 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24944 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24945 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24946
14ac6aa2
RB		 249472013-02-05 Richard Biener <rguenther@suse.de>
24948
24949 PR tree-optimization/53342
24950 PR tree-optimization/53185
24951 * tree-vectorizer.h (vect_check_strided_load): Remove.
24952 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24953 not disallow peeling for vectorized strided loads.
24954 (vect_check_strided_load): Make static and simplify.
24955 (vect_analyze_data_refs): Adjust.
24956 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24957 correctly when vectorizing strided loads.
24958
6f22445a
RB		 249592013-02-05 Richard Biener <rguenther@suse.de>
24960
24961 * doc/install.texi: Refer to ISL, not PPL.
24962
39f9719e
JH		 249632013-02-05 Jan Hubicka <jh@suse.cz>
24964
ec4224ac
JH		 24965 PR tree-optimization/55789
24966 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24967
249682013-02-05 Jan Hubicka <jh@suse.cz>
24969
24970 PR tree-optimization/55789
39f9719e
JH		 24971 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24972 the dead call anyway.
24973
956d3b33
EB		 249742013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24975
24976 PR sanitizer/55374
24977 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24978
249792013-02-04 Alexander Potapenko <glider@google.com>
24980 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 24981 Jakub Jelinek <jakub@redhat.com>
24982
24983 PR sanitizer/55617
24984 * config/darwin.c (sort_ctor_records): Stabilized qsort
24985 on constructor priority by using original position.
24986 (finalize_ctors): New routine to sort constructors by
24987 priority before use in assemble_integer.
24988 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24989
7ac3af38
JJ		 249902013-02-04 Jakub Jelinek <jakub@redhat.com>
24991
24992 PR libstdc++/54314
24993 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24994 about visibility on artificial decls.
24995 * config/sol2.c (solaris_assemble_visibility): Likewise.
24996
152689dc
KT		 249972013-02-04 Kai Tietz <ktietz@redhat.com>
24998
24999 PR target/56186
25000 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25001 argument and improve checking of return-argument types for 16-byte
25002 modes.
25003 (ix86_function_value_1): Add additional valtype argument on call
25004 of function_value_64.
25005 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25006 handling infunction_value_64 function.
25007
2b5987b5
MGD		 250082013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25009
7ac3af38 25010 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 25011
14597080
RB		 250122013-02-04 Richard Biener <rguenther@suse.de>
25013
25014 PR tree-optimization/56188
25015 * tree-ssa-structalias.c (label_visit): Consider case with
25016 initially non-empty points-to set.
25017 (perform_var_substitution): Dump node mapping and clean up.
25018
ed73881e
RG		 250192013-02-04 Richard Guenther <rguenther@suse.de>
25020
25021 PR lto/56168
25022 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25023 node prevail as last resort.
25024 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 25025 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 25026
139a0707
RB		 250272013-02-04 Richard Biener <rguenther@suse.de>
25028
25029 PR tree-optimization/56113
25030 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25031 Merge into ...
25032 (equiv_class_lookup_or_add): ... this.
25033 (label_visit): Adjust and fix error in previous patch.
25034 (perform_var_substitution): Adjust.
25035
50fe8924
OE		 250362013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25037
25038 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25039 * config/sh/sh4-300.md: Likewise.
25040 * config/sh/sh4a.md: Likewise.
25041 * config/sh/constraints.md: Likewise.
25042 * config/sh/sh.md: Likewise.
25043 * config/sh/netbsd-elf.h: Likewise.
25044 * config/sh/predicates.md: Likewise.
25045 * config/sh/sh-protos.h: Likewise.
25046 * config/sh/ushmedia.h: Likewise.
25047 * config/sh/linux.h: Likewise.
25048 * config/sh/sh.c: Likewise.
25049 * config/sh/superh.h: Likewise.
25050 * config/sh/elf.h: Likewise.
25051 * config/sh/sh4.md: Likewise.
25052 * config/sh/sh.h: Likewise.
25053
1a04ac2b
JDA		 250542013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25055
25056 * config/pa/constraints.md: Adjust unused letters. Change "T"
25057 constraint to match_test floating_point_store_memory_operand().
25058 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25059 (base14_operand): New.
25060 (floating_point_store_memory_operand): New.
25061 (integer_store_memory_operand): Revise to use base14_operand and
25062 reg_plus_base_memory_operand.
25063 (move_dest_operand): Allow symbolic_memory_operands.
25064 (symbolic_memory_operand): Check for LO_SOM.
25065 (symbolic_operand): Change default case to break.
25066 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25067 CONST_DOUBLE values to be reloaded by putting them into memory when
25068 the destination is a floating point register.
25069 (movdf): Remove code to handle CONST_DOUBLE.
25070 (movsf): Likewise.
25071 (reload_indf_r1): New.
25072 (reload_insf_r1): New.
25073 Consistently use "Q" and "T" constraints with integer and floating
25074 point move instructions, respectively.
25075 (movdi): Remove FAIL.
25076 Change predicate for source operand unamed DImode move from
25077 general_operand to move_src_operand.
25078 (umulsidi3): Change predicate for destination operand to
25079 register_operand.
25080 Likewise for similar unamed patterns.
25081 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25082 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25083 (hppa_legitimize_address): Simplify mask calculation.
25084 (pa_emit_move_sequence): Revised handling of secondary reloads from
25085 REG+D addresses for floating point loads and stores. Directly handle
25086 loading CONST0_RTX (mode) to a floating point register.
25087 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25088 to floating point registers. Don't restrict secondary reloads to
25089 floating point registers to integer modes. Revise some comments and
25090 cleanup some code.
25091 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25092 (pa_legitimate_address_p): New.
25093 (pa_legitimize_reload_address): New.
25094 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25095 (STRICT_REG_OK_FOR_BASE_P): New.
25096 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25097 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25098
3bf65591
DE		 250992013-02-03 David Edelsohn <dje.gcc@gmail.com>
25100 Andrew Dixie <andrewd@gentrack.com>
25101
25102 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25103 flag set.
25104
c0a8a3e6
RS		 251052013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25106
25107 * expmed.c (extract_bit_field_1): Pass the full width of the
25108 structure to get_best_reg_extraction_insn.
25109
99113dff
DE		 251102013-02-01 David Edelsohn <dje.gcc@gmail.com>
25111
25112 PR target/54601
25113 * configure.ac (use_cxa_atexit): Add AIX.
25114 * configure: Regenerate.
25115
25116 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25117
ddd84654
JJ		 251182013-02-01 Jakub Jelinek <jakub@redhat.com>
25119
25120 PR debug/54793
25121 * final.c (need_profile_function): New variable.
25122 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25123 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25124 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25125 notes, targetm.asm_out.function_prologue doesn't emit anything,
25126 HAVE_prologue and profiler should be emitted before prologue,
25127 set need_profile_function instead of emitting it.
25128 (final_scan_insn): If need_profile_function, emit
25129 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25130 NOTE_INSN_FUNCTION_BEG note.
25131
251322013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 25133
25134 * config/rs6000/rs6000.md (smulditi3): New.
25135 (umulditi3): New.
25136
ff2a9d88
RH		 25137 * config/alpha/alpha.md (umulditi3): New.
25138
14d52b90
DE		 251392013-02-01 David Edelsohn <dje.gcc@gmail.com>
25140
25141 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25142 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25143
8c7ca45c
RB		 251442013-02-01 Richard Biener <rguenther@suse.de>
25145
25146 PR tree-optimization/56113
25147 * tree-ssa-structalias.c (label_visit): Reduce work for
25148 single-predecessor nodes.
25149
9f419393
EB		 251502013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25151
25152 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25153 range isn't testing for zero.
25154
c5f4be84
SB		 251552013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25156
25157 PR middle-end/56113
25158 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25159
dbdbd982
NC		 251602013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25161 Nick Clifton <nickc@redhat.com>
25162
25163 * config/v850/constraints.md (Q): Define as a memory constraint.
25164 * config/v850/predicates.md (label_ref_operand): New predicate.
25165 (e3v5_shift_operand): New predicate.
25166 (ior_operator): New predicate.
25167 * config/v850/t-v850: Add e3v5 multilib.
25168 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25169 (v850_gen_movdi): Prototype.
25170 * config/v850/v850.c: Add support for e3v5 architecture.
25171 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25172 TARGET_V850E_UP.
25173 (construct_save_jarl): Add e3v5 long JARL support.
25174 (v850_adjust_insn_length): New function. Adjust length of call
25175 insns when using e3v5 instructions.
25176 (v850_gen_movdi): New function: Generate instructions to move a
25177 DImode value.
25178 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25179 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25180 (TARGET_USE_FPU): Enable for e3v5.
25181 (CONST_OK_FOR_W): New macro.
25182 (ADJUST_INSN_LENGTH): Define.
25183 * config/v850/v850.md (UNSPEC_LOOP): Define.
25184 (attr cpu): Add v850e3v5.
25185 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25186 (movdi): New pattern.
25187 (movdi_internal): New pattern.
25188 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25189 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25190 (cstoresf4): Likewise.
25191 (cstoredf4): Likewise.
25192 (insv): New pattern.
25193 (rotlso3_a): New pattern.
25194 (rotlsi3_b): New pattern
25195 (rotlsi3_v850e3v5): New pattern.
25196 (doloop_begin): New pattern.
25197 (fix_loop_counter): New pattern.
25198 (doloop_end): New pattern.
25199 (branch_normal): Add e3v5 long branch support.
25200 (branch_invert): Likewise.
25201 (branch_z_normal): Likewise.
25202 (branch_z_invert): Likewise.
25203 (branch_nz_normal): Likewise.
25204 (branch_nz_invert): Likewise.
25205 (call_internal_short): Add e3v5 register-indirect JARL support.
25206 (call_internal_long): Likewise.
25207 (call_value_internal_short): Likewise.
25208 (call_value_internal_long): Likewise.
25209 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25210 (mloop): New option.
25211 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 25212 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 25213
73861a41
PK		 252142013-01-31 Paul Koning <ni1d@arrl.net>
25215
25216 PR debug/55059
25217 PR debug/54508
25218 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25219 children if parent is a class.
25220 (prune_unused_types_prune): Don't add DW_AT_declaration.
25221
e44978dc
RB		 252222013-01-31 Richard Biener <rguenther@suse.de>
25223
25224 PR tree-optimization/56157
25225 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25226 match up operand with SLP child.
25227
90b10dec
JM		 252282013-01-31 Jason Merrill <jason@redhat.com>
25229
04d2dadd 25230 PR debug/54410
90b10dec
JM		 25231 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25232 parameters the first time.
25233 (gen_scheduled_generic_parms_dies): Check completeness here.
25234
6e616110
RB		 252352013-01-31 Richard Biener <rguenther@suse.de>
25236
25237 PR middle-end/53073
25238 * common.opt (faggressive-loop-optimizations): New flag,
25239 enabled by default.
25240 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25241 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25242 infer_loop_bounds_from_undefined by it.
25243
636f59cf
RB		 252442013-01-31 Richard Biener <rguenther@suse.de>
25245
25246 PR tree-optimization/56150
25247 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25248 visit virtual operands.
25249 (find_uses_to_rename_bb): Likewise.
25250
252512013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 25252
25253 PR tree-optimization/56150
25254 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25255 mixed store non-store stmts.
25256
32887460
JJ		 252572013-01-30 Jakub Jelinek <jakub@redhat.com>
25258
e60e09a0
JJ		 25259 PR sanitizer/55374
25260 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25261 LIBASAN_EARLY_SPEC is defined.
25262 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25263 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25264 before %o.
25265 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25266
32887460
JJ		 25267 PR c++/55742
25268 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25269 invalid args instead of ICEing on it.
25270 (ix86_valid_target_attribute_tree): Return error_mark_node if
25271 ix86_valid_target_attribute_inner_p failed.
25272 (ix86_valid_target_attribute_p): Return false only if
25273 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25274 target("default") attribute.
25275 (sorted_attr_string): Change argument from const char * to tree,
25276 merge in all target attribute arguments rather than just one.
25277 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25278 instead of free. Avoid using strcat.
25279 (ix86_mangle_function_version_assembler_name): Mangle
25280 target("default") as if no target attribute is present. Adjust
25281 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25282 instead of xmalloc and XDELETEVEC instead of free.
25283 (ix86_function_versions): Don't return true if one of the decls
25284 doesn't have target attribute. If they don't and one of the decls
25285 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25286 sorted_attr_string caller. Use XDELETEVEC instead of free.
25287 (ix86_supports_function_versions): Remove.
25288 (make_name): Fix up formatting.
25289 (make_dispatcher_decl): Remove resolver_name and its initialization.
25290 Avoid leaking memory.
25291 (is_function_default_version): Return true if there is
25292 target("default") attribute rather than no target attribute at all.
25293 (make_resolver_func): Avoid leaking memory.
25294 (ix86_generate_version_dispatcher_body): Likewise.
25295 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25296 * target.def (supports_function_versions): Remove.
25297 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25298 * doc/tm.texi: Regenerated.
25299
73cca0cc
VM		 253002013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25301
25302 PR rtl-optimization/56144
25303 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25304 for values with side effects.
25305
11452e7b
RB		 253062013-01-30 Richard Biener <rguenther@suse.de>
25307
25308 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25309 (sparseset_pop): Likewise.
25310 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25311 to be able to use quick_push in the worker loop.
25312
01cb1ef5
MP		 253132013-01-30 Marek Polacek <polacek@redhat.com>
25314
25315 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25316
e892936e
RB		 253172013-01-30 Richard Biener <rguenther@suse.de>
25318
25319 PR lto/56147
6edc3e32 25320 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 25321
cc06c01d
GJL		 253222013-01-30 Georg-Johann Lay <avr@gjlay.de>
25323
25324 PR tree-optimization/56064
25325 * fixed-value.c (fixed_from_double_int): New function.
25326 * fixed-value.h (fixed_from_double_int): New prototype.
25327 (const_fixed_from_double_int): New static inline function.
25328 * fold-const.c (native_interpret_fixed): New static function.
25329 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25330 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25331 (native_encode_fixed): New static function.
25332 (native_encode_expr) <FIXED_CST>: Use it.
25333 (native_interpret_int): Move double_int worker code to...
25334 * double-int.c (double_int::from_buffer): ...this new static method.
25335 * double-int.h (double_int::from_buffer): Prototype it.
25336
d394a308
RB		 253372013-01-30 Richard Biener <rguenther@suse.de>
25338
25339 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25340 New pointer-map and obstack.
25341 (init_alias_vars): Allocate pointer-map and obstack.
25342 (delete_points_to_sets): Free them.
25343 (find_what_var_points_to): Cache result.
25344 (find_what_p_points_to): Adjust for changed interface of
25345 find_what_var_points_to.
25346 (compute_points_to_sets): Likewise.
25347 (ipa_pta_execute): Likewise.
25348
20804d96
RO		 253492013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25350
25351 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25352 * configure: Regenerate.
25353 * config.in: Regenerate.
25354 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25355 #nobits/#progbits if supported.
25356
a7ad88a2
OE		 253572013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25358
25359 PR target/56121
25360 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25361 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25362 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25363
7af79f92
GY		 253642013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25365
91bfca59
OE		 25366 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25367 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 25368
753bcf7b
GY		 253692013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25370
91bfca59
OE		 25371 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25372 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 25373
8cbc2ea8
GY		 253742013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25375
91bfca59
OE		 25376 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25377 declaration.
25378 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25379 * config/arm/cortex-a7.md: New bypasses using
25380 arm_mac_accumulator_is_result.
8cbc2ea8 25381
697a3325
GY		 253822013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25383
91bfca59 25384 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 25385 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 25386 (cortex_a7_fpfmad): New reservation.
25387 (cortex_a7_fpmacs): Use ffmas and update required units.
25388 (cortex_a7_fpmuld): Update required units and latency.
25389 (cortex_a7_fpmacd): Likewise.
25390 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25391 (cortex_a7_neon). Likewise.
25392 (bypass) Update participating units.
697a3325 25393
29637783
GY		 253942013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25395
91bfca59
OE		 25396 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25397 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25398 from fmac to ffma.
25399 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25400 (vfp_fmul): Use ffmad.
25401 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25402 (cortex_r4_fmacd): Use ffmad.
25403 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25404 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25405 (cortex_a9_fmacd): Use ffmad.
25406 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25407 (cortex_a8_vfp_macd): Use ffmad.
25408 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25409 (cortex_a5_fpmacd): Use ffmad.
25410 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25411 (cortex_a15_vfp_macd): Use ffmad.
25412 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 25413
2ee8a2d5
JM		 254142013-01-29 Jason Merrill <jason@redhat.com>
25415
25416 PR libstdc++/54314
25417 * varasm.c (default_assemble_visibility): Don't warn about
25418 visibility on artificial decls.
25419
78d087bc
RB		 254202013-01-29 Richard Biener <rguenther@suse.de>
25421
25422 PR tree-optimization/56113
25423 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25424 the bitmap leader.
25425 (label_visit): Free duplicate bitmaps and record the leader instead.
25426 (perform_var_substitution): Adjust.
25427
83ba4d6f
RB		 254282013-01-29 Richard Biener <rguenther@suse.de>
25429
25430 PR tree-optimization/55270
25431 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25432 the CFG, schedule loops for fixup.
25433
66dfe13f
NC		 254342013-01-29 Nick Clifton <nickc@redhat.com>
25435
25436 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25437 SP_REG.
25438
5a579c3b
LE		 254392013-01-28 Leif Ekblad <leif@rdos.net>
25440
25441 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25442 * config/i386/i386.h (TARGET_RDOS): New macro.
25443 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25444 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25445 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25446 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25447 DEFAULT_LARGE_SECTION_THRESHOLD.
25448 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25449 * config/i386/rdos.h: New file.
25450 * config/i386/rdos64.h: New file.
25451
51e44392
BS		 254522013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25453
25454 PR other/54814
25455 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25456 TEST_HARD_REG_BIT.
25457
db1fb332
JJ		 254582013-01-28 Jakub Jelinek <jakub@redhat.com>
25459
25460 PR rtl-optimization/56117
25461 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25462 call cselib_lookup_from_insn on the MEM before calling
25463 add_insn_mem_dependence.
25464
16917761
RB		 254652013-01-28 Richard Biener <rguenther@suse.de>
25466
25467 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25468 to a stmt that didn't have one.
25469 (copy_phis_for_bb): Likewise for PHI arguments.
25470 (copy_debug_stmt): Likewise for debug stmts.
25471
b9fc0497
RB		 254722013-01-28 Richard Biener <rguenther@suse.de>
25473
25474 PR tree-optimization/56034
6edc3e32 25475 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 25476 (partition_builtin_p): Adjust.
25477 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25478 it is the last partition.
25479 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25480 up the vertex for the definition.
25481 (classify_partition): Classify whether a partition is a
25482 PKIND_REDUCTION, thus has uses outside of the loop.
25483 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25484 Merge all PKIND_REDUCTION partitions into the last partition.
25485 (tree_loop_distribution): Seed partitions from reductions as well.
25486
aa710d25
JJ		 254872013-01-28 Jakub Jelinek <jakub@redhat.com>
25488
0bfbca58
JJ		 25489 PR tree-optimization/56125
25490 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25491 pow(x,c) into sqrt(x) * powi(x, n/2) or
25492 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25493 optimizing for size.
25494 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25495 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25496 integer.
25497
aa710d25
JJ		 25498 PR tree-optimization/56094
25499 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25500 to UNKNOWN_LOCATION while gimplifying expr.
25501
77dc5297
UB		 255022013-01-27 Uros Bizjak <ubizjak@gmail.com>
25503
25504 PR target/56114
25505 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25506 operand 0 in movabs insn template for -masm=intel asm alternative.
25507 (*movabs<mode>_2): Ditto for operand 1.
25508
0bfbca58 255092013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 25510
25511 PR target/54663
25512 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25513 of microblaze-c.o
25514
0bfbca58 255152013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 25516
25517 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25518 tm_file.
25519
0bfbca58 255202013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 25521
25522 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25523 Undef to avoid warning.
25524
478f60f9
MH		 255252013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25526
25527 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25528 * configure: Regenerate.
25529
d7fa6ee2
JJ		 255302013-01-25 Jakub Jelinek <jakub@redhat.com>
25531
25532 PR tree-optimization/56098
25533 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25534 for stmts with volatile ops.
25535 (cond_store_replacement): Don't optimize if assign has volatile ops.
25536 (cond_if_else_store_replacement_1): Don't optimize if either
25537 then_assign or else_assign have volatile ops.
25538 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25539 volatile ops.
25540
f8fe87bd
GJL		 255412013-01-25 Georg-Johann Lay <avr@gjlay.de>
25542
25543 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25544
28398d0d
GJL		 255452013-01-25 Georg-Johann Lay <avr@gjlay.de>
25546
25547 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25548 missing ':' in asm example.
25549
b7d7d917
TB		 255502013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25551
25552 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25553 entries into lane and laneq entries.
77dc5297
UB		 25554 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25555 Remove AdvSIMD scalar modes.
b7d7d917
TB		 25556 (aarch64_sq<r>dmulh_laneq<mode>): New.
25557 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25558 modes.
25559 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25560 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25561 * config/aarch64/iterators.md (VCOND): New.
25562 (VCONQ): New.
25563
556f9906
GJL		 255642013-01-25 Georg-Johann Lay <avr@gjlay.de>
25565
25566 PR target/54222
25567 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25568 Add NULL LIBNAME argument to existing definitions.
25569 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25570 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25571 * config/avr/avr.c (DEF_BUILTIN): Same.
25572 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25573 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25574 implementation is available (DECL_ASSEMBLER_NAME is set).
25575 (avr_fold_absfx): New static function.
25576 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25577 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25578 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25579 AVR_BUILTIN_ABSLLK.
25580 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25581 (abshk, absk, abslk, absllk): Provide as static inline functions.
25582
1f546bbb
MP		 255832013-01-25 Marek Polacek <polacek@redhat.com>
25584
25585 PR tree-optimization/56035
25586 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25587
7e184bd7
UB		 255882012-01-24 Uros Bizjak <ubizjak@gmail.com>
25589
25590 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25591 (*movtf_internal_rex64): Add (!o,C) alternative
25592 (*movxf_internal_rex64): Ditto.
25593 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25594
f5ba49ac
SM		 255952013-01-24 Shenghou Ma <minux.ma@gmail.com>
25596
25597 * doc/invoke.texi: fix typo.
25598 * doc/objc.texi: fix typo.
25599
a6343728
RS		 256002013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25601
25602 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25603 for the first two alternatives.
25604
cd030c07
DN		 256052013-01-24 Diego Novillo <dnovillo@google.com>
25606
77dc5297 25607 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 25608 (ggc-zone.o): Remove.
25609 * configure.ac: Remove option --with-gc.
25610 * configure: Re-generate.
25611 * doc/install.texi: Remove documentation for --with-gc.
25612 * gengtype.c (write_enum_defn): Remove. Update all users.
25613 (write_Types_process_field): Remove generation of gt_e_* argument.
25614 (output_type_enum): Remove. Update all users.
25615 (write_enum_defn): Remove. Update all users.
25616 (enum alloc_zone): Remove. Update all users.
77dc5297 25617 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 25618 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25619 Update all callers.
25620 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25621 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 25622 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25623 Update all users.
aaf1e810 25624 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 25625 (struct alloc_zone): Remove.
25626 (ggc_internal_alloc_zone_stat): Remove.
25627 (ggc_internal_cleared_alloc_zone_stat): Remove.
25628 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 25629 (ggc_pch_count_object): Remove last argument. Update all users.
25630 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 25631 (struct alloc_zone): Remove.
25632 * ggc-zone.c: Remove.
77dc5297 25633 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 25634 (struct alloc_zone): Remove.
25635 (ggc_alloc_typed_stat): Remove.
25636 (ggc_alloc_typed): Remove.
25637 (ggc_splay_alloc): Remove first argument.
25638 (rtl_zone): Remove. Update all users.
25639 (tree_zone): Remove. Update all users.
25640 (tree_id_zone): Remove. Update all users.
25641 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25642 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25643 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25644 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25645
a861ffa4
GJL		 256462013-01-24 Georg-Johann Lay <avr@gjlay.de>
25647
25648 * config/avr/avr.c (avr_out_fract): Make register numbers that
25649 might be outside of source operand signed.
25650
593c0ddd
UB		 256512013-01-24 Uros Bizjak <ubizjak@gmail.com>
25652
25653 * config/i386/constraints.md (Yf): New constraint.
25654 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25655 of f constraint to conditionaly disable x87 register preferences.
25656 (*movdf_internal): Ditto.
25657 (*movsf_internal): Ditto.
25658
e86c0101
SB		 256592013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25660
25661 PR inline-asm/55934
25662 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25663 that have operands with impossible constraints.
25664 Add a FIXME for a speed-up opportunity.
25665 * lra-constraints.c (process_alt_operands): Verify that a class
25666 selected from constraints on asms is valid for the operand mode.
25667 (curr_insn_transform): Remove incorrect comment.
25668
f6fee35f
DE		 256692013-01-23 David Edelsohn <dje.gcc@gmail.com>
25670
25671 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25672 TOC operand is a valid symbol ref in the constant pool.
25673
256742013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 25675
aaf1e810 25676 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 25677
dc62d7d1
GJL		 256782013-01-23 Georg-Johann Lay <avr@gjlay.de>
25679
25680 PR target/54222
25681 * config/avr/stdfix.h: New file.
25682 * t-avr (stdfix-gcc.h): New rule to build it.
25683 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25684
2660d12d
KS		 256852013-01-23 Kostya Serebryany <kcc@google.com>
25686
77dc5297
UB		 25687 * config/darwin.h: remove dependency on
25688 CoreFoundation (asan on Mac OS).
2660d12d 25689
a70418fc
JJ		 256902013-01-23 Jakub Jelinek <jakub@redhat.com>
25691
25692 PR target/49069
25693 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25694 instead of cmpdi_operand for first comparison operand.
25695 Don't assert that comparison operands aren't both constants.
25696
47876a2a
JW		 256972013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25698
25699 * doc/install.texi (Downloading the Source): Update references to
25700 downloading separate components.
25701
257022013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 25703
25704 * doc/extend.texi (__int128): Improve grammar.
25705
47876a2a 257062013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 25707
25708 PR target/56028
25709 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25710 alternative to (o,r).
25711 (*movdi_internal_rex64): Remove (!o,n) alternative.
25712 (DImode immediate->memory splitter): Remove.
25713 (DImode immediate->memory peephole2): Remove.
25714 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25715 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25716 alternative to (!o,*r).
25717 (*movtf_internal_sse): New pattern.
25718 (*movxf_internal_rex64): New pattern.
25719 (*movxf_internal): Disable for TARGET_64BIT.
25720 (*movdf_internal_rex64): Remove (!o,F) alternative.
25721
3a984f10
JJ		 257222013-01-22 Jakub Jelinek <jakub@redhat.com>
25723
502498d5
JJ		 25724 PR middle-end/56074
25725 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25726 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25727 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25728 stmt locations where LOCATION_LOCUS of the stmt location is
25729 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25730
3a984f10
JJ		 25731 PR target/55686
25732 * config/i386/i386.md (UNSPEC_STOS): New.
25733 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25734 *strsetqi_1): Add UNSPEC_STOS.
25735
fa817f7f
PC		 257362013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25737
25738 PR c++/56067
25739 * doc/invoke.texi: Remove left over -Wsynth example.
25740
8f498c1b
JJ		 257412013-01-21 Jakub Jelinek <jakub@redhat.com>
25742
25743 PR tree-optimization/56051
25744 * fold-const.c (fold_binary_loc): Don't fold
25745 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25746 a narrowing conversion, or widening conversion from signed
25747 to unsigned.
25748
47876a2a 257492013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 25750
25751 PR rtl-optimization/56023
25752 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25753 dependent on debug instruction.
25754
5b9db1bc
MJ		 257552013-01-21 Martin Jambor <mjambor@suse.cz>
25756
25757 PR middle-end/56022
25758 * function.c (allocate_struct_function): Call
25759 invoke_set_current_function_hook earlier.
25760
e8bb7d68
JJ		 257612013-01-21 Jakub Jelinek <jakub@redhat.com>
25762
25763 * reload1.c (init_reload): Only initialize reload_obstack
25764 during the first call.
25765
616a4e32
MP		 257662013-01-21 Marek Polacek <polacek@redhat.com>
25767
25768 * cfgloop.c (verify_loop_structure): Fix up grammar.
25769
4401981b
YHH		 257702013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25771
25772 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25773 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25774
8e87740b
RR		 257752013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25776
25777 PR target/56058
25778 * config/arm/marvell-pj4.md: Update copyright year.
25779 Fix up use of alu to alu_reg and simple_alu_imm.
25780
47876a2a 257812013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 25782
25783 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25784
89d56d79
VM		 257852013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25786
25787 PR target/55433
25788 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 25789 insn for secondary memory move when memory mode should be different.
89d56d79 25790
fe603553
JDA		 257912013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25792
25793 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25794 atomic_storedi_1): New patterns.
25795
01284895
VK		 257962013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25797
25798 btver2 pipeline descriptions.
25799 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25800 descriptions.
25801 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 25802 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 25803 type attributes.
25804 * config/i386/btver2.md: New file describing btver2 pipelines.
25805
5630e3e1
JL		 258062013-01-19 Andrew Pinski <apinski@cavium.com>
25807
25808 PR tree-optimization/52631
25809 * tree-ssa-sccvn (visit_use): Before looking up the original
25810 statement, try looking up the simplified expression.
25811
650ae806
AG		 258122013-01-19 Anthony Green <green@moxielogic.com>
25813
25814 * config/moxie/moxie.c (moxie_expand_prologue): Set
25815 current_function_static_stack_size.
25816
e300ec2d
JJ		 258172013-01-18 Jakub Jelinek <jakub@redhat.com>
25818
25819 PR tree-optimization/56029
25820 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25821 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25822
a3d7af04
SS		 258232013-01-18 Sharad Singhai <singhai@google.com>
25824
25825 PR tree-optimization/55995
25826 * dumpfile.c (dump_loc): Print location only if available.
25827 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25828
66aa7879
VM		 258292013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25830
25831 PR target/55433
25832 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25833 secondary memory move.
25834 (inherit_reload_reg): Use rclass instead of cl for
25835 check_secondary_memory_needed_p.
25836
3f0fee7b
JJ		 258372013-01-18 Jakub Jelinek <jakub@redhat.com>
25838
25839 PR middle-end/56015
25840 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 25841 the case where writing real complex part of target modifies op1.
3f0fee7b 25842
70c67693
JG		 258432013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25844
25845 * config/aarch64/aarch64-simd.md
25846 (aarch64_vcond_internal<mode>): Handle unordered cases.
25847 * config/aarch64/iterators.md (v_cmp_result): New.
25848
df8de9b3
YHH		 258492013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25850 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25851
25852 * config/arm/marvell-pj4.md: New file.
25853 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25854 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25855 (generic_vfp): Likewise.
25856 * config/arm/arm-cores.def: Add marvell-pj4.
25857 * config/arm/arm-tune.md: Regenerate.
25858 * config/arm/arm-tables.opt: Regenerate.
25859 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25860 * doc/invoke.texi: Document marvell-pj4.
25861
be30c356
TB		 258622013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25863
25864 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25865
0bfbca58 258662013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 25867
25868 PR debug/54114
25869 PR debug/54402
25870 PR debug/49888
25871 * var-tracking.c (negative_power_of_two_p): New.
25872 (global_get_addr_cache, local_get_addr_cache): New.
25873 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25874 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25875 heading comment.
25876 (vt_stack_offset_p): Remove.
25877 (vt_canon_true_dep): Always canonicalize loc's address.
25878 (clobber_overlapping_mems): Make sure we have a MEM.
25879 (local_get_addr_clear_given_value): New.
25880 (val_reset): Clear local cached entries.
25881 (compute_bb_dataflow): Create and release the local cache.
25882 Disable duplicate MEMs clobbering.
25883 (emit_notes_in_bb): Clobber MEMs likewise.
25884 (vt_emit_notes): Create and release the local cache.
25885 (vt_initialize, vt_finalize): Create and release the global
25886 cache, respectively.
1f6bc337 25887 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 25888
0bfbca58 258892013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 25890
25891 PR libmudflap/53359
25892 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25893 not found in the symtab.
25894
0bfbca58 258952013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 25896
c350ba53 25897 PR debug/56006
3aa03517
AO		 25898 PR rtl-optimization/55547
25899 PR rtl-optimization/53827
25900 PR debug/53671
25901 PR debug/49888
25902 * alias.c (offset_overlap_p): New, factored out of...
25903 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25904 the conservative special case for symbolic constants. Don't
25905 adjust zero sizes on alignment.
25906
c664546f
JL		 259072013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25908
25909 PR rtl-optimization/52573
25910 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25911 REG_UNUSED for the same register.
25912
1bd3f750
MP		 259132013-01-17 Richard Biener <rguenther@suse.de>
25914 Marek Polacek <polacek@redhat.com>
25915
25916 PR rtl-optimization/55833
25917 * loop-unswitch.c (unswitch_loops): Move loop verification...
25918 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25919 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25920 Set it to true when we're removing a loop from hierarchy tree in
25921 an irreducible region.
25922 (fix_bb_placements): Adjust caller.
25923 (fix_loop_placements): Likewise.
25924
e52a8b71
GJL		 259252013-01-17 Georg-Johann Lay <avr@gjlay.de>
25926
25927 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25928 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25929 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25930 Remove ID. Adjust comments.
25931 * config/avr/avr-c.c (avr_builtin_name): Remove.
25932 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25933 * config/avr/avr.c (avr_tolower): New static function.
25934 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25935 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25936 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25937 default expansion.
25938
8386a7ea
JH		 259392013-01-17 Jan Hubicka <jh@suse.cz>
25940
610fb637 25941 PR tree-optimization/55273
8386a7ea
JH		 25942 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25943
47876a2a 259442013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 25945
25946 PR target/55981
25947 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25948 store through atomic_store<mode>_1.
25949 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25950
8222c37e
MJ		 259512013-01-17 Martin Jambor <mjambor@suse.cz>
25952
25953 PR tree-optimizations/55264
25954 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25955 for virtual methods.
25956 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25957 virtual methods before inlining is over.
25958 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25959 virtual functions.
25960 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25961 non-virtual.
25962
79f01c76
VM		 259632013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25964
25965 PR rtl-optimization/56005
25966 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25967 pending reads for prefetch.
25968
d0b6bb1b
IB		 259692013-01-16 Ian Bolton <ian.bolton@arm.com>
25970
aaf1e810 25971 * config/aarch64/aarch64.md
d0b6bb1b
IB		 25972 (*cstoresi_neg_uxtw): New pattern.
25973 (*cmovsi_insn_uxtw): New pattern.
25974 (*<optab>si3_uxtw): New pattern.
25975 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25976 (*<optab>si3_insn_uxtw): New pattern.
25977 (*bswapsi2_uxtw): New pattern.
25978
cb9cf03b
RB		 259792013-01-16 Richard Biener <rguenther@suse.de>
25980
25981 * tree-inline.c (tree_function_versioning): Remove set but
25982 never used variable.
25983
2cfc56b9
RB		 259842013-01-16 Richard Biener <rguenther@suse.de>
25985
25986 PR tree-optimization/55964
25987 * tree-flow.h (rename_variables_in_loop): Remove.
25988 (rename_variables_in_bb): Likewise.
25989 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25990 (copy_loop_before): Adjust and delete update-ssa status.
25991 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 25992 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 25993 (rename_variables_in_loop): Remove.
25994 (slpeel_update_phis_for_duplicate_loop): Likewise.
25995 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25996 use available cfg machinery instead of duplicating it.
25997 Update PHI nodes and perform poor-mans SSA update here.
25998 (slpeel_tree_peel_loop_to_edge): Adjust.
25999
c25a0c60
RB		 260002013-01-16 Richard Biener <rguenther@suse.de>
26001
26002 PR tree-optimization/54767
26003 PR tree-optimization/53465
26004 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26005 (vrp_visit_phi_node): For PHI arguments coming via backedges
26006 drop all symbolical range information.
26007 (execute_vrp): Compute backedges.
26008
04b535af
RB		 260092013-01-16 Richard Biener <rguenther@suse.de>
26010
26011 * doc/install.texi: Update CLooG and ISL requirements to
26012 0.18.0 and 0.11.1.
26013
8b0a1e0b
CB		 260142013-01-16 Christian Bruel <christian.bruel@st.com>
26015
26016 PR target/55301
26017 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26018 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26019 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26020
260212013-01-16 DJ Delorie <dj@redhat.com>
26022
26023 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26024 (UNSPECV_SP_SWITCH_E): New.
26025 (sp_switch_1): Change to an unspec.
26026 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26027 replace $r15.
26028
260292013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 26030
26031 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26032 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26033 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26034 (expand_mem_thread_fence): Ditto.
26035 (expand_mem_signal_fence): Ditto.
26036 (expand_atomic_load): Ditto.
26037 (expand_atomic_store): Ditto.
26038
0bfbca58 260392013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 26040
26041 PR rtl-optimization/55547
26042 PR rtl-optimization/53827
26043 PR debug/53671
26044 PR debug/49888
26045 * alias.c (memrefs_conflict_p): Set sizes to negative after
26046 AND adjustments.
26047
305e3ac1
JJ		 260482013-01-15 Jakub Jelinek <jakub@redhat.com>
26049
26050 PR target/55940
26051 * function.c (thread_prologue_and_epilogue_insns): Always
26052 add crtl->drap_reg to set_up_by_prologue.set, even if
26053 stack_realign_drap is false.
26054
f78ac4f2
JBG		 260552013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26056
26057 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26058 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26059 *call): Fix indention.
26060
a78a8cc4
TV		 260612013-01-15 Tom de Vries <tom@codesourcery.com>
26062
26063 PR target/55876
26064 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26065 Update comment.
26066
0e80383f
VM		 260672013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26068
305e3ac1 26069 PR rtl-optimization/55153
0e80383f
VM		 26070 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26071
207b5956
MJ		 260722013-01-15 Martin Jambor <mjambor@suse.cz>
26073
26074 PR tree-optimization/55920
26075 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26076 accesses as grp_to_be_debug_replaced.
26077
a7818b54
JJ		 260782013-01-15 Jakub Jelinek <jakub@redhat.com>
26079
26080 PR tree-optimization/55920
26081 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26082 there is non-useless type conversion needed from debug rhs to lhs,
26083 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26084
b0fe107e
JM		 260852013-01-15 Joseph Myers <joseph@codesourcery.com>
26086 Mikael Pettersson <mikpe@it.uu.se>
26087
26088 PR target/43961
26089 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26090 Thumb.
26091 (ASM_OUTPUT_CASE_LABEL): Remove.
26092 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26093 * final.c (shorten_branches): Update alignment of labels before
26094 jump tables if CASE_VECTOR_SHORTEN_MODE.
26095
34ab62ee
RB		 260962013-01-15 Richard Biener <rguenther@suse.de>
26097
26098 PR bootstrap/55961
26099 * system.h: Do not include gmp.h for building host tools.
26100
783a3a05
RB		 261012013-01-15 Richard Biener <rguenther@suse.de>
26102
26103 PR middle-end/55882
26104 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26105 account for bitpos when computing alignment.
26106
3a579e09
VY		 261072013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26108
26109 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26110 (ix86_target_macros_internal): Likewise.
26111
26112 * config/i386/i386.c (m_CORE2I7): Removed.
26113 (m_CORE_HASWELL): New macro.
26114 (m_CORE_ALL): Likewise.
26115 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26116 (initial_ix86_arch_features): Likewise.
26117 (processor_target_table): Initializations for Core avx2.
26118 (cpu_names): New names "core-avx2".
26119 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26120 PROCESSOR_CORE_HASWELL.
26121 (ix86_issue_rate): New case.
26122 (ia32_multipass_dfa_lookahead): Likewise.
26123 (ix86_sched_init_global): Likewise.
26124
26125 * config/i386/i386.h (TARGET_HASWELL): New macro.
26126 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26127 (processor_type): New PROCESSOR_HASWELL.
26128
955f5a07
JJ		 261292013-01-15 Jakub Jelinek <jakub@redhat.com>
26130
ff784829
JJ		 26131 PR tree-optimization/55955
26132 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26133 *SHIFT_EXPR and *ROTATE_EXPR codes.
26134
955f5a07
JJ		 26135 PR tree-optimization/48766
26136 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26137 -ftrapv disable -fwrapv.
26138
aeb8b4e9
GJL		 261392013-01-14 Georg-Johann Lay <avr@gjlay.de>
26140
26141 PR target/55974
26142 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26143 etc. to 1 and not to __flash.
26144 Use LL suffix for __INT24_MAX__ with -mint8.
26145 Use ULL suffix for __UINT24_MAX__ with -mint8.
26146
1c494c6a
GJL		 261472013-01-14 Georg-Johann Lay <avr@gjlay.de>
26148
26149 * config/avr/avr-arch.h
26150 (struct base_arch_s): Use typedef avr_arch_t instead.
26151 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26152 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26153 * config/avr/avr.c: Same.
26154 * config/avr/avr-devices.c: Same.
26155 * config/avr/driver-avr.c: Same.
26156 * config/avr/gen-avr-mmcu-texi.c: Same.
26157 * config/avr/avr-mcus.def: Adjust comment.
26158
a50344cb
TB		 261592013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26160
88e784e6
UB		 26161 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26162 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 26163
47876a2a 261642013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 26165 Andi Kleen <ak@linux.intel.com>
26166
26167 PR target/55948
26168 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26169 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26170 memmodel flag.
26171
00892272
GJL		 261722013-01-14 Georg-Johann Lay <avr@gjlay.de>
26173
26174 * config/avr/avr-stdint.h: Remove trailing blanks.
26175 * config/avr/avr-log.h: Same.
26176 * config/avr/avr-arch.h: Same.
26177 * config/avr/avr-devices.c: Same.
26178 * config/avr/avr-dimode.md: Same.
26179 * config/avr/predicates.md: Same.
26180 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 26181
00892272
GJL		 26182 * config/avr/avr-protos.h: Same. And:
26183 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26184 (init_cumulative_args): Rename to avr_init_cumulative_args.
26185 (expand_prologue): Rename to avr_expand_prologue.
26186 (expand_epilogue): Rename to avr_expand_epilogue.
26187 (adjust_insn_length): Rename to avr_adjust_insn_length.
26188 (notice_update_cc): Rename to avr_notice_update_cc.
26189 (final_prescan_insn): Rename to avr_final_prescan_insn.
26190 * config/avr/avr.c: Same.
26191 * config/avr/avr.h: Same.
26192 * config/avr/avr.md: Remove trailing blanks.
26193 (prologue): Use avr_expand_prologue.
26194 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26195
7bb01996
RB		 261962013-01-14 Richard Biener <rguenther@suse.de>
26197
26198 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26199 verify_location, collect_subblocks): New functions.
26200 (verify_gimple_in_cfg): Verify that locations only reference
26201 BLOCKs in the functions BLOCK tree.
26202
2724573f
RB		 262032013-01-14 Richard Biener <rguenther@suse.de>
26204
26205 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26206 PHI argument.
26207 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26208 unshare reference.
26209 (insert_out_of_ssa_copy_on_edge): Likewise.
26210 (rewrite_close_phi_out_of_ssa): Likewise.
26211 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26212 debug expressions.
26213 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26214 propagated constants.
26215 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26216 can not be shared.
26217
9a0bbab6
GJL		 262182013-01-14 Georg-Johann Lay <avr@gjlay.de>
26219
26220 * config/avr/avr-modes.def: Add GPL copyright notice.
26221
45805f17
UB		 262222013-01-13 Uros Bizjak <ubizjak@gmail.com>
26223
26224 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26225 MEMMODEL_MASK to determine memory model.
26226 (atomic_store<mode>): Ditto from operands[2].
26227 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26228
9d60be38
JJ		 262292013-01-13 Jakub Jelinek <jakub@redhat.com>
26230
26231 PR fortran/55935
45805f17 26232 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 26233 (fold_gimple_assign): Don't call unshare_expr here.
26234 (fold_ctor_reference): Call unshare_expr.
26235
e7f49d92
TG		 262362013-01-13 Terry Guo <terry.guo@arm.com>
26237
aaf1e810
EB		 26238 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26239 * doc/fragments.texi: Document MULTILIB_REUSE.
26240 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 26241 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 26242 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 26243 (tmpmultilib4): Ditto.
26244 (multilib_reuse): New multilib argument.
26245
fbd03a27
RS		 262462013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26247
26248 * Makefile.in: Update copyright.
26249
b3681f13
TV		 262502013-01-12 Tom de Vries <tom@codesourcery.com>
26251
26252 PR middle-end/55890
26253 * calls.c (expand_call): Check if arg_nr is valid.
26254
3f287e4b
MM		 262552013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26256
26257 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26258 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 26259 documentation. Add missing '__' in front of
26260 __builtin_ia32_packssdw256.
3f287e4b 26261
1abcd5eb
AK		 262622013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26263
26264 PR target/55719
26265 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26266 NO_REGS for larl operands.
26267 (s390_reload_larl_operand): Use s390_load_address instead of
26268 emit_move_insn.
26269
980d0812
RB		 262702013-01-11 Richard Biener <rguenther@suse.de>
26271
26272 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26273 (verify_node_sharing): ... here.
26274 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26275
7cb9fd07
EB		 262762013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26277
26278 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26279 Substitute TREECHECKING.
26280 * configure: Regenerate.
26281 * Makefile.in (TREECHECKING): New.
26282
47876a2a 262832013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 26284
26285 PR tree-optimization/44061
26286 * tree-vrp.c (extract_range_basic): Compute zero as
26287 value-range for __builtin_constant_p of function parameters.
26288
d1e082c2
RS		 262892013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26290
45805f17 26291 Update copyright years.
d1e082c2 26292
f85021b0
VM		 262932013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26294
9d60be38 26295 PR rtl-optimization/55672
0160442c
VM		 26296 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26297 const to be eliminable.
f85021b0 26298
7a8b1ec4
DE		 262992013-01-10 David Edelsohn <dje.gcc@gmail.com>
26300
26301 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26302 * configure: Regenerate.
26303
ad2e5b71
RB		 263042013-01-10 Richard Biener <rguenther@suse.de>
26305
ddf9322d 26306 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 26307
ffc5b2bb
RB		 263082013-01-10 Richard Biener <rguenther@suse.de>
26309
26310 PR bootstrap/55792
26311 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26312 locations for virtual PHI arguments.
26313 (rewrite_update_phi_arguments): Likewise.
26314
e1f674e4
JS		 263152013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26316
26317 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26318 on to assembler.
26319
a57fc743
JJ		 263202013-01-10 Jakub Jelinek <jakub@redhat.com>
26321
26322 PR tree-optimization/55921
26323 * tree-complex.c (expand_complex_asm): New function.
26324 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26325
0ff4390d
AK		 263262013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26327
26328 PR target/55718
26329 * config/s390/s390.c (s390_symref_operand_p)
26330 (s390_loadrelative_operand_p): Merge the two functions.
26331 (s390_check_qrst_address, print_operand_address): Add parameters
26332 to s390_loadrelative_operand_p invokation.
26333 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26334 (s390_reload_larl_operand, s390_secondary_reload): Use
26335 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26336 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26337
6f557e0e
MS		 263382013-01-09 Mike Stump <mikestump@comcast.net>
26339
26340 * dse.c (record_store): Remove unnecessary assert.
26341
7770c9e9
JH		 263422013-01-09 Jan Hubicka <jh@suse.cz>
26343
26344 PR tree-optimization/55569
26345 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26346 * cfgloop.h (scale_loop_profile): Likewise.
26347
a19b1432
JH		 263482013-01-09 Jan Hubicka <jh@suse.cz>
26349
26350 PR lto/45375
ddf9322d
UB		 26351 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26352 functions.
a19b1432
JH		 26353 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26354
9a002da8
RS		 263552013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26356
26357 PR middle-end/55114
26358 * expr.h (maybe_emit_group_store): Declare.
26359 * expr.c (maybe_emit_group_store): New function.
26360 * builtins.c (expand_builtin_int_roundingfn): Call it.
26361 (expand_builtin_int_roundingfn_2): Likewise.
26362
511dcace
VM		 263632013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26364
e1f2b729 26365 PR rtl-optimization/55829
511dcace
VM		 26366 * lra-constraints.c (match_reload): Add code for absent output.
26367 (curr_insn_transform): Add code for reloads of matched inputs
26368 without output.
26369
7b0fe4f4
UB		 263702013-01-09 Uros Bizjak <ubizjak@gmail.com>
26371
26372 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26373 attribute of movddup insn to DF.
26374 (*vec_interleave_lowv2df): Ditto.
26375 (vec_dupv2df): Ditto.
26376
870ca331
JH		 263772013-01-09 Jan Hubicka <jh@suse.cz>
26378
26379 PR tree-optimiation/55875
26380 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26381 EVERY_ITERATION parameter.
7b0fe4f4 26382 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 26383 (idx_infer_loop_bounds): Similarly here.
26384 (n_of_executions_at_most): Simplify
26385 to only test for cases where statement is dominated by the
7b0fe4f4 26386 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 26387 (scev_probably_wraps_p): Use max loop iterations info
26388 as a global bound first.
26389
6f575fe4 263902013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 26391 Nick Clifton <nickc@redhat.com>
26392
26393 * config/v850/v850.md (cbranchsf4): New pattern.
26394 (cstoresf4): New pattern.
26395 (cbranchdf4): New pattern.
26396 (cstoredf4): New pattern.
26397 (movsicc): Disallow floating point comparisons.
26398 (cmpsf_le_insn): Fix order of operators.
26399 (cmpsf_lt_insn): Likewise.
26400 (cmpsf_eq_insn): Likewise.
26401 (cmpdf_le_insn): Likewise.
26402 (cmpdf_lt_insn): Likewise.
26403 (cmpdf_eq_insn): Likewise.
26404 (cmpsf_ge_insn): Use LE comparison.
26405 (cmpdf_ge_insn): Likewise.
26406 (cmpsf_gt_insn): Use LT comparison.
26407 (cmpdf_gt_insn): Likewise.
26408 (cmpsf_ne_insn): Delete pattern.
26409 (cmpdf_ne_insn): Delete pattern.
26410 * config/v850/v850.c (v850_gen_float_compare): Use
26411 gen_cmpdf_eq_insn for NE comparison.
26412 (v850_float_z_comparison_operator)
26413 (v850_float_nz_comparison_operator): Move from here ...
26414 * config/v850/predicates.md: ... to here. Move GT and GE
26415 comparisons into v850_float_z_comparison_operator.
26416 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26417 Delete prototype.
26418 (v850_float_nz_comparison_operator): Likewise.
26419
f0d54148
JDA		 264202013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26421
26422 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26423 with calls to gen_insvsi/gen_insvdi.
26424
8f01beca
VK		 264252013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26426
26427 * config/i386/i386.c (initial_ix86_tune_features): Set up
26428 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26429
2aa579ad
JJ		 264302013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26431 Jakub Jelinek <jakub@redhat.com>
26432
26433 PR tree-optimization/48189
26434 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26435 If nitercst is 0, don't predict the exit edge.
26436
6edc3e32 264372013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 26438
26439 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26440 in asm_fprintf with reg_names.
26441 (aarch64_print_operand_address): Likewise.
26442 (aarch64_return_addr): Likewise.
26443 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26444
f6f94d94
JDA		 264452013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26446
26447 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26448 (INT_U6_BITS): Likewise.
26449 * config/pa/predicates.md (uint6_operand): New predicate.
26450 (shift5_operand, shift6_operand): Likewise.
26451 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26452 arith32_operand.
26453 (lshrdi3): Use shift6_operand.
26454 (shrpsi4, shrpdi4): New insn patterns.
26455 (extzv): Delete expander.
26456 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26457 predicates in unamed zero extract patterns. Tighten common constraint.
26458 (extv): Delete expander.
26459 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26460 predicates in unamed sign extract patterns. Tighten common constraint.
26461 (insv): Delete expander.
26462 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26463 predicates in unamed insert patterns. Tighten common constraint.
26464 Change uint32_operand predicate to uint6_operand predicate in unamed
26465 DImode pattern to insert constant values of type 1...1xxxx.
26466
36b72910
JH		 264672013-01-04 Jan Hubicka <jh@suse.cz>
26468
26469 PR tree-optimization/55823
7b0fe4f4
UB		 26470 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26471 issue.
36b72910 26472
47876a2a 264732013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 26474 Uros Bizjak <ubizjak@gmail.com>
26475
26476 PR rtl-optimization/55845
26477 * df-problems.c (can_move_insns_across): Stop scanning at
26478 volatile_insn_p source instruction or give up if
26479 across_from .. across_to range contains any volatile_insn_p
26480 instructions.
26481
4369c11e
TB		 264822013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26483
7b0fe4f4
UB		 26484 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26485 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26486 Declare.
4369c11e 26487 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26488 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26489
aeb7e7c1
JJ		 264902013-01-08 Jakub Jelinek <jakub@redhat.com>
26491
26492 PR fortran/55341
26493 * asan.c (asan_clear_shadow): New function.
26494 (asan_emit_stack_protection): Use it.
26495
a02ad1aa
TB		 264962013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26497
26498 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26499 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26500 with tab instead of space.
26501
f90d87f5
NC		 265022013-01-08 Nick Clifton <nickc@redhat.com>
26503
26504 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26505 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 26506 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26507 MDBH registers.
f90d87f5 26508
385eb93d
JG		 265092013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26510
26511 * config/aarch64/aarch64-simd.md
26512 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26513 (aarch64_simd_bsl): Likewise.
26514 (aarch64_vcond_internal<mode>): Likewise.
26515 (vcond<mode><mode>): Likewise.
26516 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26517 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26518
4dcd1054
JG		 265192013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26520
26521 * config/aarch64/aarch64-builtins.c
26522 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26523
4267a4a6
MJ		 265242013-01-08 Martin Jambor <mjambor@suse.cz>
26525
26526 PR debug/55579
26527 * tree-sra.c (analyze_access_subtree): Return true also after
26528 potentially creating a debug-only replacement.
26529
5f4e6de3
JJ		 265302013-01-08 Jakub Jelinek <jakub@redhat.com>
26531
3138f224
JJ		 26532 PR middle-end/55890
26533 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26534
5f4e6de3
JJ		 26535 PR tree-optimization/54120
26536 * tree-vrp.c (range_fits_type_p): Don't allow
26537 src_precision < precision from signed vr to unsigned_p
26538 if vr->min or vr->max is negative.
26539 (simplify_float_conversion_using_ranges): Test can_float_p
26540 against CODE_FOR_nothing.
26541
f3ef18ff
JJ		 265422013-01-08 Jakub Jelinek <jakub@redhat.com>
26543 Richard Biener <rguenther@suse.de>
26544
26545 PR middle-end/55851
26546 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26547 types instead of just INTEGER_TYPE types.
26548
25c210f9
MK		 265492013-01-07 Mark Kettenis <kettenis@openbsd.org>
26550
26551 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26552 TF_SIZE): Define.
f90d87f5 26553
81826a7b
SE		 265542013-01-07 Steve Ellcey <sellcey@mips.com>
26555
26556 PR target/42661
26557 * config/mips/mips.opt: Change mad to mmad to match documentation.
26558
35678514
GJL		 265592013-01-07 Georg-Johann Lay <avr@gjlay.de>
26560
26561 PR target/55897
26562 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26563 .progmemx.data now.
26564
e5669488
GJL		 265652013-01-07 Georg-Johann Lay <avr@gjlay.de>
26566
26567 PR target/55897
26568 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26569 (avr_addrspace_t): Add .section_name field.
26570 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26571 array size.
26572 (avr_addrspace): Same. Initialize .section_name. Remove last
26573 NULL entry. Put __memx into .progmemx.data.
26574 (progmem_section_prefix): Remove.
26575 (avr_asm_init_sections): No need to initialize progmem_section.
26576 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26577 section name prefix.
26578 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26579 retrieve the progmem section.
26580 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26581 boundary to run over avr_addrspace[].
26582 (avr_register_target_pragmas): Ditto.
26583
656e6f37
JJ		 265842013-01-06 Jakub Jelinek <jakub@redhat.com>
26585
26586 * varasm.c (output_constant_def_contents): For asan_protect_global
26587 protected strings, adjust DECL_ALIGN if needed, before testing for
26588 anchored symbols.
26589 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26590 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26591 normal decls.
26592 (output_object_block): For asan protected decls, emit asan padding
26593 after their contents.
26594 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26595 (asan_finish_file): Test it here instead.
26596
6edc3e32
UB		 265972013-01-07 Nick Clifton <nickc@redhat.com>
26598 Matthias Klose <doko@debian.org>
26599 Doug Kwan <dougkwan@google.com>
26600 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 26601
26602 PR driver/55470
26603 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26604
26605 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26606
26607 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26608
7b0fe4f4 26609 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 26610
26611 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26612
2a095093
GJL		 266132013-01-07 Georg-Johann Lay <avr@gjlay.de>
26614
f3b503f4 26615 PR target/54461
2a095093
GJL		 26616 * doc/install.texi (Cross-Compiler-Specific Options): Document
26617 --with-avrlibc.
26618
383f9b34
TB		 266192013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26620
26621 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26622 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26623 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26624 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26625 vqmovun_high_s64): Fix source operand number and update copyright.
26626
3626621a
RB		 266272013-01-07 Richard Biener <rguenther@suse.de>
26628
26629 PR middle-end/55890
26630 * gimple.h (gimple_call_builtin_p): New overload.
26631 * gimple.c (validate_call): New function.
26632 (gimple_call_builtin_p): Likewise.
26633 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26634 Use gimple_call_builtin_p.
26635 (find_func_clobbers): Likewise.
26636 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26637 (strlen_optimize_stmt): Likewise.
26638
8b2ea410
JG		 266392013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26640
26641 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26642 (vld1q_dup_*): Likewise.
26643 (vld1_*): Likewise.
26644 (vld1q_*): Likewise.
26645 (vld1_lane_*): Likewise.
26646 (vld1q_lane_*): Likewise.
26647
9713d329
RB		 266482013-01-07 Richard Biener <rguenther@suse.de>
26649
26650 * lto-streamer.h (LTO_minor_version): Bump to 2.
26651
3520f7cc
JG		 266522013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26653
26654 * config/aarch64/aarch64-protos.h
26655 (aarch64_const_double_zero_rtx_p): Rename to...
26656 (aarch64_float_const_zero_rtx_p): ...this.
26657 (aarch64_float_const_representable_p): New.
26658 (aarch64_output_simd_mov_immediate): Likewise.
26659 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26660 move immediate case.
26661 * config/aarch64/aarch64.c
26662 (aarch64_const_double_zero_rtx_p): Rename to...
26663 (aarch64_float_const_zero_rtx_p): ...this.
26664 (aarch64_print_operand): Allow printing of new constants.
26665 (aarch64_valid_floating_const): New.
26666 (aarch64_legitimate_constant_p): Check for valid floating-point
26667 constants.
26668 (aarch64_simd_valid_immediate): Likewise.
26669 (aarch64_vect_float_const_representable_p): New.
26670 (aarch64_float_const_representable_p): Likewise.
26671 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26672 (aarch64_output_simd_mov_immediate): New.
26673 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26674 (*movdf_aarch64): Likewise.
26675 * config/aarch64/constraints.md (Ufc): New.
26676 (Y): call aarch64_float_const_zero_rtx.
26677 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26678
e167c04d
RB		 266792013-01-07 Richard Biener <rguenther@suse.de>
26680
26681 PR tree-optimization/55888
26682 PR tree-optimization/55862
26683 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26684 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26685 not if it is contained therein.
26686
0139aaab
GJL		 266872013-01-07 Georg-Johann Lay <avr@gjlay.de>
26688
26689 * config/avr/t-avr: Typo.
26690
4a176b23
GJL		 266912013-01-07 Georg-Johann Lay <avr@gjlay.de>
26692
26693 PR55243
26694 * config/avr/t-avr: Don't automatically rebuild
26695 $(srcdir)/config/avr/t-multilib
26696 $(srcdir)/config/avr/avr-tables.opt
26697 $(srcdir)/doc/avr-mmcu.texi
26698 (avr-mcus): New phony target to build them on request.
26699 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26700 * avr/avr-mcus.def: Adjust comments.
26701
c7afdc98
UB		 267022013-01-07 Uros Bizjak <ubizjak@gmail.com>
26703
26704 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26705
1ab05c31
RS		 267062013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26707
26708 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26709
488e3acc
RS		 267102013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26711
26712 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26713
a89599d2
DE		 267142013-01-05 David Edelsohn <dje.gcc@gmail.com>
26715
26716 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26717 to generate profiling.
26718 * config/rs6000/aix64.h (LIB_SPEC): Same.
26719
70f09188
AP		 267202013-01-04 Andrew Pinski <apinski@cavium.com>
26721
26722 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26723 New function.
26724 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26725
918d445f
UB		 267262013-01-04 Uros Bizjak <ubizjak@gmail.com>
26727
26728 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26729 unconditionally.
26730 (ix86_expand_move): Ditto.
26731 (ix86_zero_extend_to_Pmode): Ditto.
26732 (ix86_expand_call): Ditto.
26733 (ix86_expand_special_args_builtin): Ditto.
26734 (ix86_expand_builtin): Ditto.
26735
361618ec
RB		 267362013-01-04 Richard Biener <rguenther@suse.de>
26737
26738 PR tree-optimization/55862
26739 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26740 translating them through PHI nodes.
26741
87eab554
MJ		 267422013-01-04 Martin Jambor <mjambor@suse.cz>
26743
26744 PR tree-optimization/55755
26745 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26746 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26747
33879b9f
MP		 267482013-01-04 Marek Polacek <polacek@redhat.com>
26749
26750 PR middle-end/55859
26751 * opts.c (default_options_optimization): Clarify error message.
26752
3068819a
RB		 267532013-01-04 Richard Biener <rguenther@suse.de>
26754
26755 PR middle-end/55863
26756 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26757 reassociation.
26758
0e4ae794
JDA		 267592013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26760
26761 PR target/53789
26762 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 26763 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26764 references.
0e4ae794 26765
65c88cef
RH		 267662013-01-03 Richard Henderson <rth@redhat.com>
26767
26768 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26769 after eliminating TLS symbols.
26770
8ac16127
MG		 267712013-01-03 Marc Glisse <marc.glisse@inria.fr>
26772
26773 PR bootstrap/50167
26774 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26775 * graphite-poly.c (debug_gmp_value): Likewise.
26776
bb664f09
UB		 267772013-01-03 Uros Bizjak <ubizjak@gmail.com>
26778
26779 PR target/55712
26780 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26781 selected code model, define __code_mode_small__, __code_model_medium__,
26782 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26783 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26784 xchg temporary register with %k. Declare temporary register as
26785 early clobbered.
26786 [__x86_64__]: For medium and large code models, preserve %rbx register.
26787
bcf1ef00
RB		 267882013-01-03 Richard Biener <rguenther@suse.de>
26789
0506634a 26790 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 26791 (dump_subscript): Adjust.
26792 (finalize_ddr_dependent): Do not dump redundant info.
26793 (analyze_siv_subscript): Adjust.
26794 (subscript_dependence_tester): Likewise.
26795 (compute_affine_dependence): Likewise.
26796
59fd17e3
RB		 267972013-01-03 Richard Biener <rguenther@suse.de>
26798
26799 Revert
26800 2013-01-03 Richard Biener <rguenther@suse.de>
26801
26802 PR tree-optimization/55857
26803 * tree-vect-stmts.c (vectorizable_load): Do not setup
26804 re-alignment for invariant loads.
26805
26806 2013-01-02 Richard Biener <rguenther@suse.de>
26807
26808 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26809 invariant load do not generate a vector load from the scalar location.
59fd17e3 26810
595c2679
RB		 268112013-01-03 Richard Biener <rguenther@suse.de>
26812
26813 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26814 for not vectorizing.
26815 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26816 not build INDIRECT_REFs, call get_name once only.
26817 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26818 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26819
90bb7d7a
RB		 268202013-01-03 Richard Biener <rguenther@suse.de>
26821
26822 PR tree-optimization/55857
26823 * tree-vect-stmts.c (vectorizable_load): Do not setup
26824 re-alignment for invariant loads.
26825
f09b77ca
RB		 268262013-01-03 Richard Biener <rguenther@suse.de>
26827
26828 PR lto/55848
26829 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26830 prefer a built-in decl.
26831
3461a16e
JJ		 268322013-01-03 Jakub Jelinek <jakub@redhat.com>
26833
df8e2b4f
JJ		 26834 * gcc.c (process_command): Update copyright notice dates.
26835 * gcov.c (print_version): Likewise.
26836 * gcov-dump.c (print_version): Likewise.
26837
3461a16e
JJ		 26838 PR rtl-optimization/55838
26839 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26840 iv0.step, iv1.step and step.
26841
8b5546d6
JJ		 268422013-01-03 Jakub Jelinek <jakub@redhat.com>
26843 Marc Glisse <marc.glisse@inria.fr>
26844
26845 PR tree-optimization/55832
26846 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26847 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26848 integer_{one,zero}_node.
26849
8ab1d2e9
JJ		 268502013-01-03 Jakub Jelinek <jakub@redhat.com>
26851
26852 PR debug/54402
26853 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26854 * var-tracking.c (reverse_op): Don't add reverse ops to
26855 VALUEs that have already
0506634a 26856 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 26857
5b9ad1d4
GP		 268582013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26859
26860 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26861
e25a6711
TJ		 268622013-01-02 Teresa Johnson <tejohnson@google.com>
26863
26864 * dumpfile.c (dump_loc): Print filename with location.
26865 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26866 new location_t parameter to emit complete unroll message with
26867 new dump framework.
26868 (canonicalize_loop_induction_variables): Compute loops location
26869 and pass to try_unroll_loop_completely.
26870 * loop-unroll.c (report_unroll_peel): New function.
26871 (peel_loops_completely): Use new dump format with location
26872 for main dumpfile message, and invoke report_unroll_peel on success.
26873 (decide_unrolling_and_peeling): Ditto.
26874 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26875 by report_unroll_peel.
26876 (decide_peel_completely): Ditto.
26877 (decide_unroll_constant_iterations): Ditto.
26878 (decide_unroll_runtime_iterations): Ditto.
26879 (decide_peel_simple): Ditto.
26880 (decide_unroll_stupid): Ditto.
26881 * cfgloop.c (get_loop_location): New function.
26882 * cfgloop.h (get_loop_location): Declare.
26883
77878621
ST		 268842013-01-02 Sriraman Tallam <tmsriram@google.com>
26885
26886 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26887 NULL.
26888
9e65d03e
JDA		 268892013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26890
26891 PR middle-end/55198
26892 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26893 BLKmode objects when EXPAND_MEMORY is specified.
26894
6a7da30f
ST		 268952013-01-02 Sriraman Tallam <tmsriram@google.com>
26896
26897 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26898 in loop predicate.
26899 (fold_builtin_cpu): Do not share cpu model decls across statements.
26900
e78167f2
JM		 269012013-01-02 Jason Merrill <jason@redhat.com>
26902
26903 PR c++/55804
26904 * tree.c (build_array_type_1): Revert earlier change.
26905
8c075fb4
YZ		 269062013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26907
26908 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26909 "cortex-a57".
26910 * config/aarch64/aarch64-tune.md: Re-generate.
26911
0682ed3e
RB		 269122013-01-02 Richard Biener <rguenther@suse.de>
26913
26914 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26915 invariant load do not generate a vector load from the scalar location.
0682ed3e 26916
b8f6a302
RB		 269172013-01-02 Richard Biener <rguenther@suse.de>
26918
26919 PR bootstrap/55784
26920 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26921 * configure: Regenerate.
26922
04b80dbb
RS		 269232013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26924
26925 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26926 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26927 (expand_builtin_int_roundingfn_2): Keep the original target around
26928 for the fallback case.
26929
635b0b0c
RS		 269302013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26931
26932 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26933 to be clear for sign changes.
26934
42d57399
JH		 269352013-01-01 Jan Hubicka <jh@suse.cz>
26936
26937 * ipa-inline-analysis.c: Fix formatting.
26938
5bb6669d
JJ		 269392013-01-01 Jakub Jelinek <jakub@redhat.com>
26940
26941 PR tree-optimization/55831
26942 * tree-vect-loop.c (get_initial_def_for_induction): Use
26943 gsi_after_labels instead of gsi_start_bb.
ad41bd84 26944\f
86a2db33 26945Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 26946
26947Copying and distribution of this file, with or without modification,
26948are permitted in any medium without royalty provided the copyright
26949notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git