]> 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
i386.c (ix86_fixup_binary_operands): When both source operands are in memory...
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
6f1abb55
UB		 12013-10-17 Uros Bizjak <ubizjak@gmail.com>
2
3 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
4 operands are in memory, prefer to force non-matched operand 1 to
5 the register.
6
1ddbbbc0
MM		 72013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
8
9 PR target/58673
10 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
11 restrict TImode addresses to single indirect registers if both
12 -mquad-memory and -mvsx-timode are used.
13 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
14 we should emit load/store quad. Remove using %y for quad memory
15 addresses.
16
17 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
18 constraints to allow load/store quad on machines where TImode is
19 not allowed in VSX registers. Use 'n' instead of 'F' constraint
20 for TImode to load integer constants.
21
f541a481
KT		 222013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23
24 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
25
51bb310d
MS		 262013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
27
28 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
29 handling of STACK_REG.
30
f4bed77b
RB		 312013-10-17 Richard Biener <rguenther@suse.de>
32
33 PR tree-optimization/58143
34 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
35 New function.
36 (rewrite_to_defined_overflow): Likewise.
37 (move_computations_dom_walker::before_dom): Rewrite stmts
38 with undefined signed overflow that are not always executed
39 into unsigned arithmetic.
40
d358fb9d
MM		 412013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
42
43 PR target/57756
44 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
45 explicit isa flag to be an options variable, instead of using
46 global_options_set. Remove define from rs6000.h.
47 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
48
49 * config/rs6000/rs6000.c (rs6000_option_override_internal):
50 Initialize rs6000_isa_flags_explicit.
51 (rs6000_function_specific_save): Add gcc_options* parameter, so
52 that the powerpc builds after the 2013-10-15 changes.
53 (rs6000_function_specific_restore): Likewise.
54
c71aed47
DD		 552013-10-16 DJ Delorie <dj@redhat.com>
56
57 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
58 op is a REG before checking REGNO.
59 (rl78_alloc_physical_registers): Verify pattern is a SET before
60 checking SET_SRC.
61
90d3bd51
BS		 622013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
63
64 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
65 endianness.
66 (vec_unpacks_lo_v4sf): Likewise.
67 (vec_unpacks_float_hi_v4si): Likewise.
68 (vec_unpacks_float_lo_v4si): Likewise.
69 (vec_unpacku_float_hi_v4si): Likewise.
70 (vec_unpacku_float_lo_v4si): Likewise.
71
de75c876
BS		 722013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
73
74 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
75 (vsx_concat_v2sf): Likewise.
76
21216998
JG		 772013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
78
79 * config/aarch64/aarch64.md
80 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
81
1ef88893 822013-10-16 Andrew MacLeod <amacleod@redhat.com>
7ab4168e 83
1ef88893 84 PR tree-optimization/58697
7ab4168e 85 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1ef88893
AM		 86 estimated_loop_iterations_int.
87 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7ab4168e 88 (get_max_loop_iterations_int): New. HWINT version of
1ef88893
AM		 89 get_max_loop_iterations.
90 * cfgloop.h: Add prototypes.
91 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
7ab4168e 92 * loop-unroll.c (decide_peel_once_rolling): Call
1ef88893
AM		 93 get_estimated_loop_iterations_int.
94 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
95 * tree-ssa-loop-niter.h: Tweak prototypes.
96
c22df64f
DM		 972013-10-16 David Malcolm <dmalcolm@redhat.com>
98
99 * gengtype-parse.c (struct_field_seq): Ignore access-control
100 keywords ("public:" etc).
101
f24bb080
MS		 1022013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
103
104 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
105 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
106
c281bf26
YR		 1072013-10-16 Yvan Roux <yvan.roux@linaro.org>
108
109 * config/arm/arm.opt (mlra): New option.
110 * config/arm/arm.c (arm_lra_p): New function.
111 (TARGET_LRA_P): Define.
112
5806f481
PM		 1132013-10-16 Paulo Matos <pmatos@broadcom.com>
114
115 * tree-core.h (tree_code_name): Remove.
116 * tree.h (get_tree_code_name): New prototype.
117 * tree.c (tree_code_name): Make static.
118 (get_tree_code_name): New function.
119 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
120 tree_class_check_failed, tree_range_check_failed,
121 tree_not_class_check_failed, omp_clause_check_failed,
122 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
123 wrapper get_tree_code_name instead of calling tree_code_name directly.
124 * tree-vrp.c (dump_asserts_for): Likewise.
125 * tree-dump.c (dequeue_and_dump): Likewise.
126 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
127 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
128 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
129 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
130 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
131 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
132 dump_gimple_omp_for): Likewise.
133 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
134 * tree-ssa-pre.c (print_pre_expr): Likewise.
135 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
136 * print-tree.c (print_node_brief, print_node): Likewise.
137 * gimple.c (gimple_check_failed): Likewise.
138 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
139 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
140 * config/mep/mep.c (mep_validate_vliw): Likewise.
141 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
142 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
143
82c9df56
GG		 1442013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
145
146 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
147 for AMD bdver3.
148
b82d0df9
HPN		 1492013-10-16 Hans-Peter Nilsson <hp@axis.com>
150
151 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
152 (MULTILIB_MATCHES): Add multilib for -march=v8.
153
bf7b5747
ST		 1542013-10-15 Sriraman Tallam <tmsriram@google.com>
155
156 PR target/57756
157 * optc-save-gen.awk: Add extra parameter to the save and restore
158 target calls.
159 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
160 * tree.c (build_optimization_node): New parameter. Add extra parameter
161 to call to cl_optimization_save.
162 (build_target_option_node): New parameter. Add extra parameter
163 to call to cl_target_option_save.
164 * tree.h (build_optimization_node): New parameter.
165 (build_target_option_node): New parameter.
166 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
167 build_optimization_node and build_target_option_node.
168 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
169 (handle_pragma_push_options): Ditto.
170 * toplev.c (process_options): Ditto.
171 * opts.c (init_options_struct): Check for opts_set non-null.
172 * target.def (target_option.save): New parameter.
173 (target_option.restore): New parameter.
174 * tm.texi: Generate.
175 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
176 (ix86_pragma_target_parse): Ditto.
177 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
178 parameters.
179 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
180 to build_optimization_node and build_target_option_node.
181 (rs6000_valid_attribute_p): Ditto.
182 (rs6000_pragma_target_parse): Ditto.
183 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
184 data.
185 * config/i386/i386.h:
186 TARGET_64BIT_P: New Macro
187 TARGET_MMX_P: New Macro.
188 TARGET_3DNOW_P: New Macro.
189 TARGET_3DNOW_A_P: New Macro.
190 TARGET_SSE_P: New Macro.
191 TARGET_SSE2_P: New Macro.
192 TARGET_SSE3_P: New Macro.
193 TARGET_SSSE3_P: New Macro.
194 TARGET_SSE4_1_P: New Macro.
195 TARGET_SSE4_2_P: New Macro.
196 TARGET_AVX_P: New Macro.
197 TARGET_AVX2_P: New Macro.
198 TARGET_AVX512F_P: New Macro.
199 TARGET_AVX512PF_P: New Macro.
200 TARGET_AVX512ER_P: New Macro.
201 TARGET_AVX512CD_P: New Macro.
202 TARGET_FMA_P: New Macro.
203 TARGET_SSE4A_P: New Macro.
204 TARGET_FMA4_P: New Macro.
205 TARGET_XOP_P: New Macro.
206 TARGET_LWP_P: New Macro.
207 TARGET_ABM_P: New Macro.
208 TARGET_BMI_P: New Macro.
209 TARGET_BMI2_P: New Macro.
210 TARGET_LZCNT_P: New Macro.
211 TARGET_TBM_P: New Macro.
212 TARGET_POPCNT_P: New Macro.
213 TARGET_SAHF_P: New Macro.
214 TARGET_MOVBE_P: New Macro.
215 TARGET_CRC32_P: New Macro.
216 TARGET_AES_P: New Macro.
217 TARGET_PCLMUL_P: New Macro.
218 TARGET_CMPXCHG16B_P: New Macro.
219 TARGET_FSGSBASE_P: New Macro.
220 TARGET_RDRND_P: New Macro.
221 TARGET_F16C_P: New Macro.
222 TARGET_RTM_P: New Macro.
223 TARGET_HLE_P: New Macro.
224 TARGET_RDSEED_P: New Macro.
225 TARGET_PRFCHW_P: New Macro.
226 TARGET_ADX_P: New Macro.
227 TARGET_FXSR_P: New Macro.
228 TARGET_XSAVE_P: New Macro.
229 TARGET_XSAVEOPT_P: New Macro.
230 TARGET_LP64_P: New Macro.
231 TARGET_X32_P: New Macro.
232 TARGET_FPMATH_DEFAULT_P: New Macro.
233 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
234 * config/i386/i386.c (ix86_option_override_internal): New parameters.
235 opts and opts_set.
236 Change ix86_tune_string to access opts->x_ix86_tune_string.
237 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
238 Change ix86_arch_string to access opts->x_ix86_arch_string.
239 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
240 Change ix86_pmode to access opts->x_ix86_pmode.
241 Change ix86_abi to access opts->x_ix86_abi.
242 Change ix86_cmodel to access opts->x_ix86_cmodel.
243 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
244 Change ix86_isa_flags_explicit to access
245 opts->x_ix86_isa_flags_explicit.
246 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
247 Change ix86_regparm to access opts->x_ix86_regparm.
248 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
249 Change ix86_preferred_stack_boundary_arg to access
250 opts->x_ix86_preferred_stack_boundary_arg.
251 Change ix86_force_align_arg_pointer to access
252 opts->x_ix86_force_align_arg_pointer.
253 Change ix86_incoming_stack_boundar_arg to access
254 opts->x_ix86_incoming_stack_boundar_arg.
255 Change ix86_fpmath to access opts->x_ix86_fpmath.
256 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
257 Change ix86_recip_name to access opts->x_ix86_recip_name.
258 Change ix86_stack_protector_guard to access
259 opts->x_ix86_stack_protector_guard.
260 Change ix86_tune_memcpy_strategy to access
261 opts->x_ix86_tune_memcpy_strategy.
262 Change ix86_tune_memset_strategy to access
263 opts->x_ix86_tune_memset_strategy.
264 Change global_options to access opts.
265 Change global_options_set to access opts_set.
6f1abb55
UB		 266 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
267 Change TARGET_MMX to TARGET_MMX_P (opts->...).
268 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
269 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
270 Change TARGET_SSE to TARGET_SSE_P (opts->...).
271 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
272 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
273 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
274 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
275 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
276 Change TARGET_AVX to TARGET_AVX_P (opts->...).
277 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
278 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
279 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
280 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
281 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
282 Change TARGET_FMA to TARGET_FMA_P (opts->...).
283 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
284 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
285 Change TARGET_XOP to TARGET_XOP_P (opts->...).
286 Change TARGET_LWP to TARGET_LWP_P (opts->...).
287 Change TARGET_ABM to TARGET_ABM_P (opts->...).
288 Change TARGET_BMI to TARGET_BMI_P (opts->...).
289 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
290 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
291 Change TARGET_TBM to TARGET_TBM_P (opts->...).
292 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
293 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
294 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
295 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
296 Change TARGET_AES to TARGET_AES_P (opts->...).
297 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
298 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
299 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
300 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
301 Change TARGET_F16C to TARGET_F16C_P (opts->...).
302 Change TARGET_RTM to TARGET_RTM_P (opts->...).
303 Change TARGET_HLE to TARGET_HLE_P (opts->...).
304 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
305 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
306 Change TARGET_ADX to TARGET_ADX_P (opts->...).
307 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
308 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
309 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
310 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
311 Change TARGET_X32 to TARGET_X32_P (opts->...).
312 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
313 Change TARGET_FLOAT_RETURNS_IN_80387 to
314 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
bf7b5747
ST		 315 (ix86_function_specific_save): New parameter. Use opts-> fields
316 to replace global fields.
317 (ix86_function_specific_restore): Ditto.
318 (ix86_valid_target_attribute_inner_p): New parameters.
319 Fix recursive call.
320 Fix call to ix86_handle_option and set_option.
321 (ix86_valid_target_attribute_tree): New parameters.
322 Change global_options to access opts.
323 Change global_options_set to access opts_set.
324 Fix call to ix86_valid_target_attribute_inner_p.
325 Change ix86_tune_string to access opts->x_ix86_tune_string.
326 Change ix86_arch_string to access opts->x_ix86_arch_string.
327 Change ix86_fpmath to access opts->x_ix86_fpmath
328 Fix call to ix86_option_override_internal.
329 Fix call to ix86_add_new_builtins.
330 Fix calls to build_optimization_node and build_target_option_node.
331 (ix86_valid_target_attribute_p): Remove access to global_options.
332 Use new gcc_options structure func_options.
333 Fix call to ix86_valid_target_attribute_tree.
334 Fix call to build_optimization_node.
335 (get_builtin_code_for_version): Fix call to
336 ix86_valid_target_attribute_tree.
337
459260ec
DM		 3382013-10-15 David Malcolm <dmalcolm@redhat.com>
339
340 * Makefile.in (PICFLAG): New.
341 (enable_host_shared): New.
342 (INTERNAL_CFLAGS): Use PICFLAG.
343 (LIBIBERTY): Use pic build of libiberty.a if configured with
344 --enable-host-shared.
345 * configure.ac: Add --enable-host-shared, setting up new
346 PICFLAG variable.
347 * configure: Regenerate.
6f1abb55 348 * doc/install.texi (--enable-shared): Add note contrasting it with ...
459260ec
DM		 349 (--enable-host-shared): New option.
350
16022041
RS		 3512013-10-15 Richard Biener <rguenther@suse.de>
352
353 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
354 for built-in functions.
355
b114bfb4
ZC		 3562013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
357
358 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
359 (optimize_range_tests_1): New function,
360 extracted from optimize_range_tests.
361 (optimize_range_tests_xor): Similarly.
362 (optimize_range_tests_diff): New function.
363 (optimize_range_tests): Use optimize_range_tests_1.
364
69b8f2f9
CH		 3652013-10-15 Cong Hou <congh@google.com>
366
7ab4168e
MS		 367 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
368 requirement of the reduction pattern so that one operand of the
69b8f2f9
CH		 369 reduction operation can come from outside of the loop.
370
8ff5a4e4
JG		 3712013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
372
373 * config/arm/neon-schedgen.ml: Remove.
374 * config/arm/cortex-a9-neon.md: Remove comment regarding
375 neon-schedgen.ml.
376
52b69ad2
JG		 3772013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
378
379 * config/arm/types: Remove old neon types.
380
e7c6a556
JG		 3812013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
382
383 * config/arm/cortex-a7.md
384 (cortex_a7_neon_type): New.
385 (cortex_a7_neon_mul): Update for new types.
386 (cortex_a7_neon_mla): Likewise.
387 (cortex_a7_neon): Likewise.
388
795dd8d5
JG		 3892013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
390
391 * config/arm/cortex-a15-neon.md
392 (cortex_a15_neon_type): New,
393
394 (cortex_a15_neon_int_1): Remove.
395 (cortex_a15_neon_int_2): Likewise.
396 (cortex_a15_neon_int_3): Likewise.
397 (cortex_a15_neon_int_4): Likewise.
398 (cortex_a15_neon_int_5): Likewise.
399 (cortex_a15_neon_vqneg_vqabs): Likewise.
400 (cortex_a15_neon_vmov): Likewise.
401 (cortex_a15_neon_vaba): Likewise.
402 (cortex_a15_neon_vaba_qqq): Likewise.
403 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
404 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
405 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
406 Likewise.
407 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
408 (cortex_a15_neon_mla_qqq_8_16): Likewise.
409 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
410 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
411 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
412 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
413 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
414 (cortex_a15_neon_shift_1): Likewise.
415 (cortex_a15_neon_shift_2): Likewise.
416 (cortex_a15_neon_shift_3): Likewise.
417 (cortex_a15_neon_vshl_ddd): Likewise.
418 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
419 (cortex_a15_neon_vsra_vrsra): Likewise.
420 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
421 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
422 (cortex_a15_neon_bp_3cycle): Likewise.
423 (cortex_a15_neon_ldm_2): Likewise.
424 (cortex_a15_neon_stm_2): Likewise.
425 (cortex_a15_neon_mcr): Likewise.
426 (cortex_a15_neon_mrc): Likewise.
427 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
428 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
429 (cortex_a15_neon_fp_vmul_ddd): Likewise.
430 (cortex_a15_neon_fp_vmul_qqd): Likewise.
431 (cortex_a15_neon_fp_vmla_ddd): Likewise.
432 (cortex_a15_neon_fp_vmla_qqq): Likewise.
433 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
434 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
435 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
436 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
437 (cortex_a15_neon_bp_simple): Likewise.
438 (cortex_a15_neon_bp_2cycle): Likewise.
439 (cortex_a15_neon_bp_3cycle): Likewise.
440 (cortex_a15_neon_vld1_1_2_regs): Likewise.
441 (cortex_a15_neon_vld1_3_4_regs): Likewise.
442 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
443 (cortex_a15_neon_vld2_4_regs): Likewise.
444 (cortex_a15_neon_vld3_vld4): Likewise.
445 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
446 (cortex_a15_neon_vst1_3_4_regs): Likewise.
447 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
448 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
449 (cortex_a15_neon_vst3_vst4): Rename to...
450 (cortex_a15_neon_vst4): This, update for new attributes.
451 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
452 (cortex_a15_neon_vld3_vld4_lane): Likewise.
453 (cortex_a15_neon_vst1_vst2_lane): Likewise.
454 (cortex_a15_neon_vst3_vst4_lane): Likewise.
455 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
456 (cortex_a15_neon_ldm_2): Likewise.
457 (cortex_a15_neon_stm_2): Likewise.
458 (cortex_a15_neon_mcr): Likewise.
459 (cortex_a15_neon_mcr_2_mcrr): Likewise.
460 (cortex_a15_neon_mrc): Likewise.
461 (cortex_a15_neon_mrrc): Likewise.
462
463 (cortex_a15_neon_abd): New.
464 (cortex_a15_neon_abd_q): Likewise.
465 (cortex_a15_neon_aba): Likewise.
466 (cortex_a15_neon_aba_q): Likewise.
467 (cortex_a15_neon_acc): Likewise.
468 (cortex_a15_neon_acc_q): Likewise.
469 (cortex_a15_neon_arith_basic): Likewise.
470 (cortex_a15_neon_arith_complex): Likewise.
471 (cortex_a15_neon_multiply): Likewise.
472 (cortex_a15_neon_multiply_q): Likewise.
473 (cortex_a15_neon_mla): Likewise.
474 (cortex_a15_neon_mla_q): Likewise.
475 (cortex_a15_neon_sat_mla_long): Likewise.
476 (cortex_a15_neon_shift_acc): Likewise.
477 (cortex_a15_neon_shift_imm_basic): Likewise.
478 (cortex_a15_neon_shift_imm_complex): Likewise.
479 (cortex_a15_neon_shift_reg_basic): Likewise.
480 (cortex_a15_neon_shift_reg_basic_q): Likewise.
481 (cortex_a15_neon_shift_reg_complex): Likewise.
482 (cortex_a15_neon_shift_reg_complex_q): Likewise.
483 (cortex_a15_neon_fp_negabs): Likewise
484 (cortex_a15_neon_fp_arith): Likewise
485 (cortex_a15_neon_fp_arith_q): Likewise
486 (cortex_a15_neon_fp_cvt_int): Likewise
487 (cortex_a15_neon_fp_cvt_int_q): Likewise
488 (cortex_a15_neon_fp_cvt_16): Likewise
489 (cortex_a15_neon_fp_mul): Likewise
490 (cortex_a15_neon_fp_mul_q): Likewise
491 (cortex_a15_neon_fp_mla): Likewise
492 (cortex_a15_neon_fp_mla_q): Likewise
493 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
494 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
495 (cortex_a15_neon_bitops): Likewise.
496 (cortex_a15_neon_bitops_q): Likewise.
497 (cortex_a15_neon_from_gp): Likewise.
498 (cortex_a15_neon_from_gp_q): Likewise.
499 (cortex_a15_neon_tbl3_tbl4): Likewise.
500 (cortex_a15_neon_zip_q): Likewise.
501 (cortex_a15_neon_to_gp): Likewise.
502 (cortex_a15_neon_load_a): Likewise.
503 (cortex_a15_neon_load_b): Likewise.
504 (cortex_a15_neon_load_c): Likewise.
505 (cortex_a15_neon_load_d): Likewise.
506 (cortex_a15_neon_load_e): Likewise.
507 (cortex_a15_neon_load_f): Likewise.
508 (cortex_a15_neon_store_a): Likewise.
509 (cortex_a15_neon_store_b): Likewise.
510 (cortex_a15_neon_store_c): Likewise.
511 (cortex_a15_neon_store_d): Likewise.
512 (cortex_a15_neon_store_e): Likewise.
513 (cortex_a15_neon_store_f): Likewise.
514 (cortex_a15_neon_store_g): Likewise.
515 (cortex_a15_neon_store_h): Likewise.
516 (cortex_a15_vfp_to_from_gp): Likewise.
517
bc5e57e3
JG		 5182013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
519
520 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
521
522 (cortex_a9_neon_vshl_ddd): Remove.
523 (cortex_a9_neon_vst3_vst4): Likewise.
524 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
525
526 (cortex_a9_neon_bit_ops_q): New.
527
528 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
529 (cortex_a9_neon_int_2): Likewise.
530 (cortex_a9_neon_int_3): Likewise.
531 (cortex_a9_neon_int_4): Likewise.
532 (cortex_a9_neon_int_5): Likewise.
533 (cortex_a9_neon_vqneg_vqabs): Likewise.
534 (cortex_a9_neon_vmov): Likewise.
535 (cortex_a9_neon_vaba): Likewise.
536 (cortex_a9_neon_vaba_qqq): Likewise.
537 (cortex_a9_neon_shift_1): Likewise.
538 (cortex_a9_neon_shift_2): Likewise.
539 (cortex_a9_neon_shift_3): Likewise.
540 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
541 (cortex_a9_neon_vsra_vrsra): Likewise.
542 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
543 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
544 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
545 Likewise.
546 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
547 (cortex_a9_neon_mla_qqq_8_16): Likewise.
548 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
549 Likewise.
550 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
551 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
552 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
553 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
554 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
555 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
556 (cortex_a9_neon_fp_vsum): Likewise.
557 (cortex_a9_neon_fp_vmul_ddd): Likewise.
558 (cortex_a9_neon_fp_vmul_qqd): Likewise.
559 (cortex_a9_neon_fp_vmla_ddd): Likewise.
560 (cortex_a9_neon_fp_vmla_qqq): Likewise.
561 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
562 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
563 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
564 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
565 (cortex_a9_neon_bp_simple): Likewise.
566 (cortex_a9_neon_bp_2cycle): Likewise.
567 (cortex_a9_neon_bp_3cycle): Likewise.
568 (cortex_a9_neon_ldr): Likewise.
569 (cortex_a9_neon_str): Likewise.
570 (cortex_a9_neon_vld1_1_2_regs): Likewise.
571 (cortex_a9_neon_vld1_3_4_regs): Likewise.
572 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
573 (cortex_a9_neon_vld2_4_regs): Likewise.
574 (cortex_a9_neon_vld3_vld4): Likewise.
575 (cortex_a9_neon_vld1_vld2_lane): Likewise.
576 (cortex_a9_neon_vld3_vld4_lane): Likewise.
577 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
578 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
579 (cortex_a9_neon_vst1_3_4_regs): Likewise.
580 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
581 (cortex_a9_neon_vst1_vst2_lane): Likewise.
582 (cortex_a9_neon_vst3_vst4_lane): Likewise.
583 (cortex_a9_neon_mcr): Likewise.
584 (cortex_a9_neon_mcr_2_mcrr): Likewise.
585 (cortex_a9_neon_mrc): Likewise.
586 (cortex_a9_neon_mrrc): Likewise.
587
9deda8cb
JG		 5882013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
589
590 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
591
592 (cortex_a8_neon_vshl_ddd): Remove.
593 (cortex_a8_neon_vst3_vst4): Likewise.
594 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
595
596 (cortex_a8_neon_bit_ops_q): New.
597
598 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
599 (cortex_a8_neon_int_2): Likewise..
600 (cortex_a8_neon_int_3): Likewise.
601 (cortex_a8_neon_int_5): Likewise.
602 (cortex_a8_neon_vqneg_vqabs): Likewise.
603 (cortex_a8_neon_int_4): Likewise.
604 (cortex_a8_neon_vaba): Likewise.
605 (cortex_a8_neon_vaba_qqq): Likewise.
606 (cortex_a8_neon_shift_1): Likewise.
607 (cortex_a8_neon_shift_2): Likewise.
608 (cortex_a8_neon_shift_3): Likewise.
609 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
610 (cortex_a8_neon_vsra_vrsra): Likewise.
611 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
612 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
613 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
614 Likewise.
615 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
616 (cortex_a8_neon_mla_qqq_8_16): Likewise.
617 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
618 Likewise.
619 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
620 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
621 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
622 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
623 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
624 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
625 (cortex_a8_neon_fp_vsum): Likewise.
626 (cortex_a8_neon_fp_vmul_ddd): Likewise.
627 (cortex_a8_neon_fp_vmul_qqd): Likewise.
628 (cortex_a8_neon_fp_vmla_ddd): Likewise.
629 (cortex_a8_neon_fp_vmla_qqq): Likewise.
630 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
631 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
632 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
633 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
634 (cortex_a8_neon_bp_simple): Likewise.
635 (cortex_a8_neon_bp_2cycle): Likewise.
636 (cortex_a8_neon_bp_3cycle): Likewise.
637 (cortex_a8_neon_ldr): Likewise.
638 (cortex_a8_neon_str): Likewise.
639 (cortex_a8_neon_vld1_1_2_regs): Likewise.
640 (cortex_a8_neon_vld1_3_4_regs): Likewise.
641 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
642 (cortex_a8_neon_vld2_4_regs): Likewise.
643 (cortex_a8_neon_vld3_vld4): Likewise.
644 (cortex_a8_neon_vld1_vld2_lane): Likewise.
645 (cortex_a8_neon_vld3_vld4_lane): Likewise.
646 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
647 (cortex_a8_neon_vst1_3_4_regs): Likewise.
648 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
649 (cortex_a8_neon_vst1_vst2_lane): Likewise.
650 (cortex_a8_neon_vst3_vst4_lane): Likewise.
651 (cortex_a8_neon_mcr): Likewise.
652 (cortex_a8_neon_mcr_2_mcrr): Likewise.
653 (cortex_a8_neon_mrc): Likewise.
654 (cortex_a8_neon_mrrc): Likewise.
655
0f686aa9
JG		 6562013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
657
658 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
659 (fp): New.
660 * config/aarch64/aarch64-simd.md (neon_type): Remove.
661 (aarch64_simd_dup<mode>): Add "type" attribute.
662 (aarch64_dup_lane<mode>): Likewise.
663 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
664 (*aarch64_simd_mov<mode>): Likewise.
665 (aarch64_simd_mov_from_<mode>low): Likewise.
666 (aarch64_simd_mov_from_<mode>high): Likewise.
667 (orn<mode>3): Likewise.
668 (bic<mode>3): Likewise.
669 (add<mode>3): Likewise.
670 (sub<mode>3): Likewise.
671 (mul<mode>3): Likewise.
672 (*aarch64_mul3_elt<mode>): Likewise.
673 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
674 (*aarch64_mul3_elt_to_128df): Likewise.
675 (*aarch64_mul3_elt_to_64v2df): Likewise.
676 (neg<mode>2): Likewise.
677 (abs<mode>2): Likewise.
678 (abd<mode>_3): Likewise.
679 (aba<mode>_3): Likewise.
680 (fabd<mode>_3): Likewise.
681 (*fabd_scalar<mode>3): Likewise.
682 (and<mode>3): Likewise.
683 (ior<mode>3): Likewise.
684 (xor<mode>3): Likewise.
685 (one_cmpl<mode>2): Likewise.
686 (aarch64_simd_vec_set<mode>): Likewise.
687 (aarch64_simd_lshr<mode>): Likewise.
688 (aarch64_simd_ashr<mode>): Likewise.
689 (aarch64_simd_imm_shl<mode>): Likewise.
690 (aarch64_simd_reg_sshl<mode): Likewise.
691 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
692 (aarch64_simd_reg_shl<mode>_signed): Likewise.
693 (aarch64_simd_vec_setv2di): Likewise.
694 (aarch64_simd_vec_set<mode>): Likewise.
695 (aarch64_mla<mode>): Likewise.
696 (*aarch64_mla_elt<mode>): Likewise.
697 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
698 (aarch64_mls<mode>): Likewise.
699 (*aarch64_mls_elt<mode>): Likewise.
700 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
701 (<su><maxmin><mode>3): Likewise.
702 (move_lo_quad_<mode>): Likewise.
703 (aarch64_simd_move_hi_quad_<mode>): Likewise.
704 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
705 (vec_pack_trunc_<mode>): Likewise.
706 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
707 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
708 (*aarch64_<su>mlal_lo<mode>): Likewise.
709 (*aarch64_<su>mlal_hi<mode>): Likewise.
710 (*aarch64_<su>mlsl_lo<mode>): Likewise.
711 (*aarch64_<su>mlsl_hi<mode>): Likewise.
712 (*aarch64_<su>mlal<mode>): Likewise.
713 (*aarch64_<su>mlsl<mode>): Likewise.
714 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
715 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
716 (add<mode>3): Likewise.
717 (sub<mode>3): Likewise.
718 (mul<mode>3): Likewise.
719 (div<mode>3): Likewise.
720 (neg<mode>2): Likewise.
721 (abs<mode>2): Likewise.
722 (fma<mode>4): Likewise.
723 (*aarch64_fma4_elt<mode>): Likewise.
724 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
725 (*aarch64_fma4_elt_to_128df): Likewise.
726 (*aarch64_fma4_elt_to_64v2df): Likewise.
727 (fnma<mode>4): Likewise.
728 (*aarch64_fnma4_elt<mode>): Likewise.
729 (*aarch64_fnma4_elt_<vswap_width_name><mode>
730 (*aarch64_fnma4_elt_to_128df): Likewise.
731 (*aarch64_fnma4_elt_to_64v2df): Likewise.
732 (<frint_pattern><mode>2): Likewise.
733 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
734 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
735 (vec_unpacks_lo_v4sf): Likewise.
736 (aarch64_float_extend_lo_v2df): Likewise.
737 (vec_unpacks_hi_v4sf): Likewise.
738 (aarch64_float_truncate_lo_v2sf): Likewise.
739 (aarch64_float_truncate_hi_v4sf): Likewise.
740 (aarch64_vmls<mode>): Likewise.
741 (<su><maxmin><mode>3): Likewise.
742 (<maxmin_uns><mode>3): Likewise.
743 (reduc_<sur>plus_<mode>): Likewise.
744 (reduc_<sur>plus_v2di): Likewise.
745 (reduc_<sur>plus_v2si): Likewise.
746 (reduc_<sur>plus_<mode>): Likewise.
747 (aarch64_addpv4sf): Likewise.
748 (clz<mode>2): Likewise.
749 (reduc_<maxmin_uns>_<mode>): Likewise.
750 (reduc_<maxmin_uns>_v2di): Likewise.
751 (reduc_<maxmin_uns>_v2si): Likewise.
752 (reduc_<maxmin_uns>_<mode>): Likewise.
753 (reduc_<maxmin_uns>_v4sf): Likewise.
754 (aarch64_simd_bsl<mode>_internal): Likewise.
755 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
756 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
757 (aarch64_get_lane<mode>): Likewise.
758 (*aarch64_combinez<mode>): Likewise.
759 (aarch64_combine<mode>): Likewise.
760 (aarch64_simd_combine<mode>): Likewise.
761 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
762 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
763 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
764 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
765 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
766 (aarch64_<sur>h<addsub><mode>): Likewise.
767 (aarch64_<sur><addsub>hn<mode>): Likewise.
768 (aarch64_<sur><addsub>hn2<mode>): Likewise.
769 (aarch64_pmul<mode>): Likewise.
770 (aarch64_<su_optab><optab><mode>): Likewise.
771 (aarch64_<sur>qadd<mode>): Likewise.
772 (aarch64_sqmovun<mode>): Likewise.
773 (aarch64_<sur>qmovn<mode>): Likewise.
774 (aarch64_s<optab><mode>): Likewise.
775 (aarch64_sq<r>dmulh<mode>): Likewise.
776 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
777 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
778 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
779 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
780 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
781 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
782 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
783 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
784 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
785 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
786 (aarch64_sqdmull<mode>): Likewise.
787 (aarch64_sqdmull_lane<mode>_internal): Likewise.
788 (aarch64_sqdmull_n<mode>): Likewise.
789 (aarch64_sqdmull2<mode>_internal): Likewise.
790 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
791 (aarch64_sqdmull2_n<mode>_internal): Likewise.
792 (aarch64_<sur>shl<mode>): Likewise.
793 (aarch64_<sur>q<r>shl<mode>
794 (aarch64_<sur>shll_n<mode>): Likewise.
795 (aarch64_<sur>shll2_n<mode>): Likewise.
796 (aarch64_<sur>shr_n<mode>): Likewise.
797 (aarch64_<sur>sra_n<mode>): Likewise.
798 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
799 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
800 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
801 (aarch64_cm<optab><mode>): Likewise.
802 (aarch64_cm<optab>di): Likewise.
803 (aarch64_cm<optab><mode>): Likewise.
804 (aarch64_cm<optab>di): Likewise.
805 (aarch64_cmtst<mode>): Likewise.
806 (aarch64_cmtstdi): Likewise.
807 (aarch64_cm<optab><mode>): Likewise.
808 (*aarch64_fac<optab><mode>): Likewise.
809 (aarch64_addp<mode>): Likewise.
810 (aarch64_addpdi): Likewise.
811 (sqrt<mode>2): Likewise.
812 (vec_load_lanesoi<mode>): Likewise.
813 (vec_store_lanesoi<mode>): Likewise.
814 (vec_load_lanesci<mode>): Likewise.
815 (vec_store_lanesci<mode>): Likewise.
816 (vec_load_lanesxi<mode>): Likewise.
817 (vec_store_lanesxi<mode>): Likewise.
818 (*aarch64_mov<mode>): Likewise.
819 (aarch64_ld2<mode>_dreg): Likewise.
820 (aarch64_ld2<mode>_dreg): Likewise.
821 (aarch64_ld3<mode>_dreg): Likewise.
822 (aarch64_ld3<mode>_dreg): Likewise.
823 (aarch64_ld4<mode>_dreg): Likewise.
824 (aarch64_ld4<mode>_dreg): Likewise.
825 (aarch64_tbl1<mode>): Likewise.
826 (aarch64_tbl2v16qi): Likewise.
827 (aarch64_combinev16qi): Likewise.
828 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
829 (aarch64_st2<mode>_dreg): Likewise.
830 (aarch64_st2<mode>_dreg): Likewise.
831 (aarch64_st3<mode>_dreg): Likewise.
832 (aarch64_st3<mode>_dreg): Likewise.
833 (aarch64_st4<mode>_dreg): Likewise.
834 (aarch64_st4<mode>_dreg): Likewise.
835 (*aarch64_simd_ld1r<mode>): Likewise.
836 (aarch64_frecpe<mode>): Likewise.
837 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
838 (aarch64_frecps<mode>): Likewise.
839
f7379e5e
JG		 8402013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
841
842 * config/arm/iterators.md (V_elem_ch): New.
843 (q): Likewise.
844 (VQH_type): Likewise.
845 * config/arm/arm.md (is_neon_type): New.
846 (conds): Use is_neon_type.
847 (anddi3_insn): Update type attribute.
848 (xordi3_insn): Likewise.
849 (one_cmpldi2): Likewise.
850 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
851 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
852 (*movmisalign<mode>_neon_store): Likewise.
853 (*movmisalign<mode>_neon_load): Likewise.
854 (vec_set<mode>_internal): Likewise.
855 (vec_set<mode>_internal): Likewise.
856 (vec_setv2di_internal): Likewise.
857 (vec_extract<mode>): Likewise.
858 (vec_extract<mode>): Likewise.
859 (vec_extractv2di): Likewise.
860 (*add<mode>3_neon): Likewise.
861 (adddi3_neon): Likewise.
862 (*sub<mode>3_neon): Likewise.
863 (subdi3_neon): Likewise.
864 (fma<VCVTF:mode>4): Likewise.
865 (fma<VCVTF:mode>4_intrinsic): Likewise.
866 (*fmsub<VCVTF:mode>4): Likewise.
867 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
868 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
869 (ior<mode>3): Likewise.
870 (and<mode>3): Likewise.
871 (orn<mode>3_neon): Likewise.
872 (orndi3_neon): Likewise.
873 (bic<mode>3_neon): Likewise.
874 (bicdi3_neon): Likewise.
875 (xor<mode>3): Likewise.
876 (one_cmpl<mode>2): Likewise.
877 (abs<mode>2): Likewise.
878 (neg<mode>2): Likewise.
879 (negdi2_neon): Likewise.
880 (*umin<mode>3_neon): Likewise.
881 (*umax<mode>3_neon): Likewise.
882 (*smin<mode>3_neon): Likewise.
883 (*smax<mode>3_neon): Likewise.
884 (vashl<mode>3): Likewise.
885 (vashr<mode>3_imm): Likewise.
886 (vlshr<mode>3_imm): Likewise.
887 (ashl<mode>3_signed): Likewise.
888 (ashl<mode>3_unsigned): Likewise.
889 (neon_load_count): Likewise.
890 (ashldi3_neon_noclobber): Likewise.
891 (ashldi3_neon): Likewise.
892 (signed_shift_di3_neon): Likewise.
893 (unsigned_shift_di3_neon): Likewise.
894 (ashrdi3_neon_imm_noclobber): Likewise.
895 (lshrdi3_neon_imm_noclobber): Likewise.
896 (<shift>di3_neon): Likewise.
897 (widen_ssum<mode>3): Likewise.
898 (widen_usum<mode>3): Likewise.
899 (quad_halves_<code>v4si): Likewise.
900 (quad_halves_<code>v4sf): Likewise.
901 (quad_halves_<code>v8hi): Likewise.
902 (quad_halves_<code>v16qi): Likewise.
903 (reduc_splus_v2di): Likewise.
904 (neon_vpadd_internal<mode>): Likewise.
905 (neon_vpsmin<mode>): Likewise.
906 (neon_vpsmax<mode>): Likewise.
907 (neon_vpumin<mode>): Likewise.
908 (neon_vpumax<mode>): Likewise.
909 (*ss_add<mode>_neon): Likewise.
910 (*us_add<mode>_neon): Likewise.
911 (*ss_sub<mode>_neon): Likewise.
912 (*us_sub<mode>_neon): Likewise.
913 (neon_vadd<mode>_unspec): Likewise.
914 (neon_vaddl<mode>): Likewise.
915 (neon_vaddw<mode>): Likewise.
916 (neon_vhadd<mode>): Likewise.
917 (neon_vqadd<mode>): Likewise.
918 (neon_vaddhn<mode>): Likewise.
919 (neon_vmul<mode>): Likewise.
920 (neon_vfms<VCVTF:mode>): Likewise.
921 (neon_vmlal<mode>): Likewise.
922 (neon_vmls<mode>): Likewise.
923 (neon_vmlsl<mode>): Likewise.
924 (neon_vqdmulh<mode>): Likewise.
925 (neon_vqdmlal<mode>): Likewise.
926 (neon_vqdmlsl<mode>): Likewise.
927 (neon_vmull<mode>): Likewise.
928 (neon_vqdmull<mode>): Likewise.
929 (neon_vsub<mode>_unspec): Likewise.
930 (neon_vsubl<mode>): Likewise.
931 (neon_vsubw<mode>): Likewise.
932 (neon_vqsub<mode>): Likewise.
933 (neon_vhsub<mode>): Likewise.
934 (neon_vsubhn<mode>): Likewise.
935 (neon_vceq<mode>): Likewise.
936 (neon_vcge<mode>): Likewise.
937 (neon_vcgeu<mode>): Likewise.
938 (neon_vcgt<mode>): Likewise.
939 (neon_vcgtu<mode>): Likewise.
940 (neon_vcle<mode>): Likewise.
941 (neon_vclt<mode>): Likewise.
942 (neon_vcage<mode>): Likewise.
943 (neon_vcagt<mode>): Likewise.
944 (neon_vtst<mode>): Likewise.
945 (neon_vabd<mode>): Likewise.
946 (neon_vabdl<mode>): Likewise.
947 (neon_vaba<mode>): Likewise.
948 (neon_vabal<mode>): Likewise.
949 (neon_vmax<mode>): Likewise.
950 (neon_vmin<mode>): Likewise.
951 (neon_vpaddl<mode>): Likewise.
952 (neon_vpadal<mode>): Likewise.
953 (neon_vpmax<mode>): Likewise.
954 (neon_vpmin<mode>): Likewise.
955 (neon_vrecps<mode>): Likewise.
956 (neon_vrsqrts<mode>): Likewise.
957 (neon_vqabs<mode>): Likewise.
958 (neon_vqneg<mode>): Likewise.
959 (neon_vcls<mode>): Likewise.
960 (clz<mode>2): Likewise.
961 (popcount<mode>2): Likewise.
962 (neon_vrecpe<mode>): Likewise.
963 (neon_vrsqrte<mode>): Likewise.
964 (neon_vget_lane<mode>_sext_internal): Likewise.
965 (neon_vget_lane<mode>_zext_internal): Likewise.
966 (neon_vdup_n<mode>): Likewise.
967 (neon_vdup_n<mode>): Likewise.
968 (neon_vdup_nv2di): Likewise.
969 (neon_vdup_lane<mode>_interal): Likewise.
970 (*neon_vswp<mode>): Likewise.
971 (neon_vcombine<mode>): Likewise.
972 (float<mode><V_cvtto>2): Likewise.
973 (floatuns<mode><V_cvtto>2): Likewise.
974 (fix_trunc<mode><V_cvtto>2): Likewise.
975 (fixuns_trunc<mode><V_cvtto>2
976 (neon_vcvt<mode>): Likewise.
977 (neon_vcvt<mode>): Likewise.
978 (neon_vcvtv4sfv4hf): Likewise.
979 (neon_vcvtv4hfv4sf): Likewise.
980 (neon_vcvt_n<mode>): Likewise.
981 (neon_vcvt_n<mode>): Likewise.
982 (neon_vmovn<mode>): Likewise.
983 (neon_vqmovn<mode>): Likewise.
984 (neon_vqmovun<mode>): Likewise.
985 (neon_vmovl<mode>): Likewise.
986 (neon_vmul_lane<mode>): Likewise.
987 (neon_vmul_lane<mode>): Likewise.
988 (neon_vmull_lane<mode>): Likewise.
989 (neon_vqdmull_lane<mode>): Likewise.
990 (neon_vqdmulh_lane<mode>): Likewise.
991 (neon_vqdmulh_lane<mode>): Likewise.
992 (neon_vmla_lane<mode>): Likewise.
993 (neon_vmla_lane<mode>): Likewise.
994 (neon_vmlal_lane<mode>): Likewise.
995 (neon_vqdmlal_lane<mode>): Likewise.
996 (neon_vmls_lane<mode>): Likewise.
997 (neon_vmls_lane<mode>): Likewise.
998 (neon_vmlsl_lane<mode>): Likewise.
999 (neon_vqdmlsl_lane<mode>): Likewise.
1000 (neon_vext<mode>): Likewise.
1001 (neon_vrev64<mode>): Likewise.
1002 (neon_vrev32<mode>): Likewise.
1003 (neon_vrev16<mode>): Likewise.
1004 (neon_vbsl<mode>_internal): Likewise.
1005 (neon_vshl<mode>): Likewise.
1006 (neon_vqshl<mode>): Likewise.
1007 (neon_vshr_n<mode>): Likewise.
1008 (neon_vshrn_n<mode>): Likewise.
1009 (neon_vqshrn_n<mode>): Likewise.
1010 (neon_vqshrun_n<mode>): Likewise.
1011 (neon_vshl_n<mode>): Likewise.
1012 (neon_vqshl_n<mode>): Likewise.
1013 (neon_vqshlu_n<mode>): Likewise.
1014 (neon_vshll_n<mode>): Likewise.
1015 (neon_vsra_n<mode>): Likewise.
1016 (neon_vsri_n<mode>): Likewise.
1017 (neon_vsli_n<mode>): Likewise.
1018 (neon_vtbl1v8qi): Likewise.
1019 (neon_vtbl2v8qi): Likewise.
1020 (neon_vtbl3v8qi): Likewise.
1021 (neon_vtbl4v8qi): Likewise.
1022 (neon_vtbl1v16qi): Likewise.
1023 (neon_vtbl2v16qi): Likewise.
1024 (neon_vcombinev16qi): Likewise.
1025 (neon_vtbx1v8qi): Likewise.
1026 (neon_vtbx2v8qi): Likewise.
1027 (neon_vtbx3v8qi): Likewise.
1028 (neon_vtbx4v8qi): Likewise.
1029 (*neon_vtrn<mode>_insn): Likewise.
1030 (*neon_vzip<mode>_insn): Likewise.
1031 (*neon_vuzp<mode>_insn): Likewise.
1032 (neon_vld1<mode>): Likewise.
1033 (neon_vld1_lane<mode>): Likewise.
1034 (neon_vld1_lane<mode>): Likewise.
1035 (neon_vld1_dup<mode>): Likewise.
1036 (neon_vld1_dup<mode>): Likewise.
1037 (neon_vld1_dupv2di): Likewise.
1038 (neon_vst1<mode>): Likewise.
1039 (neon_vst1_lane<mode>): Likewise.
1040 (neon_vst1_lane<mode>): Likewise.
1041 (neon_vld2<mode>): Likewise.
1042 (neon_vld2<mode>): Likewise.
1043 (neon_vld2_lane<mode>): Likewise.
1044 (neon_vld2_lane<mode>): Likewise.
1045 (neon_vld2_dup<mode>): Likewise.
1046 (neon_vst2<mode>): Likewise.
1047 (neon_vst2<mode>): Likewise.
1048 (neon_vst2_lane<mode>): Likewise.
1049 (neon_vst2_lane<mode>): Likewise.
1050 (neon_vld3<mode>): Likewise.
1051 (neon_vld3qa<mode>): Likewise.
1052 (neon_vld3qb<mode>): Likewise.
1053 (neon_vld3_lane<mode>): Likewise.
1054 (neon_vld3_lane<mode>): Likewise.
1055 (neon_vld3_dup<mode>): Likewise.
1056 (neon_vst3<mode>): Likewise.
1057 (neon_vst3qa<mode>): Likewise.
1058 (neon_vst3qb<mode>): Likewise.
1059 (neon_vst3_lane<mode>): Likewise.
1060 (neon_vst3_lane<mode>): Likewise.
1061 (neon_vld4<mode>): Likewise.
1062 (neon_vld4qa<mode>): Likewise.
1063 (neon_vld4qb<mode>): Likewise.
1064 (neon_vld4_lane<mode>): Likewise.
1065 (neon_vld4_lane<mode>): Likewise.
1066 (neon_vld4_dup<mode>): Likewise.
1067 (neon_vst4<mode>): Likewise.
1068 (neon_vst4qa<mode>): Likewise.
1069 (neon_vst4qb<mode>): Likewise.
1070 (neon_vst4_lane<mode>): Likewise.
1071 (neon_vst4_lane<mode>): Likewise.
1072 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1073 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1074 (neon_vec_<US>mult_lo_<mode>): Likewise.
1075 (neon_vec_<US>mult_hi_<mode>): Likewise.
1076 (neon_vec_<US>shiftl_<mode>): Likewise.
1077 (neon_unpack<US>_<mode>): Likewise.
1078 (neon_vec_<US>mult_<mode>): Likewise.
1079 (vec_pack_trunc_<mode>): Likewise.
1080 (neon_vec_pack_trunc_<mode>): Likewise.
1081 (neon_vabd<mode>_2): Likewise.
1082 (neon_vabd<mode>_3): Likewise.
1083
a9e66678
JG		 10842013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1085
1086 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1087 (load_pair): Update type attribute.
1088 (store_pair): Update type attribute.
1089 * config/aarch64/iterators.md (q): New.
1090
20445ed6
JG		 10912013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1092
1093 * config/arm/types.md: Add new types for Neon insns.
1094
afb4ac68
AI		 10952013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1096 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1097 Sergey Lega <sergey.s.lega@intel.com>
1098 Anna Tikhonova <anna.tikhonova@intel.com>
1099 Ilya Tocar <ilya.tocar@intel.com>
1100 Andrey Turetskiy <andrey.turetskiy@intel.com>
1101 Ilya Verbin <ilya.verbin@intel.com>
1102 Kirill Yukhin <kirill.yukhin@intel.com>
1103 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1104
1105 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1106 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1107 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1108 (rcp14<mode>): New.
1109 (srcp14<mode>): Ditto.
1110 (rsqrt14<mode>): Ditto.
1111 (rsqrt14<mode>): Ditto.
1112 (avx512f_vmscalef<mode>): Ditto.
1113 (avx512f_scalef<mode>): Ditto.
1114 (avx512f_getexp<mode>): Ditto.
1115 (avx512f_sgetexp<mode>): Ditto.
1116 (avx512f_fixupimm<mode>): Ditto.
1117 (avx512f_sfixupimm<mode>): Ditto.
1118 (avx512f_rndscale<mode>): Ditto.
1119 (*avx512er_exp2<mode>): Ditto.
1120 (*avx512er_rcp28<mode>): Ditto.
1121 (avx512er_rsqrt28<mode>): Ditto.
1122 (avx512f_getmant<mode>): Ditto.
1123 (avx512f_getmant<mode>): Ditto.
1124 (avx512f_rndscale<mode>): Fix formatting.
1125
a1b499a3
JL		 11262013-10-15 Martin Jambor <mjambor@suse.cz>
1127
1128 * ipa-utils.h (ipa_edge_within_scc): Declare.
1129 * ipa-cp.c (edge_within_scc): Moved...
1130 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 1131
2e2206fa
AI		 11322013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1133 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1134 Sergey Lega <sergey.s.lega@intel.com>
1135 Anna Tikhonova <anna.tikhonova@intel.com>
1136 Ilya Tocar <ilya.tocar@intel.com>
1137 Andrey Turetskiy <andrey.turetskiy@intel.com>
1138 Ilya Verbin <ilya.verbin@intel.com>
1139 Kirill Yukhin <kirill.yukhin@intel.com>
1140 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1141
1142 * config/i386/predicates.md (const_8_to_15_operand): New.
1143 (const_16_to_31_operand): Ditto.
1144 * config/i386/sse.md (V8FI): New.
1145 (V16FI): Ditto.
1146 (reduc_splus_v8df): Ditto.
1147 (reduc_splus_v16sf): Ditto.
1148 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1149 (vec_extract_hi_<mode>): Ditto.
1150 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1151 (vec_set_lo_<mode>): Ditto.
1152 (vec_set_hi_<mode>): Ditto.
1153 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1154 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1155 (avx512f_pshufd_1): Ditto.
1156 (avx512f_broadcast<mode>): Ditto.
1157 (avx512f_broadcast<mode>): Ditto.
1158 (define_split): Split vec_extract_lo into move.
1159 (ssequartermode): Ditto.
1160 (ssedoublemode): Extened with wider modes.
1161 (vec_extract_lo_<mode>): Ditto.
1162
0fe65b75
AI		 11632013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1164 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1165 Sergey Lega <sergey.s.lega@intel.com>
1166 Anna Tikhonova <anna.tikhonova@intel.com>
1167 Ilya Tocar <ilya.tocar@intel.com>
1168 Andrey Turetskiy <andrey.turetskiy@intel.com>
1169 Ilya Verbin <ilya.verbin@intel.com>
1170 Kirill Yukhin <kirill.yukhin@intel.com>
1171 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1172
1173 * config/i386/predicates.md (register_or_constm1_operand): New.
1174 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1175 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1176 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1177 UNSPEC_SCATTER_PREFETCH
1178 (VI48_512): New.
1179 (avx512f_ucmp<mode>3): Ditto.
1180 (avx512f_vternlog<mode>): Ditto.
1181 (avx512f_align<mode>): Ditto.
1182 (<shift_insn><mode>3): Ditto.
1183 (avx512f_<rotate>v<mode>): Ditto.
1184 (avx512f_<rotate><mode>): Ditto.
1185 (avx512f_eq<mode>3): Ditto.
1186 (avx512f_eq<mode>3_1): Ditto.
1187 (avx512f_gt<mode>3): Ditto.
1188 (avx512f_testm<mode>3): Ditto.
1189 (avx512f_testnm<mode>3): Ditto.
1190 (avx512pf_gatherpf<mode>): Ditto.
1191 (*avx512pf_gatherpf<mode>_mask): Ditto.
1192 (*avx512pf_gatherpf<mode>): Ditto.
1193 (avx512pf_scatterpf<mode>): Ditto.
1194 (*avx512pf_scatterpf<mode>_mask): Ditto.
1195 (*avx512pf_scatterpf<mode>): Ditto.
1196 (avx512f_vec_dup_gpr<mode>): Ditto.
1197 (clz<mode>2): Ditto.
1198 (conflict<mode>): Ditto.
1199 (REDUC_SMINMAX_MODE): Extened with wider modes.
1200 (reduc_<code>_<mode>): Ditto.
1201 (vlshr<mode>3): Ditto.
1202 (vashl<mode>3): Ditto.
1203
ab931c71
AI		 12042013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1205 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1206 Sergey Lega <sergey.s.lega@intel.com>
1207 Anna Tikhonova <anna.tikhonova@intel.com>
1208 Ilya Tocar <ilya.tocar@intel.com>
1209 Andrey Turetskiy <andrey.turetskiy@intel.com>
1210 Ilya Verbin <ilya.verbin@intel.com>
1211 Kirill Yukhin <kirill.yukhin@intel.com>
1212 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1213
1214 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1215 UNSPEC_SCATTER.
1216 (VI48F_512): New.
1217 (avx512fmaskmode): Ditto.
1218 (bcstscalarsuff): Ditto.
1219 (avx512f_blendm<mode>): Ditto.
1220 (cmp_imm_predicate): Ditto.
1221 (avx512f_cmp<mode>3): Ditto.
1222 (avx512f_vec_dup<mode>): Ditto.
1223 (avx512f_vec_dup_mem<mode>): Ditto.
1224 (avx512f_vpermi2var<mode>3): Ditto.
1225 (avx512f_vpermt2var<mode>3): Ditto.
1226 (vec_init<mode>): Ditto.
1227 (avx512f_gathersi<mode>): Ditto.
1228 (*avx512f_gathersi<mode>): Ditto.
1229 (*avx512f_gathersi<mode>_2): Ditto.
1230 (avx512f_gatherdi<mode>): Ditto.
1231 (*avx512f_gatherdi<mode>): Ditto.
1232 (*avx512f_gatherdi<mode>_2): Ditto.
1233 (avx512f_scattersi<mode>): Ditto.
1234 (*avx512f_scattersi<mode>): Ditto.
1235 (avx512f_scatterdi<mode>): Ditto.
1236 (*avx512f_scatterdi<mode>): Ditto.
1237 (sseintprefix): Extened with wider modes.
1238 (VEC_GATHER_IDXSI): Ditto.
1239 (VEC_GATHER_IDXDI): Ditto.
1240 (VEC_GATHER_SRCDI): Ditto.
1241
c9acb877
MGD		 12422013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1243 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1244
1245 * config/arm/t-aprofile: New file.
1246 * config.gcc: Handle --with-multilib-list option.
1247
ecbda484
BS		 12482013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1249
6f1abb55
UB		 1250 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1251 the input for r2.
ecbda484 1252
24f161fd
RB		 12532013-10-15 Richard Biener <rguenther@suse.de>
1254
1255 * tree-loop-distribution.c (build_empty_rdg): Inline into
1256 single user.
1257 (rdg_flag_vertex): Inline into single user.
1258 (rdg_flag_vertex_and_dependent): Likewise.
1259 (build_rdg_partition_for_vertex): Remove processed bitmap.
1260 (rdg_build_partitions): Simplify.
1261
5de989ed
RB		 12622013-10-15 Richard Biener <rguenther@suse.de>
1263
1264 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1265 Restructure forwarding through conversions and copies to
1266 avoid performing copy-propagation the wrong way. Adjust
1267 recursion invocations.
1268 (forward_propagate_addr_expr): Add argument stating if we
1269 are recursing from a single-use.
1270 (ssa_forward_propagate_and_combine): Adjust.
1271
47e0da37
DM		 12722013-10-14 David Malcolm <dmalcolm@redhat.com>
1273
1274 * dumpfile.h (gcc::dump_manager): New class, to hold state
1275 relating to dumpfile management.
1276 (get_dump_file_name): Remove in favor of method of dump_manager.
1277 (dump_initialized_p): Likewise.
1278 (dump_start): Likewise.
1279 (dump_finish): Likewise.
1280 (dump_switch_p): Likewise.
1281 (dump_register): Likewise.
1282 (get_dump_file_info): Likewise.
1283 * context.c (gcc::context::context): Construct the dump_manager
1284 instance.
1285 * context.h (gcc::context::get_dumps): New.
1286 (gcc::context::m_dumps): New.
1287 * coverage.c (coverage_init): Port to dump_manager API.
6f1abb55 1288 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
47e0da37
DM		 1289 (extra_dump_files_in_use): Likewise.
1290 (extra_dump_files_alloced): Likewise.
1291 (gcc::dump_manager::dump_manager): New.
1292 (dump_register): Convert to...
1293 (gcc::dump_manager::dump_register): ...method, replacing
1294 function-static next_dump with m_next_dump field.
1295 (get_dump_file_info): Convert to...
1296 (gcc::dump_manager::get_dump_file_info): ...method.
1297 (get_dump_file_name): Convert to...
1298 (gcc::dump_manager::get_dump_file_name): ...method.
1299 (dump_start): Convert to...
1300 (gcc::dump_manager::dump_start): ...method.
1301 (dump_finish): Convert to...
1302 (gcc::dump_manager::dump_finish): ...method.
1303 (dump_begin): Replace body with...
1304 (gcc::dump_manager::dump_begin): ...new method.
1305 (dump_phase_enabled_p): Convert to...
1306 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1307 (dump_phase_enabled_p): Convert to...
1308 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1309 (dump_initialized_p): Convert to...
1310 (gcc::dump_manager::dump_initialized_p): ...method.
1311 (dump_flag_name): Replace body with...
1312 (gcc::dump_manager::dump_flag_name): ...new method.
1313 (dump_enable_all): Convert to...
1314 (gcc::dump_manager::dump_enable_all): ...new method.
1315 (opt_info_enable_passes): Convert to...
1316 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1317 (dump_switch_p_1): Convert to...
1318 (gcc::dump_manager::dump_switch_p_1): ...new method.
1319 (dump_switch_p): Convert to...
1320 (gcc::dump_manager::dump_switch_p): ...new method.
1321 (opt_info_switch_p): Port to dump_manager API.
1322 (enable_rtl_dump_file): Likewise.
1323 * opts-global.c (handle_common_deferred_options): Port to new
1324 dump_manager API.
1325 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1326 (pass_manager::register_one_dump_file): Likewise.
1327 (pass_manager::register_pass): Likewise.
1328 (pass_init_dump_file): Likewise.
1329 (pass_fini_dump_file): Likewise.
1330 * statistics.c (statistics_early_init): Likewise.
1331
83fd5d11
RB		 13322013-10-14 Richard Biener <rguenther@suse.de>
1333
1334 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1335 iterative_hash_canonical_type, gimple_canonical_type_hash,
1336 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1337 gimple_register_canonical_type, print_gimple_types_stats,
1338 free_gimple_type_tables): Move to lto/lto.c
1339 (gt-gimple.h): Do not include.
1340 * gimple.h (gimple_register_canonical_type,
1341 print_gimple_types_stats, free_gimple_type_tables): Remove.
1342 * Makefile.in (GTFILES): Remove gimple.c.
1343
a6dbd0f7
TS		 13442013-10-14 Travis Snoozy <quandary@remstate.com>
1345
1346 PR target/58716
1347 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1348 scanning for msp430x targets.
1349
b5ccb9ed
EB		 13502013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1351
1352 PR bootstrap/58509
1353 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1354 (registers_ok_for_ldd_peep): Move around.
1355 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1356 * config/sparc/sparc.md (widening peepholes): Use it.
1357
fd7e5084
RB		 13582013-10-14 Richard Biener <rguenther@suse.de>
1359
1360 PR middle-end/58712
1361 PR middle-end/55358
1362 * gimple.c (iterative_hash_canonical_type): Make sure to
1363 record the hash into the correct hashtable slot.
1364
3246afaa
EB		 13652013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1366
1367 PR rtl-optimization/58662
1368 * combine.c (try_combine): Take into account death nodes on I2 when
1369 splitting a PARALLEL of two independent SETs. Fix dump message.
1370
5d30dc5b
OE		 13712013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1372
1373 PR target/51244
1374 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1375 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1376 * config/sh/t-sh (sh_treg_combine.o): New entry.
1377 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1378 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1379 (register_sh_passes): New function. Register sh_treg_combine pass.
1380 (sh_option_override): Invoke it.
1381 (sh_canonicalize_comparison): Handle op0_preserve_value.
1382 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1383 opportunities. Canonicalize branch condition.
1384 (nott): Allow only if pseudos can be created for non-SH2A.
1385
37d1c5db
L		 13862013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1387
1388 PR target/58690
1389 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1390 (ix86_expand_movmem): Replace copy_addr_to_reg with
1391 ix86_copy_addr_to_reg.
1392 (ix86_expand_setmem): Likewise.
1393
b86a70ff
AM		 13942013-10-12 Alexander Monakov <amonakov@ispras.ru>
1395
1396 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1397 provided by ix86_fp_compare_mode instead of CCFPUmode.
1398
6e1d5961
JG		 13992013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1400
1401 * config/aarch64/arm_neon.h
1402 (vtbx<1,3>_<psu>8): Fix register constriants.
1403
ef3cfba2
JL		 14042013-10-11 Jeff Law <law@redhat.com>
1405
1406 PR tree-optimization/58640
adcfd489
UB		 1407 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1408 threading paths that cross over two loop entry points.
ef3cfba2 1409
6e8b7d9c
BS		 14102013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1411
1412 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1413 handle vector float as well.
1414 (*vsx_le_perm_load_v4si): Likewise.
1415 (*vsx_le_perm_store_v2di): Likewise.
1416 (*vsx_le_perm_store_v4si): Likewise.
1417
f200869a
BS		 14182013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1419
1420 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1421 directly to circumvent subtract from splat{31} workaround.
1422 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1423 prototype.
1424 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1425 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1426 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1427 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1428 separate big and little endian logic.
1429 (*altivec_vperm_<mode>_internal): New define_insn.
1430 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1431 separate big and little endian logic.
1432 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1433 (vec_permv16qi): Add little endian logic.
1434
8d28e3fc
MG		 14352013-10-11 Marc Glisse <marc.glisse@inria.fr>
1436
1437 * doc/extend.texi (returns_nonnull): Remove arguments.
1438
3c87b77b
AI		 14392013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1440 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1441 Sergey Lega <sergey.s.lega@intel.com>
1442 Anna Tikhonova <anna.tikhonova@intel.com>
1443 Ilya Tocar <ilya.tocar@intel.com>
1444 Andrey Turetskiy <andrey.turetskiy@intel.com>
1445 Ilya Verbin <ilya.verbin@intel.com>
1446 Kirill Yukhin <kirill.yukhin@intel.com>
1447 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1448
1449 * config/i386/sse.md (VI48F_256_512): New.
1450 (avx2_permvar<mode>): Change to ...
1451 (<avx2_avx512f>_permvar<mode>): This.
1452
50e60d7d
AI		 14532013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1454 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1455 Sergey Lega <sergey.s.lega@intel.com>
1456 Anna Tikhonova <anna.tikhonova@intel.com>
1457 Ilya Tocar <ilya.tocar@intel.com>
1458 Andrey Turetskiy <andrey.turetskiy@intel.com>
1459 Ilya Verbin <ilya.verbin@intel.com>
1460 Kirill Yukhin <kirill.yukhin@intel.com>
1461 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1462
6f1abb55 1463 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
50e60d7d
AI		 1464 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1465 * config/i386/sse.md (VI4_AVX): New.
1466 (sf2simodelower): Ditto.
1467 (sse2_cvtps2dq): Change to ...
1468 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1469
f62ce24f
AI		 14702013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1471 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1472 Sergey Lega <sergey.s.lega@intel.com>
1473 Anna Tikhonova <anna.tikhonova@intel.com>
1474 Ilya Tocar <ilya.tocar@intel.com>
1475 Andrey Turetskiy <andrey.turetskiy@intel.com>
1476 Ilya Verbin <ilya.verbin@intel.com>
1477 Kirill Yukhin <kirill.yukhin@intel.com>
1478 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1479
1480 * config/i386/sse.md (V_512): New.
1481 (VI_512): Ditto.
1482 (vcond<V_512:mode><VF_512:mode>): Ditto.
1483 (vcond<V_512:mode><VI_512:mode>): Ditto.
1484 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1485
558d9f79
AI		 14862013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1487 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1488 Sergey Lega <sergey.s.lega@intel.com>
1489 Anna Tikhonova <anna.tikhonova@intel.com>
1490 Ilya Tocar <ilya.tocar@intel.com>
1491 Andrey Turetskiy <andrey.turetskiy@intel.com>
1492 Ilya Verbin <ilya.verbin@intel.com>
1493 Kirill Yukhin <kirill.yukhin@intel.com>
1494 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1495
1496 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1497 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1498 (FMAMODE): Ditto.
1499 (fma<mode>4): Removed condition.
1500 (fms<mode>4): Ditto.
1501 (fnma<mode>4): Ditto.
1502 (fnms<mode>4): Ditto.
1503 (fma4i_fmadd_<mode>): Ditto.
1504 (*fma_fmadd_<mode>): Ditto.
1505 (*fma_fmsub_<mode>): Ditto.
1506 (*fma_fnmadd_<mode>): Ditto.
1507 (*fma_fnmsub_<mode>): Ditto.
1508 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1509 (*fma_fmaddsub_<mode>): Ditto.
1510 (*fma_fmsubadd_<mode>): Ditto.
1511 (*fmai_fmadd_<mode>): Ditto.
1512 (*fmai_fmsub_<mode>): Ditto.
1513 (*fmai_fnmadd_<mode>): Ditto.
1514 (*fmai_fnmsub_<mode>): Ditto.
1515
e8d08206
AI		 15162013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1517 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1518 Sergey Lega <sergey.s.lega@intel.com>
1519 Anna Tikhonova <anna.tikhonova@intel.com>
1520 Ilya Tocar <ilya.tocar@intel.com>
1521 Andrey Turetskiy <andrey.turetskiy@intel.com>
1522 Ilya Verbin <ilya.verbin@intel.com>
1523 Kirill Yukhin <kirill.yukhin@intel.com>
1524 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1525
1526 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1527 (VI124_256): Changed to ...
1528 (VI124_256_48_512): This.
1529 (ssepackmode): Extended with wider modes.
1530 (<code><mode>3): Changed iterator.
1531 (*avx2_<code><mode>3): Ditto.
1532 (vec_pack_trunc_<mode>): Ditto.
1533
16821545
AI		 15342013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1535 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1536 Sergey Lega <sergey.s.lega@intel.com>
1537 Anna Tikhonova <anna.tikhonova@intel.com>
1538 Ilya Tocar <ilya.tocar@intel.com>
1539 Andrey Turetskiy <andrey.turetskiy@intel.com>
1540 Ilya Verbin <ilya.verbin@intel.com>
1541 Kirill Yukhin <kirill.yukhin@intel.com>
1542 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1543
1544 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1545 (VI8F_256_512): Ditto.
1546 (abs<mode>2): Changed iterator.
1547 (avx2_perm<mode>): Changed to ...
1548 (<avx2_avx512f>_perm<mode>): This.
1549 (avx2_perm<mode>_1): Changed to ...
1550 (<avx2_avx512f>_perm<mode>_1): This.
1551
5348cff8
AI		 15522013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1553 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1554 Sergey Lega <sergey.s.lega@intel.com>
1555 Anna Tikhonova <anna.tikhonova@intel.com>
1556 Ilya Tocar <ilya.tocar@intel.com>
1557 Andrey Turetskiy <andrey.turetskiy@intel.com>
1558 Ilya Verbin <ilya.verbin@intel.com>
1559 Kirill Yukhin <kirill.yukhin@intel.com>
1560 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1561
1562 * config/i386/sse.md (VI48_AVX512F): New.
1563 (VI48_AVX2): Changed to ...
1564 (VI48_AVX2_48_AVX512F): This.
1565 (avx2_ashrv<mode>): Changed to ...
1566 (<avx2_avx512f>_ashrv<mode>): This.
1567 (avx2_<shift_insn>v<mode>): Changed to ...
1568 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1569
f5f41d88
AI		 15702013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1571 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1572 Sergey Lega <sergey.s.lega@intel.com>
1573 Anna Tikhonova <anna.tikhonova@intel.com>
1574 Ilya Tocar <ilya.tocar@intel.com>
1575 Andrey Turetskiy <andrey.turetskiy@intel.com>
1576 Ilya Verbin <ilya.verbin@intel.com>
1577 Kirill Yukhin <kirill.yukhin@intel.com>
1578 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1579
1580 * config/i386/sse.md (VI4_AVX512F): New.
1581 (VI8_AVX2_AVX512F): Ditto.
1582 (mul<mode>3): Extended with wider modes.
1583 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1584 (mul<mode>3): Ditto.
1585 (vec_widen_<s>mult_odd_<mode>): Ditto.
1586
3bdf6340
AI		 15872013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1588 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1589 Sergey Lega <sergey.s.lega@intel.com>
1590 Anna Tikhonova <anna.tikhonova@intel.com>
1591 Ilya Tocar <ilya.tocar@intel.com>
1592 Andrey Turetskiy <andrey.turetskiy@intel.com>
1593 Ilya Verbin <ilya.verbin@intel.com>
1594 Kirill Yukhin <kirill.yukhin@intel.com>
1595 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1596
1597 * config/i386/sse.md (VI2_AVX512F): New.
1598 (VI124_AVX512F): Ditto.
1599 (sseunpackmode): Extended with wider modes.
1600 (sseunpackfltmode): Ditto.
1601 (vec_unpacks_float_hi_<mode>): Ditto.
1602 (vec_unpacks_float_lo_<mode>): Ditto.
1603 (vec_unpacku_float_hi_<mode>): Ditto.
1604 (vec_unpacku_float_lo_<mode>): Ditto.
1605 (vec_unpacks_lo_<mode>): Ditto.
1606 (vec_unpacks_hi_<mode>): Ditto.
1607 (vec_unpacku_lo_<mode>): Ditto.
1608 (vec_unpacku_hi_<mode>): Ditto.
1609
67f783cb
AI		 16102013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1611 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1612 Sergey Lega <sergey.s.lega@intel.com>
1613 Anna Tikhonova <anna.tikhonova@intel.com>
1614 Ilya Tocar <ilya.tocar@intel.com>
1615 Andrey Turetskiy <andrey.turetskiy@intel.com>
1616 Ilya Verbin <ilya.verbin@intel.com>
1617 Kirill Yukhin <kirill.yukhin@intel.com>
1618 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1619
1620 * config/i386/i386.md (multdiv): New.
1621 (multdiv_mnemonic): Ditto.
1622 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1623 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1624 (<sse>_vmdiv<mode>3): Removed.
1625
ec5e777c
AI		 16262013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1627 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1628 Sergey Lega <sergey.s.lega@intel.com>
1629 Anna Tikhonova <anna.tikhonova@intel.com>
1630 Ilya Tocar <ilya.tocar@intel.com>
1631 Andrey Turetskiy <andrey.turetskiy@intel.com>
1632 Ilya Verbin <ilya.verbin@intel.com>
1633 Kirill Yukhin <kirill.yukhin@intel.com>
1634 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1635
1636 * config/i386/sse.md (V): Extended with wider modes.
1637 (VF2): Ditto.
1638 (ssehalfvecmode): Ditto.
1639 (i128): Ditto.
1640 (ssepackfltmode): Ditto.
1641 (avx_vec_concat<mode>): Ditto.
1642 (V_256_512): New iterator.
1643 (VF2_512_256): Ditto.
1644 (si2dfmode): New attribute.
1645 (si2dfmodelower): Ditto.
1646 (sf2dfmode): Ditto.
1647 (concat_tg_mode): Ditto.
1648 (floatv4siv4df2): Changed to ...
1649 (float<si2dfmodelower><mode>2): This.
1650 (avx_cvtps2pd256): Changed to ...
1651 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1652 (vec_pack_trunc_v4df): Changed to ...
1653 (vec_pack_trunc_<mode>): This.
1654 (avx_vpermil<mode>): Changed to ...
1655 (<sse2_avx_avx512f>_vpermil<mode>): This.
1656 (<fixsuffix>fix_truncv8dfv8si2): New.
1657 (vec_pack_sfix_trunc_v8df): Ditto.
1658 (avx512f_rndscale<mode>): Ditto.
1659 (avx512f_roundpd512): Ditto.
1660 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1661
a9ccbba2
AI		 16622013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1663 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1664 Sergey Lega <sergey.s.lega@intel.com>
1665 Anna Tikhonova <anna.tikhonova@intel.com>
1666 Ilya Tocar <ilya.tocar@intel.com>
1667 Andrey Turetskiy <andrey.turetskiy@intel.com>
1668 Ilya Verbin <ilya.verbin@intel.com>
1669 Kirill Yukhin <kirill.yukhin@intel.com>
1670 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1671
1672 * config/i386/i386.md (any_fix): New iterator.
1673 (fixsuffix): New attribute.
1674 * config/i386/sse.md (VF1): Extened with wider modes.
1675 (VI): Ditto.
1676 (VI_AVX2): Ditto.
1677 (VI8): Ditto.
1678 (sseintvecmodelower): Ditto.
1679 (ssescalarmode): Ditto.
1680 (ssescalarnum): Ditto.
1681 (VF1_128_256): New.
1682 (ssexmmmode): Ditto.
1683 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1684 (<sse>_rcp<mode>2): Change iterator.
1685 (rsqrt<mode>2): Ditto.
1686 (<sse>_rsqrt<mode>2): Ditto.
1687 (avx2_vec_dup<mode>): Ditto.
1688 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1689 (round<mode>2_sfix): Ditto.
1690 (avx2_pbroadcast<mode>): Ditto.
1691 (*andnot<mode>3): Handle XI mode.
1692 (*<code><mode>3): Ditto.
1693 (AVXTOSSEMODE): Removed.
1694 (avx_vpermil<mode>): Changed to ...
1695 (<sse2_avx_avx512f>_vpermil<mode>): This.
1696
2b1ebb0c
AI		 16972013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1698 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1699 Sergey Lega <sergey.s.lega@intel.com>
1700 Anna Tikhonova <anna.tikhonova@intel.com>
1701 Ilya Tocar <ilya.tocar@intel.com>
1702 Andrey Turetskiy <andrey.turetskiy@intel.com>
1703 Ilya Verbin <ilya.verbin@intel.com>
1704 Kirill Yukhin <kirill.yukhin@intel.com>
1705 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1706
1707 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1708 (<sse>_comi): Ditto.
1709 (<sse>_ucomi): Ditto.
1710 (sse_cvtss2siq_2): Ditto.
1711 (sse2_cvtsd2si): Ditto.
1712 (sse2_cvtsd2siq): Ditto.
1713 (sse2_cvttsd2si): Ditto.
1714 (sse2_cvttsd2siq): Ditto.
1715 (<shift_insn><mode>3): Ditto.
1716 (sse2_cvtsi2sdq): Update constraint and prefix.
1717 (sse_cvtsi2ss): Update prefix.
1718 (sse_cvtsi2ssq): Ditto.
1719
d89124ed
JJ		 17202013-10-11 Jakub Jelinek <jakub@redhat.com>
1721
1722 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1723 ignore internal calls.
1724
e6ad28c3
RB		 17252013-10-11 Richard Biener <rguenther@suse.de>
1726
6f1abb55
UB		 1727 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1728 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1729 TREE_OVERFLOW is set.
e6ad28c3 1730
d1227447
TS		 17312013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1732
1ac4b8e4
TS		 1733 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1734
7d05cebb
TS		 1735 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1736 * gimple.h: Likewise.
1737
adcfd489 1738 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 1739
1a79fb8a
TS		 1740 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1741 macros.
1742
3afd2873
TS		 1743 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1744
adcfd489 1745 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS		 1746 * configure: Regenerate.
1747
acf0174b
JJ		 17482013-10-11 Jakub Jelinek <jakub@redhat.com>
1749
1750 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1751 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1752 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1753 similarly pp_right_paren.
1754 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1755 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1756 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1757 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1758 clauses.
1759 (omp_declare_simd_clauses_equal,
1760 omp_remove_redundant_declare_simd_attrs): New functions.
1761 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1762 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1763 * tree.h (OMP_LOOP_CHECK): Define.
1764 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1765 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1766 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1767 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1768 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1769 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1770 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1771 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1772 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1773 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1774 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1775 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1776 (omp_remove_redundant_declare_simd_attrs): New prototype.
1777 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1778 GIMPLE_OMP_TEAMS): New codes.
1779 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1780 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1781 fields.
1782 (target_nesting_level): New variable.
1783 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1784 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1785 for collapse > 1 static schedule unless ordered.
1786 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1787 (determine_parallel_type): Adjust get_ws_args_for caller.
1788 (install_var_field): Handle mask & 4 for double indirection.
1789 (scan_sharing_clauses): Ignore shared clause on teams construct.
1790 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1791 (create_omp_child_function): If inside target or declare target
adcfd489 1792 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ		 1793 (find_combined_for): New function.
1794 (scan_omp_parallel): Handle combined loops.
1795 (scan_omp_target, scan_omp_teams): New functions.
1796 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1797 restrictions and set ctx->cancellable for cancellable constructs.
1798 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1799 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1800 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 1801 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b 1802 (omp_clause_aligned_alignment): New function.
6f1abb55 1803 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
acf0174b
JJ		 1804 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1805 on teams constructs. Handle user defined reductions and new
1806 OpenMP 4.0 clauses.
1807 (lower_reduction_clauses): Don't set placeholder to address of ref
1808 if it has already the right type.
1809 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1810 (expand_parallel_call): Use the new non-_start suffixed builtins,
1811 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1812 and GOMP_parallel_end after the call.
1813 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1814 (expand_omp_for_init_counts): Handle combined loops.
1815 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1816 loops.
1817 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1818 end of cancellable loops.
1819 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1820 Likewise. Handle collapse > 1 loops.
1821 (expand_omp_simd): Handle combined loops.
1822 (expand_omp_for): Add inner_stmt argument, adjust callers of
1823 expand_omp_for* functions, use expand_omp_for_static*chunk even
1824 for collapse > 1 unless ordered.
1825 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1826 of cancellable sections.
1827 (expand_omp_single): Remove need_barrier variable, just rely on
1828 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1829 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1830 (expand_omp_atomic_load, expand_omp_atomic_store,
1831 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1832 (expand_omp_target): New function.
1833 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1834 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1835 (build_omp_regions_1): Immediately close region for
1836 GF_OMP_TARGET_KIND_UPDATE.
1837 (maybe_add_implicit_barrier_cancel): New function.
1838 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1839 cancellation.
1840 (lower_omp_single): Likewise. Add clobber after the barrier.
1841 (lower_omp_taskgroup): New function.
1842 (lower_omp_for): Handle combined loops. Adjust
1843 lower_rec_input_clauses caller. Handle cancellation.
1844 (lower_depend_clauses): New function.
1845 (lower_omp_taskreg): Lower depend clauses. Adjust
1846 lower_rec_input_clauses caller. Add clobber after the call. Handle
1847 cancellation.
1848 (lower_omp_target, lower_omp_teams): New functions.
1849 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1850 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1851 and GOMP_cancellation_point calls.
1852 (lower_omp): Fold stmts inside of target region.
1853 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1854 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1855 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1856 (BT_FN_VOID_OMPFN_PTR_UINT,
1857 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1858 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1859 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1860 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1861 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1862 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1863 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1864 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1865 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1866 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1867 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1868 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1869 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1870 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1871 BUILT_IN_GOMP_PARALLEL_END.
1872 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1873 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1874 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1875 GF_OMP_FOR_KIND_DISTRIBUTE.
1876 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1877 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1878 (dump_gimple_omp_return): Print lhs if it has any.
1879 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1880 gimple_omp_atomic_seq_cst_p.
1881 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1882 and GIMPLE_OMP_TEAMS.
1883 * langhooks.c (lhd_omp_mappable_type): New function.
1884 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1885 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1886 hook.
1887 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1888 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6f1abb55 1889 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
acf0174b
JJ		 1890 (struct gimplify_omp_ctx): Add combined_loop field.
1891 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1892 on stmts inside of target region.
1893 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1894 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1895 ORT_TARGET and ORT_TARGET_DATA.
1896 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1897 Handle GOVD_MAP.
1898 (omp_notice_threadprivate_variable): Complain about threadprivate
1899 variables in target region.
1900 (omp_notice_variable): Complain about vars with non-mappable type
1901 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1902 (omp_check_private): Ignore ORT_TARGET* regions.
1903 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1904 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1905 (find_combined_omp_for): New function.
1906 (gimplify_omp_for): Handle gimplification of combined loops.
1907 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1908 OMP_TEAMS.
1909 (gimplify_omp_target_update): New function.
1910 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1911 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1912 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1913 (gimplify_body): If fndecl has "omp declare target" attribute, add
1914 implicit ORT_TARGET context around it.
1915 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1916 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1917 * tree-nested.c (convert_nonlocal_reference_stmt,
1918 convert_local_reference_stmt, convert_gimple_call): Handle
1919 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1920 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1921 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1922 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1923 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1924 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1925 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1926 BUILT_IN_GOMP_LOOP_END_CANCEL,
1927 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1928 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1929 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1930 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1931 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1932 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1933 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1934 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1935 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1936 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1937 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1938 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1939 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1940 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1941 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1942 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1943 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1944 gimple_build_omp_teams): New functions.
1945 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1946 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1947 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1948 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1949 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1950 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1951 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1952 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1953 GF_OMP_ATOMIC_SEQ_CST): New.
1954 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1955 gimple_build_omp_teams): New prototypes.
1956 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1957 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1958 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1959 GIMPLE_OMP_SINGLE as end of range.
1960 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1961 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1962 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1963 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1964 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1965 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1966 gimple_omp_target_kind, gimple_omp_target_set_kind,
1967 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1968 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1969 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1970 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1971 gimple_omp_teams_set_clauses): New inlines.
1972 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1973 and GIMPLE_OMP_TASKGROUP.
adcfd489 1974 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ		 1975 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1976 enum omp_clause_proc_bind_kind): New.
1977 (union omp_clause_subcode): Add depend_kind, map_kind and
1978 proc_bind_kind fields.
1979 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1980 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1981 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1982 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1983 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1984
67fa7880
TJ		 19852013-10-10 Teresa Johnson <tejohnson@google.com>
1986
1987 * predict.c (tree_estimate_probability): Add new parameter
1988 for estimate_bb_frequencies.
1989 (estimate_bb_frequencies): Add new parameter to force estimation.
1990 (rebuild_frequencies): When max frequency in function is small,
1991 recompute counts from frequencies.
1992 * predict.h (estimate_bb_frequencies): New parameter.
1993
b591a8b7
DM		 19942013-10-10 David Malcolm <dmalcolm@redhat.com>
1995
1996 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1997 optimizations are disabled.
1998
782f0db2
DM		 19992013-10-10 David Malcolm <dmalcolm@redhat.com>
2000
2001 * coverage.c (coverage_finish): Fix leak of da_file_name.
2002
13ef00fa
JH		 20032013-10-10 Jan Hubicka <jh@suse.cz>
2004
2005 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2006 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2007 Bobcat and generic.
2008
a1d50386
JJ		 20092013-10-10 Jakub Jelinek <jakub@redhat.com>
2010
2011 PR middle-end/58670
2012 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2013 if any labels are in FALLTHRU_BB, use a special label emitted
2014 immediately after the asm goto insn rather than label_rtx
2015 of the LABEL_DECL.
2016 (expand_asm_stmt): Adjust caller.
2017 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2018 edge if the last insn in predecessor is a jump with single successor,
2019 but it isn't simplejump_p.
2020
e2c2fde2
RB		 20212013-10-10 Richard Biener <rguenther@suse.de>
2022
2023 PR tree-optimization/58656
6f1abb55 2024 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
e2c2fde2 2025
de6a5608
AM		 20262013-10-10 Andrew MacLeod <amacleod@redhat.com>
2027
adcfd489
UB		 2028 * gimplify.c: Include expr.h and tm_p.h for targets with special
2029 va-arg padding requirements.
de6a5608 2030
b184c8f1
AM		 20312013-10-10 Andrew MacLeod <amacleod@redhat.com>
2032
2033 * tree-flow.h: Move some prototypes to gimple.h.
2034 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2035 * gimple.h: Relocate some prototypes from tree-flow.h
2036 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2037 Move to gimplify.c.
2038 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2039 (build_va_arg_indirect_ref): Relocate and make static.
2040 (std_gimplify_va_arg_expr): Relocate here.
2041 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2042 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2043
13b72c22
AK		 20442013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2045
2046 * doc/md.texi: Document the mnemonic attribute.
2047
5db40447
AK		 20482013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2049
2050 PR target/57377
2051 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2052 (set_attr_alternative x ...) when searching for user defined
2053 mnemonic attribute.
2054
ed018d4b
AM		 20552013-10-10 Andrew MacLeod <amacleod@redhat.com>
2056
2057 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2058
4d5b5e9f
ER		 20592013-10-09 Easwaran Raman <eraman@google.com>
2060
2061 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2062 * cfgexpand.c (defer_stack_allocation): ...use here
2063 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2064
c9ef86a1
ZC		 20652013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2066
2067 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2068 (operand_equal_for_value_replacement): New function, extracted from
2069 value_replacement and enhanced to catch more cases.
2070 (value_replacement): Use operand_equal_for_value_replacement.
2071
2431114f
AM		 20722013-10-09 Andrew MacLeod <amacleod@redhat.com>
2073
adcfd489 2074 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM		 2075 get_max_loop_iterations.
2076
1dd2a9a0
KT		 20772013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2078
2079 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2080 Place comment above function.
2081
c1bf2a39
AM		 20822013-10-09 Andrew MacLeod <amacleod@redhat.com>
2083
2084 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2085 are not related to tree-cfg.c.
2086 * tree-ssa-address.h: New file. Relocate prototypes.
2087 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2088 (addr_for_mem_ref): New. Combine call to get_address_description and
2089 return addr_for_mem_ref.
2090 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2091 * tree-ssa-live.h: Adjust prototypes.
2092 * passes.c: Include tree-ssa-live.h.
2093 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2094 * graphite.c (graphite_transform_loops): Make static.
2095 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 2096 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM		 2097 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2098 * ipa-pure-const.c (warn_function_noreturn): Make static.
2099 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2100 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2101 Relocate from tree-cfg.c
2102 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2103 static.
2104 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2105 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2106 Move to ipa-pure-const.c.
adcfd489
UB		 2107 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2108 Relocate from tree-optimize.c.
c1bf2a39
AM		 2109 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2110 make_pass_fixup_cfg): Move to tree-cfg.c.
2111 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2112 Relocate some prototypes.
2113 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2114 * tree-dump.c (dump_function_to_file): Remove prototype.
2115 Add tree-flow.h to the include file.
2116 * tree-dump.h: Remove prototype.
2117 * tree-parloops.h: New File. Add prototypes.
2118 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2119 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2120 from tree-ssa-loop.c.
2121 * tree-predcom.c (run_tree_predictive_commoning,
2122 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2123 Relocate here from tree-ssa-loop.c.
adcfd489 2124 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM		 2125 ssa_name_values.release ().
2126 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2127 (ssa_name_values): Relocate from tree-flow.h.
2128 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2129 * tree-ssa-loop.c (run_tree_predictive_commoning,
2130 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2131 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2132 make_pass_graphite, pass_data_graphite_transforms,
2133 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2134 tree_parallelize_loops, pass_data_parallelize_loops,
2135 make_pass_parallelize_loops): Move to other files.
2136 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2137 moved here.
2138 * tree.h: Remove prototypes from tree-address.c.
2139
cc524fc7
AM		 21402013-10-09 Andrew MacLeod <amacleod@redhat.com>
2141
2142 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2143 (struct int_tree_map): Move to tree-hasher.h
2144 (SCALE, LABEL, PERCENT): Move to gimple.h
2145 * tree-flow-inline.h: Delete. Move functions to other files.
2146 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2147 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2148 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2149 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2150 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2151 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2152 * tree-hasher.h: Don't include tree-flow.h.
2153 (struct int_tree_map): Relocate from tree-flow.h.
2154 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2155 tree-flow-inline.h and make static.
adcfd489
UB		 2156 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2157 tree-flow-inline.h.
cc524fc7
AM		 2158 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2159 tree-flow-inline.h and make static.
2160 * tree.h (is_global_var, may_be_aliased): Relocate from
2161 tree-flow-inline.h.
2162 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2163 * value-prof.c: No longer include tree-flow-inline.h.
2164 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2165
71343877
AM		 21662013-10-09 Andrew MacLeod <amacleod@redhat.com>
2167
2168 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2169 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2170 (enum move_pos): Move to tree-ssa-loop-im.h
2171 * cfgloop.h: Move some prototypes.
2172 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2173 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2174 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2175 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2176 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2177 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2178 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2179 (enum move_pos): Relocate here.
2180 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2181 tree-ssa-loop.c.
2182 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB		 2183 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2184 make_pass_lim): Relocate here from tree-ssa-loop.c.
2185 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM		 2186 tree-ssa-loop.c.
2187 (loop_edge_to_cancel, unloop_loops): Make static.
2188 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2189 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2190 (tree_complete_unroll, gate_tree_complete_unroll,
2191 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2192 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2193 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2194 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2195 (stmt_invariant_in_loop_p): Remove unused function.
2196 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2197 * tree-ssa-loop-manip.h: New file. Add prototypes.
2198 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2199 (gcov_type_to_double_int): Move to cfgloop.h.
2200 (double_int_cmp, bound_index,
2201 estimate_numbers_of_iterations_loop): Make static.
2202 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2203 (max_loop_iterations): Factor out get_max_loop_iterations.
2204 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2205 cfgloop.c.
2206 * tree-ssa-loop-niter.h: New file. Add prototypes.
2207 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2208 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2209 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2210 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2211 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2212 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2213 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2214 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2215 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2216 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2217 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2218 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2219 pass_data_complete_unroll, make_pass_complete_unroll,
2220 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2221 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2222 tree-ssa-loop-ivcanon.c.
2223 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2224 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2225 tree-ssa-loop-prefetch.c.
2226 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2227 tree-ssa-loop-im.c.
2228 (get_lsm_tmp_name): Relocate and add suffix parameter.
2229 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2230 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2231 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2232 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 2233 (get_estimated_loop_iterations): Factor out accessor from
71343877 2234 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB		 2235 (get_max_loop_iterations): Factor out accessor from
2236 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM		 2237 * loop-unroll.c (decide_unroll_constant_iterations,
2238 decide_unroll_runtime_iterations, decide_peel_simple,
2239 decide_unroll_stupid): Use new get_* accessors.
2240
826cacfe
MG		 22412013-10-09 Marc Glisse <marc.glisse@inria.fr>
2242
2243 PR tree-optimization/20318
2244 * doc/extend.texi (returns_nonnull): New function attribute.
2245 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2246 attribute.
2247 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2248 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2249
378f8976
EB		 22502013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2251
2252 PR middle-end/58570
2253 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2254 false if both components are bitfields.
2255
0fe04f5c
AV		 22562013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2257
2258 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2259 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2260 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2261 * config/aarch64/aarch64.h
2262 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2263 * config/aarch64/aarch64-simd-builtins.def
2264 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2265
bed9bae4
AV		 22662013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2267
2268 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2269
7f3d8b19
AV		 22702013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2271
2272 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2273
7df625a6
AV		 22742013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2275
2276 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2277 (vneg_f64): New intrinsic.
2278 (vneg_s8): Asm replaced with C.
2279 (vneg_s16): Likewise.
2280 (vneg_s32): Likewise.
2281 (vneg_s64): New intrinsic.
2282 (vnegq_f32): Asm replaced with C.
2283 (vnegq_f64): Likewise.
2284 (vnegq_s8): Likewise.
2285 (vnegq_s16): Likewise.
2286 (vnegq_s32): Likewise.
2287 (vnegq_s64): Likewise.
2288
2b86fca7
RL		 22892013-10-09 Renlin Li <Renlin.Li@arm.com>
2290
2291 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2292
e179df83
AK		 22932013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2294
2295 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2296 packed stack special handling.
2297 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2298 back to fixed stack slots for FPRs saved due to stdarg.
2299
74129172
AK		 23002013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2301
2302 * config/s390/s390.c (s390_frame_info): Restructure function.
2303
6455a49e
AK		 23042013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2305
2306 * config/s390/s390.c (struct s390_frame_layout): New field
2307 gpr_save_slots.
2308 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2309 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2310 regs_ever_clobbered to char*.
adcfd489
UB		 2311 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2312 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK		 2313 (s390_register_info): Enable FPR save slots. Move/Copy some
2314 functionality into ...
2315 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2316 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2317 function.
2318 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2319 not imply a stack frame.
2320 (s390_init_frame_layout): Remove variable clobbered_regs.
2321 (s390_update_register_info): Remove function.
2322 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2323 cannot be used for register renaming.
2324 (s390_hard_regno_scratch_ok): New function.
2325 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2326 (s390_initial_elimination_offset): Change offset calculation of
2327 the return address pointer.
adcfd489 2328 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 2329 (restore_gprs): Set frame related flag.
adcfd489 2330 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK		 2331 (s390_emit_prologue): Call s390_register_info instead of
2332 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2333 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2334 (s390_optimize_prologue): Call s390_optimize_register_info.
2335 Try to remove also FPR slot save/restore INSNs. Remove frame
2336 related flags from restore INSNs.
2337
59d96342
DD		 23382013-10-08 DJ Delorie <dj@redhat.com>
2339
0dc7adad
DD		 2340 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2341 (movhi): Likewise.
2342
59d96342
DD		 2343 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2344 avoid conflict with the MI use of %c.
2345 * config/rl78/rl78-real.md: change %c to %C throughout.
2346 * config/rl78/rl78-virt.md: Likewise.
2347
a81169d8
JH		 23482013-10-08 Jan Hubicka <jh@suse.cz>
2349
2350 * config/i386/i386.c (ix86_option_override_internal): Switch
2351 to SSE math for -ffast-math when target ISA supports SSE2.
2352
481d1b81
AM		 23532013-10-08 Andrew MacLeod <amacleod@redhat.com>
2354
2355 * tree-flow.h: Remove some prototypes.
2356 * tree.h: Remove some protypes, add a couple.
2357 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2358 using_eh_for_cleanups_p): Add interface routines for front ends.
2359 * tree-eh.h: New file. Add protoptyes.
2360 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2361 (add_stmt_to_eh_lp_fn): Make static.
2362 (lower_try_finally): Use new using_eh_for_cleanups_p.
2363 * emit-rtl.c: Include tree-eh.h.
2364 * gimple.h: Include tree-eh.h.
2365
0e6a0e48
MG		 23662013-10-08 Marc Glisse <marc.glisse@inria.fr>
2367
2368 PR tree-optimization/58480
2369 * tree-vrp.c (infer_nonnull_range): New function.
2370 (infer_value_range): Call infer_nonnull_range.
2371
ef6179d1
DC		 23722013-10-08 Dehao Chen <dehao@google.com>
2373
2374 PR tree-optimization/58619
2375 * tree-inline.c (copy_phis_for_bb): Combine location data
2376 only if non-null.
2377
07d964d5
ZC		 23782013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2379
2380 PR target/58423
2381 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2382 RTX_FRAME_RELATED_P on INSN.
2383
19e9b2a3
BS		 23842013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2385
2386 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2387 (altivec_expand_vec_perm_const): Call it.
2388
0cf68694
BS		 23892013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2390
2391 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2392 sequences for LE VSX loads and stores at expand time.
2393 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2394 prototype.
2395 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2396 (rs6000_gen_le_vsx_permute): New.
2397 (rs6000_gen_le_vsx_load): New.
2398 (rs6000_gen_le_vsx_store): New.
2399 (rs6000_gen_le_vsx_move): New.
2400 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2401 (*vsx_le_perm_load_v4si): New.
2402 (*vsx_le_perm_load_v8hi): New.
2403 (*vsx_le_perm_load_v16qi): New.
2404 (*vsx_le_perm_store_v2di): New.
2405 (*vsx_le_perm_store_v4si): New.
2406 (*vsx_le_perm_store_v8hi): New.
2407 (*vsx_le_perm_store_v16qi): New.
2408 (*vsx_xxpermdi2_le_<mode>): New.
2409 (*vsx_xxpermdi4_le_<mode>): New.
2410 (*vsx_xxpermdi8_le_V8HI): New.
2411 (*vsx_xxpermdi16_le_V16QI): New.
2412 (*vsx_lxvd2x2_le_<mode>): New.
2413 (*vsx_lxvd2x4_le_<mode>): New.
2414 (*vsx_lxvd2x8_le_V8HI): New.
2415 (*vsx_lxvd2x16_le_V16QI): New.
2416 (*vsx_stxvd2x2_le_<mode>): New.
2417 (*vsx_stxvd2x4_le_<mode>): New.
2418 (*vsx_stxvd2x8_le_V8HI): New.
2419 (*vsx_stxvd2x16_le_V16QI): New.
2420
9520e1eb
RL		 24212013-10-07 Renlin Li <Renlin.Li@arm.com>
2422
2423 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2424
36c0bd4f
AK		 24252013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2426
2427 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2428 loop to work also for 31bit ABI.
2429 Save the stack pointer for frame_size > 0.
2430
ee163e72
AK		 24312013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2432
2433 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2434 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2435 constraint letters from expanders.
2436 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2437 retry count to general_operand.
2438 ("tabort"): Give operand 0 a mode.
2439 ("tabort_1"): Add mode and constraint letter for operand 0.
2440 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2441
45d99234
JL		 24422013-10-04 Jeff Law <law@redhat.com>
2443
2444 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2445
adcfd489
UB		 2446 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2447 out of ...
45d99234
JL		 2448 (thread_across_edge): Here. Call it.
2449
7c327f7b
CC		 24502013-10-04 Cary Coutant <ccoutant@google.com>
2451
2452 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2453 discarding a location list expression (or a piece of one).
7c327f7b 2454
6867e128
JH		 24552013-10-03 Jan Hubicka <jh@suse.cz>
2456
65efa7e7
UB		 2457 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2458 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2459 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2460
c7f36d55
JH		 24612013-10-03 Jan Hubicka <jh@suse.cz>
2462
65efa7e7 2463 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH		 2464 accumulate-outgoing-args when producing unwind info.
2465
4f0bee4c
WM		 24662013-10-03 Wei Mi <wmi@google.com>
2467
2468 * lra-constraints.c (insert_move_for_subreg): New function
2469 extracted from simplify_operand_subreg.
2470 (simplify_operand_subreg): Add reload for paradoxical subreg.
2471
111c3f39
RX		 24722013-10-03 Rong Xu <xur@google.com>
2473
65efa7e7
UB		 2474 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2475 the candidate of builtin_expect such that we should fix the
2476 size/time estimation.
2477 (estimate_function_body_sizes): Do the acutally size/time fix-up
2478 for builtin_expect.
111c3f39 2479
942df739
RX		 24802013-10-03 Rong Xu <xur@google.com>
2481
65efa7e7
UB		 2482 * predict.c (tree_predict_by_opcode): Get the probability
2483 for builtin_expect from param builtin_expect_probability.
2484 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2485 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2486 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2487
2284b034
MG		 24882013-10-03 Marc Glisse <marc.glisse@inria.fr>
2489
2490 PR c++/19476
2491 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2492 * calls.c (alloca_call_p): Use get_callee_fndecl.
2493 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2494 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2495 Likewise.
2496 (vrp_visit_stmt): Remove duplicated code.
2497
0609bdf2
MM		 24982013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2499
2500 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2501 ceildf2, btruncdf2, instead of vsx_* name.
2502
2503 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2504 iterators to only do V2DF and V4SF here. Move the DF code to
2505 rs6000.md where it is combined with SF mode. Replace <VSv> with
2506 just 'v' since only vector operations are handled with these insns
2507 after moving the DF support to rs6000.md.
2508 (vsx_sub<mode>3): Likewise.
2509 (vsx_mul<mode>3): Likewise.
2510 (vsx_div<mode>3): Likewise.
2511 (vsx_fre<mode>2): Likewise.
2512 (vsx_neg<mode>2): Likewise.
2513 (vsx_abs<mode>2): Likewise.
2514 (vsx_nabs<mode>2): Likewise.
2515 (vsx_smax<mode>3): Likewise.
2516 (vsx_smin<mode>3): Likewise.
2517 (vsx_sqrt<mode>2): Likewise.
2518 (vsx_rsqrte<mode>2): Likewise.
2519 (vsx_fms<mode>4): Likewise.
2520 (vsx_nfma<mode>4): Likewise.
2521 (vsx_copysign<mode>3): Likewise.
2522 (vsx_btrunc<mode>2): Likewise.
2523 (vsx_floor<mode>2): Likewise.
2524 (vsx_ceil<mode>2): Likewise.
2525 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2526 (vsx_sminsf3): Likewise.
2527 (vsx_fmadf4): Likewise.
2528 (vsx_fmsdf4): Likewise.
2529 (vsx_nfmadf4): Likewise.
2530 (vsx_nfmsdf4): Likewise.
2531 (vsx_cmpdf_internal1): Likewise.
2532
2533 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2534 simpler to select whether a target has SPE or traditional floating
2535 point support in iterators.
2536 (TARGET_DF_SPE): Likewise.
2537 (TARGET_SF_FPR): Likewise.
2538 (TARGET_DF_FPR): Likewise.
2539 (TARGET_SF_INSN): Macros to say whether floating point support
2540 exists for a given operation for expanders.
2541 (TARGET_DF_INSN): Likewise.
2542
2543 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2544 combining of SF/DF mode operations, using both traditional and VSX
2545 registers.
2546 (Fvsx): Likewise.
2547 (Ff): Likewise.
2548 (Fv): Likewise.
2549 (Fs): Likewise.
2550 (Ffre): Likewise.
2551 (FFRE): Likewise.
2552 (abs<mode>2): Combine SF/DF modes using traditional floating point
2553 instructions. Add support for using the upper DF registers with
2554 VSX support, and SF registers with power8-vector support. Update
2555 expanders for operations supported by both the SPE and traditional
2556 floating point units.
2557 (abs<mode>2_fpr): Likewise.
2558 (nabs<mode>2): Likewise.
2559 (nabs<mode>2_fpr): Likewise.
2560 (neg<mode>2): Likewise.
2561 (neg<mode>2_fpr): Likewise.
2562 (add<mode>3): Likewise.
2563 (add<mode>3_fpr): Likewise.
2564 (sub<mode>3): Likewise.
2565 (sub<mode>3_fpr): Likewise.
2566 (mul<mode>3): Likewise.
2567 (mul<mode>3_fpr): Likewise.
2568 (div<mode>3): Likewise.
2569 (div<mode>3_fpr): Likewise.
2570 (sqrt<mode>3): Likewise.
2571 (sqrt<mode>3_fpr): Likewise.
2572 (fre<Fs>): Likewise.
2573 (rsqrt<mode>2): Likewise.
2574 (cmp<mode>_fpr): Likewise.
2575 (smax<mode>3): Likewise.
2576 (smin<mode>3): Likewise.
2577 (smax<mode>3_vsx): Likewise.
2578 (smin<mode>3_vsx): Likewise.
2579 (negsf2): Delete SF operations that are merged with DF.
2580 (abssf2): Likewise.
2581 (addsf3): Likewise.
2582 (subsf3): Likewise.
2583 (mulsf3): Likewise.
2584 (divsf3): Likewise.
2585 (fres): Likewise.
2586 (fmasf4_fpr): Likewise.
2587 (fmssf4_fpr): Likewise.
2588 (nfmasf4_fpr): Likewise.
2589 (nfmssf4_fpr): Likewise.
2590 (sqrtsf2): Likewise.
2591 (rsqrtsf_internal1): Likewise.
2592 (smaxsf3): Likewise.
2593 (sminsf3): Likewise.
2594 (cmpsf_internal1): Likewise.
2595 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2596 (negdf2): Delete DF operations that are merged with SF.
2597 (absdf2): Likewise.
2598 (nabsdf2): Likewise.
2599 (adddf3): Likewise.
2600 (subdf3): Likewise.
2601 (muldf3): Likewise.
2602 (divdf3): Likewise.
2603 (fred): Likewise.
2604 (rsqrtdf_internal1): Likewise.
2605 (fmadf4_fpr): Likewise.
2606 (fmsdf4_fpr): Likewise.
2607 (nfmadf4_fpr): Likewise.
2608 (nfmsdf4_fpr): Likewise.
2609 (sqrtdf2): Likewise.
2610 (smaxdf3): Likewise.
2611 (smindf3): Likewise.
2612 (cmpdf_internal1): Likewise.
2613 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2614 (btrunc<mode>2): Delete separate expander, and combine with the
2615 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2616 (btrunc<mode>2_fpr): Likewise.
2617 (ceil<mode>2): Likewise.
2618 (ceil<mode>2_fpr): Likewise.
2619 (floor<mode>2): Likewise.
2620 (floor<mode>2_fpr): Likewise.
2621 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2622 Add support for using the upper registers with VSX and
2623 power8-vector. Move insns to be closer to the define_expands. On
2624 VSX systems, prefer the traditional form of FMA over the VSX
2625 version, since the traditional form allows the target not to
2626 overlap with the inputs.
2627 (fms<mode>4_fpr): Likewise.
2628 (nfma<mode>4_fpr): Likewise.
2629 (nfms<mode>4_fpr): Likewise.
2630
5bea0c6c 26312013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 2632 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT		 2633
2634 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2635 (struct mult_cost_table): Likewise.
2636 (struct mem_cost_table): Likewise.
2637 (struct fp_cost_table): Likewise.
2638 (struct vector_cost_table): Likewise.
2639 (cpu_cost_table): Likewise.
2640 * config/arm/arm.opt (mold-rts-costs): New option.
2641 (mnew-generic-costs): Likewise.
2642 * config/arm/arm.c (generic_extra_costs): New table.
2643 (cortexa15_extra_costs): Likewise.
2644 (arm_slowmul_tune): Use NULL as new costs.
2645 (arm_fastmul_tune): Likewise.
2646 (arm_strongarm_tune): Likewise.
2647 (arm_xscale_tune): Likewise.
2648 (arm_9e_tune): Likewise.
2649 (arm_v6t2_tune): Likewise.
2650 (arm_cortex_a5_tune): Likewise.
2651 (arm_cortex_a9_tune): Likewise.
2652 (arm_v6m_tune): Likewise.
2653 (arm_fa726te_tune): Likewise.
2654 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2655 (arm_cortex_tune): Use generict_extra_costs.
2656 (shifter_op_p): New function.
2657 (arm_unspec_cost): Likewise.
2658 (LIBCALL_COST): Define.
2659 (arm_new_rtx_costs): New function.
2660 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2661 table is available. Use old costs otherwise unless mnew-generic-costs
2662 is specified.
2663 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2664 (cpu_cost_table): Declare.
2665
24c56925
MS		 26662013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2667
2668 PR target/58460
2669 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2670 (*subs_mul_imm_<mode>)
2671 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2672 (*sub_<shift>_<mode>)
2673 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2674 Remove k constraint.
2675
2ab8f063
IB		 26762013-10-03 Ian Bolton <ian.bolton@arm.com>
2677
2678 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2679 code.
2680 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2681
79221839
TJ		 26822013-10-02 Teresa Johnson <tejohnson@google.com>
2683
2684 * predict.c (probably_never_executed): New function.
2685 (probably_never_executed_bb_p): Invoke probably_never_executed.
2686 (probably_never_executed_edge_p): Ditto.
2687 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2688 Treat profile insanities conservatively.
2689
d441afe0
JDA		 26902013-10-02 John David Anglin <danglin@gcc.gnu.org>
2691
2692 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2693
80f466c4
VM		 26942013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2695
2696 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2697 use it. Use smaller increase for scratch. Don't increase reject
2698 for early clobber scratch.
2699 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2700 setting eliminated regs except setting fp from hfp.
2701 (lra_eliminate): Check lra_insn_recog_data on NULL.
2702
6e228b4b
MM		 27032013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2704
2705 PR target/58587
2706 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 2707 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM		 2708 (RS6000_CPU, power7 defaults): Likewise.
2709
11b54a5a
UB		 27102013-10-02 Uros Bizjak <ubizjak@gmail.com>
2711
2712 * config/x-linux (host-linux.o): Remove header dependencies.
2713 Use $(COMPILE) and $(POSTCOMPILE).
2714 * config/t-linux-android (linux-android.o): Ditto.
2715
bbc02b69
UB		 27162013-10-02 Uros Bizjak <ubizjak@gmail.com>
2717
2718 * Makefile.in (expmed.o-warn): Remove.
2719
440917de
AM		 27202013-10-02 Andrew MacLeod <amacleod@redhat.com>
2721
2722 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2723 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2724
99206ca9
TJ		 27252013-10-02 Teresa Johnson <tejohnson@google.com>
2726
2727 * dojump.c (do_jump_1): Divide probability between
2728 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2729
56e82b14
TT		 27302013-10-02 Tom Tromey <tromey@redhat.com>
2731
2732 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2733
3d9c733e
AM		 27342013-10-02 Andrew MacLeod <amacleod@redhat.com>
2735
2736 * tree-flow.h: Remove some prototypes.
2737 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2738 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2739 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2740 mark_virtual_phi_result_for_renaming): Relocate here.
2741 * tree-into-ssa.h: Add prototypes.
bbc02b69 2742 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM		 2743 single_pred_before_succ_order.
2744 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2745 (nonfreeing_call_p) Move to gimple.c.
2746 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2747 tree-ssa-phiopt.c.
2748 * basic-block.h (single_pred_before_succ_order): Add prototype.
2749 * gimple.c (nonfreeing_call_p): Relocate here.
2750 * gimple.h: Add prototype.
2751 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2752 * tree-ssa-dom.h: New file. Relocate prototypes here.
2753 * tree-ssa.h: Include tree-ssa-dom.h.
2754
78cedfb1
UB		 27552013-10-02 Uros Bizjak <ubizjak@gmail.com>
2756
2757 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2758 Use $(COMPILE) and $(POSTCOMPILE).
2759
2760 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2761
744730a4
AM		 27622013-10-02 Andrew MacLeod <amacleod@redhat.com>
2763
2764 * tree-flow.h: Remove some prototypes.
2765 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2766 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2767 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2768 propagate_tree_value*): Move from here to...
2769 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2770 propagate_tree_value*): Relocate here.
2771 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2772 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2773 * gimple-fold.c: Remove gimple-fold.h from include list.
2774 * tree-vrp.c: Remove gimple-fold.h from include list.
2775 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2776 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2777 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2778 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2779 * sese.c: Add tree-ssa-propagate.h to include list.
2780
826a536d
RB		 27812013-10-02 Richard Biener <rguenther@suse.de>
2782
2783 * tree-loop-distribution.c: Include tree-vectorizer.h for
2784 find_loop_location.
2785 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 2786 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB		 2787 (partition_alloc): Adjust.
2788 (partition_builtin_p): Likewise.
2789 (partition_has_writes): Remove.
2790 (partition_reduction_p): New function.
2791 (partition_merge_into): Likewise.
2792 (generate_code_for_partition): Commonize builtin partition
2793 handling tail.
2794 (rdg_cannot_recompute_vertex_p): Remove.
2795 (already_processed_vertex_p): Likewise.
2796 (rdg_flag_vertex): Do not set has_writes.
2797 (classify_partition): Adjust.
2798 (rdg_build_partitions): Do not set has_writes, treat all
2799 partitions as useful.
78cedfb1 2800 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB		 2801 (tree_loop_distribution): Report number of loops and library
2802 calls generated as opt-info.
2803
4b403ece
AM		 28042013-10-02 Andrew MacLeod <amacleod@redhat.com>
2805
2806 * tree-flow.h: Include new .h files. Move prototypes.
2807 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2808 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2809 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2810 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2811 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2812 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2813 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2814 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 2815 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM		 2816 tree-dfa.h.
2817 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2818 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2819 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2820 * tree-scalar-evolution.c: Include tree.h.
2821 * sese.c: Include tree.h.
2822 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2823 * dwarf2out.c: Include tree-dfa.h.
2824 * tree-chrec.c: Include tree.h.
2825 * tree-data-ref.c: Include tree.h.
2826
1d2151c6
YZ		 28272013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2828
2829 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2830 Fix whitespace.
2831
441ad147
RO		 28322013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2833
2834 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2835 Use $(COMPILE) and $(POSTCOMPILE).
2836 (sol2-cxx.o): Likewise.
2837 (sol2-stubs.o): Likewise.
2838 (sol2.o): Likewise.
2839 * config/x-solaris (host-solaris.o): Likewise.
2840
2841 * config/sparc/t-sparc (sparc.o): Remove.
2842 (sparc-c.o): Remove header dependencies.
2843 Use $(COMPILE) and $(POSTCOMPILE).
2844 * config/sparc/x-sparc: Likewise.
2845
2deaf8b0
JR		 28462013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2847
2848 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2849 * config/arc/arc700.md: Likewise.
2850 * config/arc/arc-modes.def: Likewise.
2851 * config/arc/arc-simd.h: Likewise.
2852 * config/arc/t-arc-uClibc: Likewise.
2853 * config/arc/t-arc-newlib: Likewise.
2854
f6fe771a
RL		 28552013-10-02 Renlin Li <renlin.li@arm.com>
2856
78cedfb1
UB		 2857 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2858 plus_constant.
f6fe771a
RL		 2859 (aarch64_expand_epilogue): Likewise.
2860
0916f876
YZ		 28612013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2862 Yufeng Zhang <yufeng.zhang@arm.com>
2863
2864 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2865 declaration.
2866 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2867 'base_in' represent a conversion and legal_cast_p_1 holds; set
2868 'base_in' with the returned value from get_unwidened.
2869
ebfcd719
KT		 28702013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2871
2872 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2873 plus_constant is not used.
2874
157ca3e9
WM		 28752013-10-01 Wei Mi <wmi@google.com>
2876
78cedfb1 2877 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM		 2878 * config/i386/i386.md: Add define_peephole2 to
2879 break partial reg stall for cvtss2sd/cvtsd2ss.
2880
cd4dd8f0
JR		 28812013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2882
2883 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 2884 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 2885
1465e5cf
JL		 28862013-10-01 Jeff Law <law@redhat.com>
2887
2888 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2889 outgoing_edge and intermediate_edge fields. Instead store the path.
2890 (redirection_data::hash): Hash on the last edge's destination index.
2891 (redirection_data::equal): Check the entire thread path.
2892 (lookup_redirectio_data): Corresponding changes.
2893 (create_edge_and_update_destination_phis): Likewise.
2894 (thread_single_edge): Likewise.
2895
25c606cb 28962013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 2897 Diego Novillo <dnovillo@google.com>
25c606cb
JR		 2898
2899 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2900 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2901 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2902 versions of these patterns.
2903
0ccbc132
JR		 2904 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2905 (__builtin_arc_aligned): Likewise.
2906
f55d4a20
JR		 2907 * config/arc/arc.md: Expand adc_0 comment stating the intended
2908 purpose and why it isn't ready.
2909 Replace commented out call_value_via_label_mixed with a
2910 plain comment about bl_s.
2911
5719867d 2912 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR		 2913 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2914 Remove if (1) condition.
2915 (arc_encode_section_info): Fix comment.
2916
6462fab0
JR		 29172013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2918
2919 * config/arc/arc.c (arc_conditional_register_usage):
2920 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2921 Also set reg_alloc_order for DMA config regs.
2922
5d5f6720
JR		 29232013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2924 Jeremy Bennett <jeremy.bennett@embecosm.com>
2925
2926 * doc/install.texi (--with-cpu): Mention ARC.
2927 (arc-*-elf32): New paragraph.
2928 (arc-linux-uclibc): Likewise.
2929 * doc/md.texi (Machine Constraints): Add ARC part.
2930 * doc/invoke.texi: (menu): Add ARC Options.
2931 (Machine Dependent Options) <ARC Options>: Add synopsis.
2932 (node ARC Options): Add.
2933 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2934 (ARC Built-in Functions): New section defining
2935 generic ARC built-in functions.
2936 (ARC SIMD Built-in Functions): New section defining SIMD specific
2937 built-in functions.
2938 (Declaring Attributes of Functions): Extended
2939 description of short_call and long_call attributes for ARC and
2940 added index entries.
2941
526b7aee
SV		 29422013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2943 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2944 Joern Rennecke <joern.rennecke@embecosm.com>
2945 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2946 Brendan Kehoe <brendan@zen.org>
2947 Michael Eager <eager@eagercon.com>
2948 Simon Cook <simon.cook@embecosm.com>
2949 Jeremy Bennett <jeremy.bennett@embecosm.com>
2950
2951 * config/arc, common/config/arc: New directories.
2952
53426f6c
JR		 29532013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2954 Brendan Kehoe <brendan@zen.org>
2955 Simon Cook <simon.cook@embecosm.com>
2956
2957 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2958
1df9f5a9
AM		 29592013-10-01 Andrew MacLeod <amacleod@redhat.com>
2960
2961 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2962 * tree-ssa-coalesce.h: New. Move prototype to here.
2963 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2964 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2965 (gimple_can_coalesce_p): Move to...
2966 * gimple.c (gimple_can_coalesce_p): Here.
2967
80560f95
AM		 29682013-10-01 Andrew MacLeod <amacleod@redhat.com>
2969
2970 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2971 (dump_decl_set): Move to gimple.c.
2972 * gimple.h: Don't include tree-ssa-operands.h.
2973 (dump_decl_set): Add prototype.
2974 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2975 Move to gimple-ssa.h.
2976 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2977 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2978 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2979 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2980 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2981 than PHI_ARG_DEF.
2982 (dump_decl_set): Relocate here.
2983 * gimple-ssa.h: New file.
2984 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2985 Relocate from gimple.h.
2986 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2987 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2988 swap_tree_operands and remove non-ssa path.
2989 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2990 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2991 swap_ssa_operands.
2992 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2993 vect_is_simple_reduction_1): Use swap_ssa_operands.
2994 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2995 (enum need_phi_state): Move to tree-into-ssa.c.
2996 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2997 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2998 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2999 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3000 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3001 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3002 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3003 num_imm_uses): Move to ssa-iterators.h.
3004 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7ab4168e 3005 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
80560f95
AM		 3006 tree-phinodes.h.
3007 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3008 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3009 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3010 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3011 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3012 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3013 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3014 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3015 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3016 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3017 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3018 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3019 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3020 (set_phi_nodes): Move to tree-phinodes.h.
3021 * tree-ssa-operands.h (enum ssa_op_iter_type,
3022 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3023 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3024 (dump_decl_set): Remove prototype.
3025 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3026 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3027 (set_phi_nodes): Relocate from tree-flow-inline.h.
3028 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3029 tree-flow-inline.h
3030 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3031 include list. Temporarily add gimple.h to include list.
3032 * ssa-iterators.h: New file.
3033 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3034 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3035 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3036 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3037 Relocate from tree-ssa-operands.h.
3038 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3039 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3040 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3041 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3042 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3043 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3044 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3045 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3046 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3047 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3048 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3049 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3050 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3051 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3052 Relocate from tree-flow-inline.h.
3053 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3054
8da00d65
VP		 30552013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3056
3057 * aarch64-simd.md
3058 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3059 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3060 Insert '\t' to output template.
3061 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3062 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3063 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3064 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3065
e2ebe1c2
UB		 30662013-10-01 Uros Bizjak <ubizjak@gmail.com>
3067
3068 * doc/install.texi (Host/target specific installation notes for GCC):
3069 Put @anchor before @heading.
3070 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3071 Use @email for email addresses.
3072
aee2d611
JL		 30732013-10-01 Jeff Law <law@redhat.com>
3074
3075 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3076 a vec. Only delete the path if we create one without successfully
3077 registering a jump thread.
3078 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3079 as a pointer.
3080 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3081 (paths): New vector of jump threading paths.
3082 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3083 (THREAD_PATH): New accessor macro for the entire thread path.
3084 (lookup_redirection_data): Get intermediate and final outgoing edge
3085 from the thread path.
3086 (create_edge_and_update_destination_phis): Copy the threading path.
3087 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3088 jump threading path.
3089 (ssa_redirect_edges): Get edges and block types from the jump threading
3090 path. Free the path vector.
3091 (thread_block): Get edges from the jump threading path. Look at the
3092 entire path to see if we thread to a loop exit. If we cancel a jump
3093 thread request, then free the path vector.
3094 (thread_single_edge): Get edges and block types from the jump threading
3095 path. Free the path vector.
3096 (thread_through_loop_header): Get edges and block types from the jump
3097 threading path. Free the path vector.
3098 (mark_threaded_blocks): Iterate over the vector of paths and store
3099 the path on the appropriate edge. Get edges and block types from the
3100 jump threading path.
3101 (mark_threaded_blocks): Get edges and block types from the jump
3102 threading path. Free the path vector.
3103 (thread_through_all_blocks): Use the vector of paths rather than
3104 a vector of 3-edge sets.
3105 (register_jump_thread): Accept pointer to a path vector rather
3106 than the path vector itself. Store the path vector for later use.
3107 Simplify.
3108
966f97ac 31092013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 3110 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ		 3111
3112 PR target/58574
3113 * config/s390/s390.c (s390_split_branches): Modify check for table
3114 jump insns.
3115 (s390_chunkify_start): Rearrange table jump insn check in order to
3116 deal with compare and branch insns correctly.
3117
3a323a38
KV		 31182013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3119
3120 PR target/58578
3121 Revert
3122 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3123 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3124 define_insn_and_split.
3125 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3126 (shiftsi3_compare): New pattern.
3127 (rrx): New pattern.
3128 * config/arm/unspecs.md (UNSPEC_RRX): New.
3129
90444831
AM		 31302013-10-01 Alan Modra <amodra@gmail.com>
3131
3132 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3133 casing inout operands.
3134
818625cf
RB		 31352013-10-01 Richard Biener <rguenther@suse.de>
3136
3137 PR tree-optimization/58553
3138 * tree-loop-distribution.c (struct partition_s): Add niter member.
3139 (classify_partition): Populate niter member for the partition
3140 and properly identify whether the relevant store happens before
3141 or after the loop exit.
3142 (generate_memset_builtin): Use niter member from the partition.
3143 (generate_memcpy_builtin): Likewise.
3144
30f641cd
RS		 31452013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3146
3147 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3148 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3149
65d3284b
RS		 31502013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3151
3152 * basic-block.h (edge_list): Prefix member names with "m_".
3153 * context.h (context): Likewise.
3154 * domwalk.h (dom_walker): Likewise.
3155 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3156 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3157 * hash-table.h (hash_table): Likewise.
3158 * machmode.h (bit_field_mode_iterator): Likewise.
3159 * pass_manager.h (pass_list): Likewise.
3160 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3161 * tree-pass.h (pass_data): Likewise.
3162 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3163 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3164 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3165 * asan.c (pass_data_asan): Update accordingly.
3166 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3167 (control_dependences::control_dependences): Likewise.
3168 (control_dependences::~control_dependences): Likewise.
3169 (control_dependences::~control_dependences): Likewise.
3170 (control_dependences::get_edges_dependent_on): Likewise.
3171 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3172 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3173 * context.c (gcc::context::context): Likewise.
3174 * cprop.c (pass_rtl_cprop::clone): Likewise.
3175 * domwalk.c (dom_walker::walk): Likewise.
3176 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3177 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3178 * mode-switching.c (pass_mode_switching::clone): Likewise.
3179 * passes.c (opt_pass::opt_pass): Likewise.
3180 (pass_manager::pass_manager): Likewise.
3181 * predict.c (pass_strip_predict_hints::clone): Likewise.
3182 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3183 (pass_split_all_insns::clone): Likewise.
3184 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3185 Likewise.
3186 (bit_field_mode_iterator::next_mode): Likewise.
3187 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3188 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3189 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3190 * tree-complex.c (pass_lower_complex::clone): Likewise.
3191 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3192 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3193 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3194 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3195 (pass_fold_builtins::clone): Likewise.
3196 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3197 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3198 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3199 (pass_cd_dce::clone): Likewise.
3200 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3201 (pass_phi_only_cprop::clone): Likewise.
3202 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3203 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3204 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3205 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3206 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3207 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3208 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3209 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3210 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3211 * tree-vrp.c (pass_vrp::clone): Likewise.
3212 * tsan.c (pass_tsan::clone): Likewise.
3213
f66d0891
JJ		 32142013-09-30 Jakub Jelinek <jakub@redhat.com>
3215
d2a365a8
JJ		 3216 PR middle-end/58564
3217 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3218 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3219
f66d0891
JJ		 3220 PR middle-end/58564
3221 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3222 optimization, punt if sign_bit_p looked through any zero extension.
3223
05357ac3
TJ		 32242013-09-30 Teresa Johnson <tejohnson@google.com>
3225
3226 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3227 Update redirected out edge count in joiner case.
3228 (ssa_redirect_edges): Common the joiner and non-joiner cases
3229 so that joiner case gets profile updates.
3230
ca406576
RB		 32312013-09-30 Richard Biener <rguenther@suse.de>
3232
3233 PR tree-optimization/58554
e2ebe1c2
UB		 3234 * tree-loop-distribution.c (classify_partition): Require
3235 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB		 3236 (tree_loop_distribution): Calculate dominance info.
3237
92d649c4
VK		 32382013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3239
3240 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3241 (NO_PROFILE_COUNTERS): Likewise.
3242 (PROFILE_HOOK): Likewise.
3243 (FUNCTION_PROFILER): Likewise.
3244 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3245
bd9534e2
IS		 32462013-09-30 Iain Sandoe <iain@codesourcery.com>
3247
3248 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3249 calls and defines in TARGET_MACHO conditional.
3250 (load_macho_picbase_di): Likewise.
3251 (reload_macho_picbase): Likewise.
3252 (reload_macho_picbase_si): Likewise.
3253 (reload_macho_picbase_di): Likewise.
3254 (nonlocal_goto_receiver): Likewise.
3255
4494fbc9
NC		 32562013-09-30 Nick Clifton <nickc@redhat.com>
3257
3258 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3259 that use the MSP430X ISA.
3260 (msp430_option_override): Scan -mmcu command line option for any
3261 MCU name that supports the MSP430X ISA.
3262 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3263 -mmcu options which enable the MSP430X ISA.
3264
791d9044
RB		 32652013-09-30 Richard Biener <rguenther@suse.de>
3266
3267 PR middle-end/58532
3268 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3269 before looking for setjmp-like calls.
3270
e1c5c877
IS		 32712013-09-29 Iain Sandoe <iain@codesourcery.com>
3272
3273 PR target/10901
3274 * config/darwin-protos.h (machopic_get_function_picbase): New.
3275 * config/darwin.c (machopic_get_function_picbase): New.
3276 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3277 label for a new func. (load_macho_picbase_di): Likewise.
3278 (reload_macho_picbase): New expand.
3279 (reload_macho_picbase_si): New insn.
3280 (reload_macho_picbase_di): New insn.
3281 (nonlocal_goto_receiver): New define and split.
3282 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3283 (unspecv enum): Add UNSPECV_NLGR.
3284
749278c5
IS		 32852013-09-29 Iain Sandoe <iain@codesourcery.com>
3286
3287 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3288 that altivec registers are correctly sized on Darwin.
3289
2c43a51c
IS		 32902013-09-29 Iain Sandoe <iain@codesourcery.com>
3291
3292 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3293 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3294 * config/x-darwin (host-darwin.o): Likewise.
3295 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3296 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3297 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3298
7a5911d3
UB		 32992013-09-29 Uros Bizjak <ubizjak@gmail.com>
3300
3301 * doc/invoke.texi: Fix usage of @tie{} command.
3302
3f67a73d
EB		 33032013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3304
3305 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3306
9ac2f538
JH		 33072013-09-28 Jan Hubicka <jh@suse.cz>
3308
3309 * config/i386/x86-tune.def: Add documentation for each of the options;
3310 add whitespace.
3311
b0ff06da
JH		 33122013-09-28 Jan Hubicka <jh@suse.cz>
3313
7a5911d3
UB		 3314 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3315 generic.
b0ff06da
JH		 3316 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3317 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3318 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3319 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3320 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3321
c3284718
RS		 33222013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3323
3324 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3325 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3326 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3327 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3328 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3329 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3330 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3331 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3332 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3333 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3334 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3335 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3336 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3337 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3338 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3339 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3340 gimple.h, godump.c, graphite-clast-to-gimple.c,
3341 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3342 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3343 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3344 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3345 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3346 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3347 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3348 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3349 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3350 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3351 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3352 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3353 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3354 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3355 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3356 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3357 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3358 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3359 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3360 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3361 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3362 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3363 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3364 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3365 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3366 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3367 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3368 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3369 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3370 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3371 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3372 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3373 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3374 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3375 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3376 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3377 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3378 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3379 whitespace before "(".
3380
c6285bd7
SL		 33812013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3382
3383 * expr.h (extract_bit_field): Remove packedp parameter.
3384 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3385 from forward declaration.
3386 (store_split_bit_field): Remove packedp arg from calls to
3387 extract_fixed_bit_field.
3388 (extract_bit_field_1): Remove packedp parameter and packedp
3389 argument from recursive calls and calls to extract_fixed_bit_field.
3390 (extract_bit_field): Remove packedp parameter and corresponding
3391 arg to extract_bit_field_1.
3392 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3393 to issue warnings.
3394 (extract_split_bit_field): Remove packedp arg from call to
3395 extract_fixed_bit_field.
3396 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3397 (copy_blkmode_from_reg): Likewise.
3398 (copy_blkmode_to_reg): Likewise.
3399 (read_complex_part): Likewise.
3400 (store_field): Likewise.
3401 (expand_expr_real_1): Likewise.
3402 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3403 to extract_bit_field.
3404 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3405 call to extract_bit_field.
3406 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3407 call to extract_bit_field.
3408 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3409 and special packedp behavior from -fstrict-volatile-bitfields
3410 documentation.
3411
522d4efc
JBG		 34122013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3413
3414 * lra-eliminations.c (init_elim_table): Guard value_p.
3415
19cdb489
MM		 34162013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3417
3418 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3419 DFmode, DImode, and SFmode in the upper VSX registers based on the
3420 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3421 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3422 of the VSX register handling.
3423
3424 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3425 (f32_sv): Likewise.
3426 (zero_extendsidi2_lfiwzx): Add support for loading into the
3427 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3428 only do VSX memory options on Altivec registers.
3429 (extendsidi2_lfiwax): Likewise.
3430 (extendsfdf2_fpr): Likewise.
3431 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3432 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3433 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3434 (movdi_internal64): Likewise.
3435
d6d11272
XDL		 34362013-09-27 Xinliang David Li <davidxl@google.com>
3437
3438 * opts.c (finish_options): Adjust parameters
3439 according to vect cost model.
3440 (common_handle_option): Set dynamic vect cost
3441 model for FDO.
3442 targhooks.c (default_add_stmt_cost): Compute stmt cost
3443 unconditionally.
3444 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3445 Use helper function.
3446 * tree-vectorizer.h (unlimited_cost_model): New function.
3447 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3448 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3449 function.
3450 (vect_enhance_data_refs_alignment): Ditto.
3451 * flag-types.h: New enum.
3452 * common/config/i386/i386-common.c (ix86_option_init_struct):
3453 No need to initialize vect_cost_model flag.
3454 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3455 unconditionally.
3456
ac1857a3
DN		 34572013-09-27 Diego Novillo <dnovillo@google.com>
3458
3459 * gimple.h (enum ssa_mode): Remove.
3460
0d0e2af6
PM		 34612013-09-27 Paulo Matos <pmatos@broadcom.com>
3462
3463 * cfgloop.h (number_of_loops): Fix typo in check for null.
3464
09dc585e
JJ		 34652013-09-27 Jakub Jelinek <jakub@redhat.com>
3466
3467 PR middle-end/58551
3468 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3469 are children of outermost saved_cfun's loop, and set it up to
3470 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3471 if loop != loop0 and SESE region contains bbs that belong to loop0.
3472
ec5a3504
RS		 34732013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3474
3475 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3476 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3477 (set_address_segment, set_address_base, set_address_index)
3478 (set_address_disp): Accept the argument unconditionally.
3479 (baseness): Remove must_be_base_p and must_be_index_p checks.
3480 (decompose_normal_address): Classify as much as possible in the
3481 main loop.
3482
f91aec98
RS		 34832013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3484
3485 * cse.c (count_reg_usage): Handle INT_LIST.
3486 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3487 * reginfo.c (reg_scan_mark_refs): Likewise.
3488 * reload1.c (eliminate_regs_1): Likewise.
3489
a9195970
IS		 34902013-09-27 Iain Sandoe <iain@codesourcery.com>
3491
3492 PR middle-end/58547
3493 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3494 signed.
3495
e6f1c509
RB		 34962013-09-27 Richard Biener <rguenther@suse.de>
3497
3498 PR tree-optimization/58459
3499 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3500 restriction not propagating into loops.
3501
84f48495
FW		 35022013-09-26 Florian Weimer <fw@deneb.enyo.de>
3503
3504 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3505 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3506 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3507
74fc8b8a
RB		 35082013-09-26 Richard Biener <rguenther@suse.de>
3509
7a5911d3 3510 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3511
b4ada065
RB		 35122013-09-26 Richard Biener <rguenther@suse.de>
3513
3514 * alias.h (component_uses_parent_alias_set): Rename to ...
3515 (component_uses_parent_alias_set_from): ... this.
3516 * alias.c (component_uses_parent_alias_set): Rename to ...
3517 (component_uses_parent_alias_set_from): ... this and return
3518 the desired parent.
3519 (reference_alias_ptr_type_1): Use the result from
3520 component_uses_parent_alias_set_from instead of stripping
3521 components one at a time.
3522 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3523
78bca40d
AM		 35242013-09-26 Andrew MacLeod <amacleod@redhat.com>
3525
7a5911d3
UB		 3526 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3527 Move prototypes to...
78bca40d
AM		 3528 * tree-ssa-ter.h: New File. Move prototypes here.
3529 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3530 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3531 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3532 from is_replaceable_p.
78bca40d
AM		 3533 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3534 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3535 (process_replaceable): Use ter_is_replaceable_p.
3536 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3537 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3538 newly refactored ssa_is_replaceable_p.
3539 * cfgexpand.c: Include tree-outof-ssa.h.
3540 * ssaexpand.h: Delete.
3541
ff2a63a7
AM		 35422013-09-26 Andrew MacLeod <amacleod@redhat.com>
3543
3544 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3545 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3546 tree-ssa.c
3547 (create_gimple_tmp): Delete.
3548 (get_expr_type, build_assign, build_type_cast): Move to...
3549 * gimple-builder.c: New File.
3550 (get_expr_type): Relocate from gimple.c.
3551 (build_assign, build_type_cast): Change to only create ssanames.
3552 * gimple.h: Move prototypes to...
3553 * gimple-builder.h: New File. Here.
3554 * tree-ssa.h: And here.
3555 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3556 count_uses_and_derefs): Relocate from gimple.c.
3557 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3558 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3559 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3560 gimple_replace_ssa_lhs.
3561 * asan.c: Include gimple-builder.h.
3562 * Makefile.in: Add gimple-builder.o.
3563
a2544177
RB		 35642013-09-26 Richard Biener <rguenther@suse.de>
3565
3566 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3567 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3568 (loe_visit_block): Use gcc_checking_assert.
3569 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3570 gimple_assign_ssa_name_copy_p.
3571 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3572 change.
3573
6d1ca9a0
DE		 35742013-09-26 David Edelsohn <dje.gcc@gmail.com>
3575
3576 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3577 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3578
46e950db
RB		 35792013-09-26 Richard Biener <rguenther@suse.de>
3580
3581 PR tree-optimization/58539
3582 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3583 the fact that debug statements are not taking part in loop-closed
3584 SSA construction.
3585
40ada30a
NC		 35862013-09-26 Nick Clifton <nickc@redhat.com>
3587
3588 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3589 time warning message.
3590 (msp430_print_operand_raw): Delete unused letter parameter.
3591 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3592 (msp430_print_operand_address): New function.
3593 (msp430_print_operand): Move address printing code from here to
3594 new function.
3595 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3596 assembler.
3597 (zero_extendpsisi2): Likewise.
3598 (extendpsisi2): New pattern.
3599 (andneghi3): New pattern.
3600
38e8f663
YR		 36012013-09-26 Yvan Roux <yvan.roux@linaro.org>
3602
3603 * config/aarch64/aarch64.opt (mlra): New option.
3604 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3605 (TARGET_LRA_P): Define.
3606
4167e8d4
EB		 36072013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3608
3609 * expr.c (expand_assignment): Remove obsolete comment.
3610
5254eac4
JL		 36112013-09-25 Jeff Law <law@redhat.com>
3612
3613 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3614 tree-ssa-threadupdate.h.
3615 (register_jump_thread): Similarly.
3616 * tree-ssa-threadupdate.h: New header file.
3617 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3618 * tree-vrp.c: Likewise.
3619 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3620 (thread_around_empty_blocks): Change type of path vector argument to
3621 an edge,type pair from just an edge. Initialize both elements when
3622 appending to a jump threading path. Tweak references to elements
3623 appropriately.
3624 (thread_across_edge): Similarly. Release memory for the elements
3625 as needed.
3626 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3627 (dump_jump_thread_path): New function broken out from
3628 register_jump_thread.
3629 (register_jump_thread): Use dump_jump_thread_path. Change type of
3630 path vector entries. Search the path for NULL edges and dump
3631 the path if one is found. Tweak the conversion of path to 3-edge
3632 form to use the block copy type information embedded in the path.
3633
20f114a3
YR		 36342013-09-25 Yvan Roux <yvan.roux@linaro.org>
3635
3636 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3637
3936bafc
YR		 36382013-09-25 Yvan Roux <yvan.roux@linaro.org>
3639 Vladimir Makarov <vmakarov@redhat.com>
3640
3641 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3642 from the least significant bit.
3643 (strip_address_mutations): Add bitfield operations handling.
3644 (must_be_index_p): Add shifting and rotate operations handling.
3645 (set_address_base): Use must_be_base_p predicate.
4167e8d4 3646 (set_address_index): Use must_be_index_p predicate.
3936bafc 3647
b86f6e9e
AI		 36482013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3649 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3650 Sergey Lega <sergey.s.lega@intel.com>
3651 Anna Tikhonova <anna.tikhonova@intel.com>
3652 Ilya Tocar <ilya.tocar@intel.com>
3653 Andrey Turetskiy <andrey.turetskiy@intel.com>
3654 Ilya Verbin <ilya.verbin@intel.com>
3655 Kirill Yukhin <kirill.yukhin@intel.com>
3656 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3657
3658 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3659 Use new names.
3660 (ix86_expand_vector_move_misalign): Support new unaligned load and
3661 stores and use new names.
3662 (CODE_FOR_sse2_storedqu): Rename to ...
3663 (CODE_FOR_sse2_storedquv16qi): ... this.
3664 (CODE_FOR_sse2_loaddqu): Rename to ...
3665 (CODE_FOR_sse2_loaddquv16qi): ... this.
3666 (CODE_FOR_avx_loaddqu256): Rename to ...
3667 (CODE_FOR_avx_loaddquv32qi): ... this.
3668 (CODE_FOR_avx_storedqu256): Rename to ...
3669 (CODE_FOR_avx_storedquv32qi): ... this.
3670 * config/i386/i386.md (fpint_logic): New.
3671 * config/i386/sse.md (VMOVE): Extend for AVX512.
3672 (VF): Ditto.
3673 (VF_128_256): New.
3674 (VF_512): Ditto.
3675 (VI_UNALIGNED_LOADSTORE): Ditto.
3676 (sse2_avx_avx512f): Ditto.
3677 (sse2_avx2): Extend for AVX512.
3678 (sse4_1_avx2): Ditto.
3679 (avx2_avx512f): New.
3680 (sse): Extend for AVX512.
3681 (sse2): Ditto.
3682 (sse4_1): Ditto.
3683 (avxsizesuffix): Ditto.
3684 (sseintvecmode): Ditto.
3685 (ssePSmode): Ditto.
3686 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3687 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3688 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3689 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3690 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3691 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3692 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3693 (STORENT_MODE): Extend for AVX512.
3694 (*absneg<mode>2): Replace constraint "x" with "v".
3695 (*mul<mode>3): Ditto.
3696 (*ieee_smin<mode>3): Ditto.
3697 (*ieee_smax<mode>3): Ditto.
3698 (avx_cmp<mode>3): Replace VF with VF_128_256.
3699 (*<sse>_maskcmp<mode>3_comm): Ditto.
3700 (<sse>_maskcmp<mode>3): Ditto.
3701 (<sse>_andnot<mode>3): Extend for AVX512.
3702 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3703 (<code><mode>3, fpint_logic): New.
3704 (*<code><mode>3): Extend for AVX512.
3705 (avx512flogicsuff): New.
3706 (avx512f_<logic><mode>): Ditto.
3707 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3708 VF_128_256.
3709 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3710 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3711 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3712 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3713 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3714 (xop_vpermil2<mode>3): Ditto.
3715 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3716 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3717 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3718 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3719
4d44d03c
TT		 37202013-09-25 Tom Tromey <tromey@redhat.com>
3721
3722 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3723 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3724 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3725 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3726 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3727 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3728 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3729 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3730 (GRAPHITE_HTAB_H): Remove.
3731
86aaf68d
TT		 37322013-09-25 Tom Tromey <tromey@redhat.com>
3733
3734 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3735
c987deb8
TT		 37362013-09-25 Tom Tromey <tromey@redhat.com>
3737
3738 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3739
9439747e
TT		 37402013-09-25 Tom Tromey <tromey@redhat.com>
3741
3742 * config/i386/t-i386 (i386.o): Remove.
3743 (i386-c.o): Use COMPILE and POSTCOMPILE.
3744
aefc31a1
TT		 37452013-09-25 Tom Tromey <tromey@redhat.com>
3746
3747 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3748
b6541edc
TT		 37492013-09-25 Tom Tromey <tromey@redhat.com>
3750
3751 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3752 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3753 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3754 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3755 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3756 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3757 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3758 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3759 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3760 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3761 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3762 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3763 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3764 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3765 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3766 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3767 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3768 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3769 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3770 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3771 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3772 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3773 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3774 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3775 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3776 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3777 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3778 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3779 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3780 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3781 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3782 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3783 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3784 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3785 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3786 (graphite-clast-to-gimple.o, graphite-dependences.o)
3787 (graphite-interchange.o, graphite-poly.o)
3788 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3789 (graphite-optimize-isl.o, tree-vect-loop.o)
3790 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3791 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3792 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3793 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3794 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3795 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3796 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3797 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3798 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3799 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3800 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3801 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3802 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3803 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3804 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3805 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3806 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3807 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3808 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3809 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3810 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3811 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3812 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3813 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3814 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3815 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3816 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3817 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3818 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3819 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3820 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3821 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3822 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3823 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3824 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3825 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3826 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3827 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3828 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3829 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3830 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3831 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3832 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3833 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3834 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3835 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3836 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3837 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3838 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3839 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3840 (gcov-dump.o): Remove.
3841 (default-c.o): Use COMPILE and POSTCOMPILE.
3842 (CFLAGS-gcc.o): New variable.
3843 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3844
8c796d19
TT		 38452013-09-25 Tom Tromey <tromey@redhat.com>
3846
3847 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3848 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3849 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3850 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3851 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3852 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3853 (c-family/array-notation-common.o, c-family/stub-objc.o)
3854 (c-family/c-ubsan.o): Remove.
3855
0a6c2227
TT		 38562013-09-25 Tom Tromey <tromey@redhat.com>
3857
3858 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3859
f3bc55f0
TT		 38602013-09-25 Tom Tromey <tromey@redhat.com>
3861
3862 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3863 to add -DENABLE_SHARED_LIBGCC.
3864 (gcc.o): Don't use subshell.
3865
c11c795e
TT		 38662013-09-25 Tom Tromey <tromey@redhat.com>
3867
3868 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3869 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3870 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3871 * configure, config.in: Rebuild.
3872
043378c3
TT		 38732013-09-25 Tom Tromey <tromey@redhat.com>
3874
3875 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3876 (COMPILE, POSTCOMPILE): New variables.
3877 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3878 (DEPFILES): New variable.
3879 Include ".Po" files.
3880 * configure.ac: Add checks for dependency checking.
3881 * configure, aclocal.m4: Regenerate.
3882
eab34643
TT		 38832013-09-25 Tom Tromey <tromey@redhat.com>
3884
3885 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 3886 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 3887
d2db5e91
TT		 38882013-09-25 Tom Tromey <tromey@redhat.com>
3889
3890 * Makefile.in (generated_files): Add options.h,
3891 target-hooks-def.h, insn-opinit.h,
3892 common/common-target-hooks-def.h, pass-instances.def,
3893 c-family/c-target-hooks-def.h.
3894
af33629e
JL		 38952013-09-25 Jeff Law <law@redhat.com>
3896
3897 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3898 than foo[foo.length () - 1] to access last member in a vec.
3899 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3900
ed0d3051
RB		 39012013-09-25 Richard Biener <rguenther@suse.de>
3902
3903 PR middle-end/58521
3904 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3905
123485ca
JH		 39062013-09-25 Jan Hubicka <jh@suse.cz>
3907
3908 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3909 test.
3910
d70b0f1f
MP		 39112013-09-25 Marek Polacek <polacek@redhat.com>
3912
3913 PR sanitizer/58420
3914 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3915 when determining the type name.
3916
e3f02798
OE		 39172013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3918
3919 * config/sh/sh.md: Fix formatting.
3920
7a5911d3 39212013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 3922
7a5911d3
UB		 3923 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3924 max peel iterations parameter.
4f17aa0b
XDL		 3925 * param.def: New parameter.
3926 * doc/invoke.texi: Document New parameter.
3927
0498471b
CL		 39282013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3929
3930 * gimple-pretty-print.c: Various whitespace tweaks.
3931 * tree-core.h: Likewise.
3932 * tree-pretty-print.c: Likewise.
3933 * tree-ssa-alias.c: Likewise.
3934 * tree-ssa-copy.c: Likewise.
3935 * tree-ssanames.c: Likewise.
3936 * tree-ssanames.h: Likewise.
3937 * tree-vrp.c: Likewise.
3938
5a8c2b57
AM		 39392013-09-24 Alan Modra <amodra@gmail.com>
3940
3941 PR middle-end/57134
3942 PR middle-end/57586
7a5911d3
UB		 3943 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3944 for output operands that disallow regs. Don't use EXPAND_WRITE on
3945 inout operands.
5a8c2b57 3946
f40333af
RB		 39472013-09-24 Richard Biener <rguenther@suse.de>
3948
3949 PR middle-end/58513
3950 * tree.c (reference_alias_ptr_type): Move ...
3951 * alias.c (reference_alias_ptr_type): ... here and implement
3952 in terms of the new reference_alias_ptr_type_1.
3953 (ref_all_alias_ptr_type_p): New helper.
3954 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3955 use ref_all_alias_ptr_type_p.
3956 (get_deref_alias_set): Add flag_strict_aliasing check here.
3957 (reference_alias_ptr_type_1): New function, split out from ...
3958 (get_alias_set): ... here.
3959 (alias_ptr_types_compatible_p): New function.
3960 * alias.h (reference_alias_ptr_type): Declare.
3961 (alias_ptr_types_compatible_p): Likewise.
3962 * tree.h (reference_alias_ptr_type): Remove.
3963 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3964 to compare MEM_REF alias types.
3965
583e8bf5
RB		 39662013-09-24 Richard Biener <rguenther@suse.de>
3967
3968 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3969
5e8586d7
MM		 39702013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3971
19c3e797
MM		 3972 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3973 reload helper function arrays into a single array reg_addr.
3974 (reload_fpr_gpr): Likewise.
3975 (reload_gpr_vsx): Likewise.
3976 (reload_vsx_gpr): Likewise.
3977 (struct rs6000_reg_addr): Likewise.
3978 (reg_addr): Likewise.
3979 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3980 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3981 (rs6000_init_hard_regno_mode_ok): Likewise.
3982 (rs6000_secondary_reload_direct_move): Likewise.
3983 (rs6000_secondary_reload): Likewise.
3984
5e8586d7
MM		 3985 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3986 constraints: wu, ww, and wy. Repurpose wv constraint added during
3987 power8 changes. Put wg constraint in alphabetical order.
3988
3989 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3990 for future work to add ISA 2.07 VSX single precision support.
3991 (-mvsx-scalar-double): Change default from -1 to 1, update
3992 documentation comment.
3993 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3994 (-mupper-regs-df): New debug switch to control whether DF values
3995 can go in the traditional Altivec registers.
3996 (-mupper-regs-sf): New debug switch to control whether SF values
3997 can go in the traditional Altivec registers.
3998
3999 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4000 and wy constraints.
4001 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4002 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4003 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4004 (rs6000_debug_legitimate_address_p): Print if we are running
4005 before, during, or after reload.
4006 (rs6000_secondary_reload): Add a comment.
4007 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4008
4009 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4010 constraints. Update documentation string.
4011 (wd constraint): Likewise.
4012 (wf constraint): Likewise.
4013 (wg constraint): Likewise.
4014 (wn constraint): Likewise.
4015 (ws constraint): Likewise.
4016 (wt constraint): Likewise.
4017 (wx constraint): Likewise.
4018 (wz constraint): Likewise.
4019 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4020 instructions.
4021 (ww constraint): Likewise.
4022 (wy constraint): Likewise.
4023 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
4024 the previous submissions.
4025 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4026
e5af9ddd
RS		 40272013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4028
4029 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4030 (REG_BR_PROB): Say that the probability is stored in an int_list.
4031 * reg-notes.def: Update commentary to mention INT_LIST.
4032 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4033 (INT_LIST): New rtx.
4034 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4035 * rtlanal.c (int_reg_note_p): New function.
4036 (alloc_reg_note): Assert that the note does not have an int argument.
4037 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4038 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4039 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4040 rather than an INSN_LIST. Handle INT_LIST.
4041 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4042 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4043 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4044 Manipulate them as ints rather than rtxes.
4045 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4046 * regmove.c (copy_src_to_dest): Likewise.
4047 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4048
4049 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4050 into the cases that need it.
4051 * config/arm/arm.c (arm_unwind_emit): Likewise.
4052
4053 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4054 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4055 * loop-doloop.c (add_test, doloop_modify): Likewise.
4056 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4057 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4058 * predict.c (combine_predictions_for_insn): Likewise.
4059 * print-rtl.c (print_rtx): Handle INT_LIST.
4060 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4061 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4062 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4063 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4064 (ix86_split_fp_branch, predict_jump): Likewise.
4065 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4066 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4067 * config/spu/spu.c (ea_load_store_inline): Likewise.
4068
4069 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4070 value of a REG_BR_PROB note.
4071 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4072 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4073 * emit-rtl.c (try_split): Likewise.
4074 * predict.c (br_prob_note_reliable_p): Likewise.
4075 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4076 * reorg.c (mostly_true_jump): Likewise.
4077 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4078 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4079 * config/i386/i386.c (ix86_print_operand): Likewise.
4080 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4081 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4082 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4083 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4084 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4085 * config/sparc/sparc.c (output_cbranch): Likewise.
4086 * config/spu/spu.c (get_branch_target): Likewise.
4087 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4088 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4089
450ad0cd
JH		 40902013-09-23 Jan Hubicka <jh@suse.cz>
4091
4092 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4093 for ipa-devirt.
4094 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4095 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4096 of external calls
4097 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4098 (gimple_fold_call): Dump inconsistent devirtualizations; add
4099 sanity check for type based devirtualizations.
4100 * ipa-prop.c: Include ipa-utils.h
4101 (ipa_intraprocedural_devirtualization): Add sanity check.
4102 (try_make_edge_direct_virtual_call): Likewise.
4103
cefce769
EB		 41042013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4105
4106 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4107 assignment statements.
4108
a895a2b8
KV		 41092013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4110
4111 * gimple-pretty-print.c (dump_ssaname_info): New function.
4112 (dump_gimple_phi): Call it.
4113 (pp_gimple_stmt_1): Likewise.
4114 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4115 (range_info_def): Declare.
4116 * tree-pretty-print.c (pp_double_int): New function.
4117 (dump_generic_node): Call it.
4118 * tree-pretty-print.h (pp_double_int): Declare.
4119 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4120 * tree-ssanames.h (range_info_def): New structure.
4121 (value_range_type): Move definition here.
4122 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4123 Declare.
4124 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4125 initialization.
4126 (set_range_info): New function.
4127 (get_range_info): Likewise.
4128 (duplicate_ssa_name_range_info): Likewise.
4129 (duplicate_ssa_name_fn): Check pointer type and call
4130 duplicate_ssa_name_range_info.
4131 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4132 * tree-vrp.c (value_range_type): Remove definition, now in
4133 tree-ssanames.h.
7a5911d3 4134 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV		 4135 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4136 (SSA_NAME_RANGE_INFO): New macro.
4137
984af6ac
RB		 41382013-09-23 Richard Biener <rguenther@suse.de>
4139
4140 PR tree-optimization/58464
4141 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 4142 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB		 4143 (phi_translate_1): Remove recursion detection here.
4144 (phi_translate): Pre-seed the cache with NULL to catch
4145 recursion here in a more generic way.
4146 (bitmap_find_leader): Adjust comment.
4147 (get_representative_for): Dump value-numbers.
4148 (create_expression_by_pieces): Likewise.
4149 (insert_into_preds_of_block): Likewise.
4150
0e26cf79
CB		 41512013-09-23 Christian Bruel <christian.bruel@st.com>
4152
4153 PR target/58475
4154 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4155 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4156
4d0e5b54
JG		 41572013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4158
4159 Revert r202780:
4160 2013-09-20 Renlin Li <renlin.li@arm.com>
4161
7a5911d3
UB		 4162 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4163 plus_constant.
4d0e5b54
JG		 4164 (aarch64_expand_epilogue): Likewise.
4165 (aarch64_legitimize_reload_address): Likewise.
4166
ca081cc8
EB		 41672013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4168
4169 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4170 NULL_TREE before pushing them onto the vector. Likewise for labels.
4171
8fc53a5f
EB		 41722013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4173
4174 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4175 comparison operators when -fno-trapping-math is in effect.
4176 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4177 comparison operators in TFmode and assert that unsupported operators
4178 cannot reach here.
4179 (ia64_print_operand): Likewise.
4180
41812013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH		 4182
4183 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4184 (sse_typeless_stores): Enable for core
4185 (sse_load0_by_pxor): Likewise.
4186 (four_jump_limit): Disable for core.
4187 (pad_returns): Likewise.
4188 (avoid_vector_decode): Likewise.
4189 (fuse_cmp_and_branch): Enable for cores.
4190 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4191
578e51e6
JDA		 41922013-09-20 John David Anglin <danglin@gcc.gnu.org>
4193
96ce28a3
JDA		 4194 PR middle-end/56791
4195 * config/pa/pa.c (pa_option_override): Disable auto increment and
4196 decrement instructions until reload is completed.
4197
578e51e6
JDA		 4198 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4199 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4200 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4201
6fcd3a13 42022013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 4203 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD		 4204
4205 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4206 (need_to_save): Save bank 0 on interrupts.
4207 (characterize_address): Strip far address wrappers.
4208 (rl78_as_legitimate_address): Likewise.
4209 (transcode_memory_rtx): Likewise.
4210 (rl78_peep_movhi_p): Disable this peephole after devirt.
4211 (rl78_propogate_register_origins): Forget all origins when a
4212 CLOBBER is seen.
4213 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4214 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4215 peephole2's.
4216 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4217 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4218 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4219 relocating.
4220 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4221
4a08db33
JDA		 42222013-09-20 John David Anglin <danglin@gcc.gnu.org>
4223
4224 * config/pa/pa.md: In "scc" insn patterns, change output template to
4225 handle const0_rtx in reg_or_0_operand operands.
4226
7691132c
JBG		 42272013-09-20 Martin Husemann <martin@NetBSD.org>
4228
4229 PR target/56875
4230 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4231 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4232
a3cc13cc
RB		 42332013-09-20 Richard Biener <rguenther@suse.de>
4234
4235 PR middle-end/58484
4236 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4237 remembering SSA name version and block index.
4238 (new_scev_info_str): Adjust.
4239 (hash_scev_info): Likewise. Also hash the block index.
4240 (eq_scev_info): Adjust.
4241 (find_var_scev_info): Likewise.
4242 (struct instantiate_cache_entry): Remove.
4243 (struct instantiate_cache_type): Use a htab to map name, block
4244 to chrec.
4245 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4246 (get_instantiated_value_entry): Likewise.
4247 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4248 (instantiate_scev_name): Adjust.
4249
925f3871
JL		 42502013-09-20 Jeff Law <law@redhat.com>
4251
4252 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4253
2888c331
YZ		 42542013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4255
4256 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4257 Call aarch64_simd_expand_args to update op[argc].
4258
0a811e96
BS		 42592013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4260
7a5911d3
UB		 4261 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4262 plugin argument.
0a811e96 4263
0078f462
BS		 42642013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4265
7a5911d3
UB		 4266 * gengtype.c (file_rules): Added rule for *.cc files.
4267 (get_output_file_with_visibility): Give fatal message when no
4268 rules found.
0078f462 4269
d20e48b0
RL		 42702013-09-20 Renlin Li <renlin.li@arm.com>
4271
4272 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4273 (aarch64_expand_epilogue): Likewise.
4274 (aarch64_legitimize_reload_address): Likewise.
4275
59bb1544
BE		 42762013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4277
4278 PR middle-end/57748
4279 * expr.c (expand_assignment): Remove misalignp code path.
4280
a5430019
MP		 42812013-09-20 Marek Polacek <polacek@redhat.com>
4282
4283 PR sanitizer/58413
4284 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4285 TYPE_PRECISION. Add asserts.
4286
9fed7f3a
RB		 42872013-09-20 Richard Biener <rguenther@suse.de>
4288
4289 PR tree-optimization/58453
4290 * tree-loop-distribution.c (distribute_loop): Apply the cost
4291 model for -ftree-loop-distribute-patterns, too.
4292
a820c834
RB		 42932013-09-20 Richard Biener <rguenther@suse.de>
4294
4295 PR middle-end/58473
4296 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4297 make type comparison less strict.
4298
3cbe17f7
AM		 42992013-09-20 Alan Modra <amodra@gmail.com>
4300
4301 * configure: Regenerate.
4302 * aclocal.m4: Regenerate.
4303
87fccdbb
MP		 43042013-09-20 Marek Polacek <polacek@redhat.com>
4305
4306 PR other/58467
4307 * doc/extend.texi: Document that attribute used is meant to be used
4308 on variables with static storage duration.
4309
8928eff3
JJ		 43102013-09-19 Jakub Jelinek <jakub@redhat.com>
4311
4312 PR tree-optimization/58472
4313 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4314 simd_lane_access set inv_p = false.
4315 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4316 the simduid magic VAR_DECL.
4317
b7b28c53
JH		 43182013-09-19 Jan Hubicka <jh@suse.cz>
4319
4320 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4321
83ae86f5
JL		 43222013-09-17 Jeff Law <law@redhat.com>
4323
4324 * tree-ssa-dom.c (record_temporary_equivalences): New function
4325 split out of dom_opt_dom_walker::after_dom_children.
4326 (dom_opt_dom_walker::thread_across_edge): Move common code
4327 in here from dom_opt_dom_walker::after_dom_children.
4328 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4329
9d532162
JH		 43302013-09-19 Jan Hubicka <jh@suse.cz>
4331
4332 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4333 (TARGET_GENERIC): Use PROCESOR_GENERIC
4334 (enum processor_type): Unify generic32 and 64.
4335 * i386.md (cpu): Likewise.
4336 * x86-tune.def (use_leave): Enable for generic32.
4337 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4338 * athlon.md: Change generic64 to generic in all occurences.
4339 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4340 (ix86_target_macros_internal): Likewise.
4341 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB		 4342 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4343 to ..
9d532162
JH		 4344 (generic_memcpy, generic_memset, generic_cost): This one.
4345 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4346 (m_GENERIC32, m_GENERIC64): Remove.
4347 (m_GENERIC): Turn into one flag.
4348 (processor_target): Unify generic tunnings.
4349 (ix86_option_override_internal): Replace generic32/64 by generic.
4350 (ix86_issue_rate): Likewise.
4351 (ix86_adjust_cost): Likewise.
4352
188c7d00
JH		 43532013-09-19 Jan Hubicka <jh@suse.cz>
4354
4355 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4356 of speculative flag.
4357
4befd127
JJ		 43582013-09-19 Jakub Jelinek <jakub@redhat.com>
4359
4360 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4361 GOMP_sections_start, even if !exit_reachable.
4362
8a26ad39
VM		 43632013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4364
4365 * lra-constraints.c (need_for_all_save_p): Use macro
4366 HARD_REGNO_CALL_PART_CLOBBERED.
4367 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4368 set up pseudo conflict hard regs.
4369
09487185
MM		 43702013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4371
4372 PR target/58452
4373 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4374 operands.
4375
080cbf9e
VM		 43762013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4377
4378 PR rtl-optimization/58438
4379 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4380 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4381 from previous subpasses.
4382
a0f70fcb
RE		 43832013-09-18 Richard Earnshaw <rearnsha@arm.com>
4384
4385 * arm.c (arm_get_frame_offsets): Validate architecture supports
4386 LDRD/STRD before accepting the tuning preference.
4387 (arm_expand_prologue): Likewise.
4388 (arm_expand_epilogue): Likewise.
4389
0547c9b6
RB		 43902013-09-18 Richard Biener <rguenther@suse.de>
4391
4392 PR tree-optimization/58417
4393 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4394 have chrecs with symbols defined in the loop as operands.
4395 (chrec_fold_multiply): Likewise.
4396 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4397 parameters before folding binary operations.
4398 (struct instantiate_cache_entry_hasher): Remove.
4399 (struct instantiate_cache_type): Use a pointer-map.
4400 (instantiate_cache_type::instantiate_cache_type): New function.
4401 (instantiate_cache_type::get): Likewise.
4402 (instantiate_cache_type::set): Likewise.
4403 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4404 (get_instantiated_value_entry): Likewise.
4405 (global_cache): New global.
4406 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4407 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4408 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4409 (instantiate_scev_name): Adjust.
4410 (instantiate_scev): Construct global instead of local cache.
4411 (resolve_mixers): Likewise.
4412
309dc1aa
DM		 44132013-09-18 Daniel Morris <danielm@ecoscentric.com>
4414 Paolo Carlini <paolo.carlini@oracle.com>
4415
4416 PR c++/58458
4417 * doc/implement-cxx.texi: Fix references to the C++ standards.
4418
995a1b4a
JJ		 44192013-09-18 Jakub Jelinek <jakub@redhat.com>
4420
4421 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4422 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4423 simd_lane_access drs, update also DR_ALIGNED_TO.
4424
ce6923c5
MP		 44252013-09-18 Marek Polacek <polacek@redhat.com>
4426
4427 PR sanitizer/58411
4428 * doc/extend.texi: Document no_sanitize_undefined attribute.
4429 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4430 no_sanitize_undefined attribute.
4431
d30d00a2
NC		 44322013-09-18 Nick Clifton <nickc@redhat.com>
4433
4434 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4435 (ASM_DECLARE_FUNCTION_NAME): Define.
4436
4d9192b5
TS		 44372013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4438
4439 * compare-elim.c (find_comparison_dom_walker): New class
4440 (find_comparisons_in_bb): Rename to
4441 find_comparison_dom_walker::before_dom_children
4442 (find_comparisons): Adjust
4443 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4444 adjust.
4445 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4446 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4447 (init_walk_dominator_tree): Remove declaration.
4448 (fini_walk_dominator_tree): Remove declaration.
4449 * fwprop.c (single_def_use_dom_walker): New class
4450 (single_def_use_enter_block): Convert to
4451 single_def_use_dom_walker::before_dom_children.
4452 (single_def_use_leave_block): Convert to
4453 single_def_use_dom_walker::after_dom_children.
4454 (build_single_def_use_links): Adjust.
4455 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4456 class.
4457 (find_candidates_in_block): Convert to
4458 find_candidates_dom_walker::before_dom_children.
4459 (execute_strength_reduction): Adjust.
4460 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4461 (sese_dom_walker): New class.
4462 (sese_dom_walker::sese_dom_walker): New constructor.
4463 (sese_dom_walker::~sese_dom_walker): New destructor.
4464 (build_sese_conditions_before): Convert to
4465 sese_dom_walker::before_dom_children.
4466 (build_sese_conditions_after): Convert to
4467 sese_dom_walker::after_dom_children.
4468 (build_poly_scop): Adjust
4469 * tree-into-ssa.c (rewrite_dom_walker): New class
4470 (rewrite_enter_block): Convert to
4471 rewrite_dom_walker::before_dom_children.
4472 (rewrite_leave_block): Convert to
4473 rewrite_dom_walker::after_dom_children.
4474 (rewrite_update_dom_walker): New class.
4475 (rewrite_update_enter_block): Convert to
4476 rewrite_update_dom_walker::before_dom_children.
4477 (rewrite_update_leave_block): Convert to
4478 rewrite_update_dom_walker::after_dom_children.
4479 (rewrite_blocks, rewrite_into_ssa): Adjust.
4480 (mark_def_dom_walker): New class.
4481 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4482 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4483 (mark_def_sites_blocks): Convert to
4484 mark_def_dom_walker::before_dom_children.
4485 (mark_def_site_blocks): Remove.
4486 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4487 (tree_ssa_dominator_optimize): Adjust.
4488 (dom_thread_across_edge): Convert to method
4489 dom_opt_dom_walker::thread_across_edge.
4490 (dom_opt_enter_block): Convert to member function
4491 dom_opt_dom_walker::before_dom_children.
4492 (dom_opt_leave_block): Convert to member function
4493 dom_opt_dom_walker::after_dom_children.
4494 * tree-ssa-dse.c (dse_dom_walker): New class.
4495 (dse_enter_block): Convert to member function
4496 dse_dom_walker::before_dom_children.
4497 (tree_ssa_dse): Adjust.
4498 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4499 (determine_invariantness_stmt): Convert to method
4500 invariantness_dom_walker::before_dom_children.
4501 (determine_invariantness): Remove
4502 (move_computations_dom_walker): New class.
4503 (move_computations_stmt): Convert to method
4504 move_computations_dom_walker::before_dom_children.
4505 (move_computations, tree_ssa_lim): Adjust.
4506 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4507 (nt_init_block): Make method
4508 notrappping_dom_walker::before_dom_children.
4509 (nt_fini_block): Make
4510 method nontrapping_dom_walker::after_dom_children.
4511 (get_non_trapping): Adjust.
4512 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4513 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4514 (eliminate_leave_block): Make method.
4515 eliminate_dom_walker::after_dom_children.
4516 (eliminate): Adjust
4517 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4518 (strlen_enter_block): Make method
4519 strlen_dom_walker::before_dom_children.
4520 (strlen_leave_block): Make
4521 method strlen_dom_walker::after_dom_children.
4522 (tree_ssa_strlen): Adjust.
4523 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4524 (tree_ssa_uncprop): Adjust.
4525 (uncprop_leave_block): Make method
4526 uncprop_dom_walker::after_dom_children.
4527 (uncprop_leave_block): Make method
4528 uncprop_dom_walker::before_dom_children.
4529
3b3cc26b
BC		 45302013-09-18 Bin Cheng <bin.cheng@arm.com>
4531
7a5911d3
UB		 4532 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4533 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4534
b5ba3a9b
NC		 45352013-09-17 Nick Clifton <nickc@redhat.com>
4536
4537 * config/rl78/rl78-real.md (bf): New pattern.
4538 (bt): New pattern.
4539 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4540 (rl78_print_operand): Do not put a # before a %B.
4541 * config/rl78/rl78.opt: Tweak doc strings.
4542
09d5094b
DD		 45432013-09-17 DJ Delorie <dj@redhat.com>
4544
4545 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4546 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4547 (rl78_option_override): Likewise, if -mallregs.
4548 (is_virtual_register): Likewise.
4549 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4550 (REGNO_OK_FOR_BASE_P): Likewise.
4551 * config/rl78/rl78.opt (-mallregs): New.
4552
2ba1dca1
NC		 45532013-09-17 Nick Clifton <nickc@redhat.com>
4554
4555 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4556 For interrupt functions: save all call clobbered registers if the
4557 interrupt handler is not a leaf function.
4558 (rl78_expand_prologue): Always recompute the frame information.
4559 For interrupt functions: only select bank 0 if one of the bank 0
4560 registers is going to be psuhed.
4561
2e7c3f21
DD		 45622013-09-17 DJ Delorie <dj@redhat.com>
4563
4564 * config/rl78/constraints.md: For each W* constraint, rename to C*
4565 and create a W* constraint that checks for an optional ES: prefix
4566 pattern also.
4567 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4568 (es_addr): New. Used to wrap far addresses.
4569 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4570 (rl78_es_base): New.
4571 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4572 wrapped far addresses.
4573 (rl78_print_operand_1): Unwrap far addresses before processing.
4574 (rl78_lo16): Wrap far addresses in unspecs.
4575 (rl78_es_addr): New.
4576 (rl78_es_base): New.
4577 (insn_ok_now): Check for not-yet-wrapped far addresses.
4578 (transcode_memory_rtx): Properly re-wrap far addresses.
4579
5c26a69a
SH		 45802013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4581
4582 * config/sparc/t-rtems: Add leon3 multilibs.
4583
181f5f3e
CH		 45842013-09-17 Cong Hou <congh@google.com>
4585
4586 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4587 when checking the dot production pattern. The type of rhs operand
4588 of multiply is now checked correctly.
4589
8d34e421
JL		 45902013-09-17 Jeff Law <law@redhat.com>
4591
4592 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4593 edge implied equivalences into successor phis.
4594 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4595 here from tree-ssa-threadedge.c.
4596 (mark_threaded_blocks): When threading through a joiner, if both
4597 successors of the joiner's clone reach the same block, verify the
4598 PHI arguments are equal. If not, cancel the jump threading request.
4599 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4600 tree-ssa-threadupdate.c
4601 (thread_across_edge): Don't check PHI argument equality when
4602 threading through joiner block here.
7a5911d3 4603
c152901f
AM		 46042013-09-17 Andrew MacLeod <amacleod@redhat.com>
4605
4606 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4607 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB		 4608 (warn_uninit, warn_uninitialized_vars,
4609 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4610 Move to tree-ssa-uninit.c.
78cedfb1 4611 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 4612 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM		 4613 ssa_undefined_value_p.
4614 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB		 4615 (warn_uninit, warn_uninitialized_vars,
4616 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4617 Move from tree-ssa.c.
4618 * tree-ssa.h: Adjust prototypes.
c152901f 4619
b631d45a
JH		 46202013-09-17 Jan Hubicka <jh@suse.cz>
4621
4622 PR middle-end/58332
4623 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4624 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4625 FUNCTION_NOT_OPTIMIZED.
4626 * ipa-inline-analysis.c (compute_inline_parameters): Function
4627 not optimized is not inlinable unless it is alwaysinline.
4628 (inline_analyze_function): Force calls in not optimized
4629 function not inlinable.
4630
5b79657a
JH		 46312013-09-17 Jan Hubicka <jh@suse.cz>
4632
4633 PR middle-end/58329
4634 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4635 to return NULL.
4636 * ipa.c (function_and_variable_visibility): Likewise.
4637 * ipa-profile.c (ipa_profile): Likewise.
4638
27442c24
BE		 46392013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4640
4641 PR ipa/58398
4642 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4643 attribute, and don't inline the resolver in this case.
4644
753b9bc3
TJ		 46452013-09-17 Teresa Johnson <tejohnson@google.com>
4646
4647 * coverage.c (get_coverage_counts): Add missing newline.
4648
e3d3cfb4
KT		 46492013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4650
4651 PR tree-optimization/58088
4652 * fold-const.c (mask_with_trailing_zeros): New function.
4653 (fold_binary_loc): Make sure we don't recurse infinitely
4654 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4655 Use mask_with_trailing_zeros where appropriate.
4656
9a0ac98f
YR		 46572013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4658
4659 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4660 of 'prev' var to get better distance estimation.
4661
78bbd765
EB		 46622013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4663
4664 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4665 * tree-inline.c (is_parameter_of): New predicate.
4666 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4667 a parameter has been remapped.
4668 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4669 (optimize_inline_calls): Initialize transform_parameter.
4670 (copy_gimple_seq_and_replace_locals): Likewise.
4671 (tree_function_versioning): Likewise.
4672 (maybe_inline_call_in_expr): Likewise.
4673
cad055a4
NC		 46742013-09-17 Nick Clifton <nickc@redhat.com>
4675
4676 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4677 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4678 interrupt handlers.
4679 (is_attr_func): New function.
4680 (msp430_is_interrupt_func): New function.
4681 (is_naked_func): New function.
4682 (is_reentrant_func): New function.
4683 (is_critical_func): New function.
4684 (msp430_start_function): Add annotations for function attributes.
4685 (msp430_attr): New function.
4686 (msp430_attribute_table): New.
4687 (msp430_function_section): New function.
4688 (TARGET_ASM_FUNCTION_SECTION): Define.
4689 (msp430_builtin): New enum.
4690 (msp430_init_builtins): New function.
4691 (msp430_builtin_devl): New function.
4692 (msp430_expand_builtin): New function.
4693 (TARGET_INIT_BUILTINS): Define.
4694 (TARGET_EXPAND_BUILTINS): Define.
4695 (TARGET_BUILTIN_DECL): Define.
4696 (msp430_expand_prologue): Add support for naked, interrupt,
4697 critical and reentrant functions.
4698 (msp430_expand_epilogue): Likewise.
4699 (msp430_print_operand): Handle 'O' character.
4700 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4701 NO_TRAMPOLINES.
4702 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4703 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4704 (pushm): Use a 'n' rather than an 'i' constraint.
4705 (msp_return): Add generation of the interrupt return instruction.
4706 (disable_interrupts): New pattern.
4707 (enable_interrupts): New pattern.
4708 (push_intr_state): New pattern.
4709 (pop_intr_state): New pattern.
4710 (bic_SR): New pattern.
4711 (bis_SR): New pattern.
4712 * doc/extend.texi: Document MSP430 function attributes and builtin
4713 functions.
4714
deb6c11a
RB		 47152013-09-17 Richard Biener <rguenther@suse.de>
4716
4717 PR tree-optimization/58432
4718 * tree-loop-distribution.c (tree_loop_distribution): Also
4719 scan PHIs for outside loop uses and seed a partition from them.
4720
c068654b
BC		 47212013-09-17 Bin Cheng <bin.cheng@arm.com>
4722
4723 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4724 (restructure_reference): Call backtrace_base_for_ref.
4725
567969e4
AM		 47262013-09-17 Alan Modra <amodra@gmail.com>
4727
4728 PR target/57589
7a5911d3
UB		 4729 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4730 patch.
567969e4 4731
5c0029de
DD		 47322013-09-16 DJ Delorie <dj@redhat.com>
4733
4734 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4735 vregs location for RL78/G10.
4736 (rl78_expand_prologue): Avoid SEL on G10.
4737 (rl78_expand_epilogue): Likewise.
4738 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4739 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4740 __RL78_G10__ when appropriate.
4741 (ASM_SPEC): Pass -mg10 along to the assembler.
4742 * config/rl78/rl78.md (sel_rb): Disable for G10.
4743 * config/rl78/rl78.opt: Add -mg10 option.
4744 * config/rl78/t-rl78: Add -mg10 multilib.
4745
ea0f3e87
XDL		 47462013-09-16 Xinliang David Li <davidxl@google.com>
4747
4748 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4749 * omp-low.c (omp_max_vf): Ditto.
4750 (expand_omp_simd): Ditto.
4751 * tree-vectorizer.c (vectorize_loops): Ditto.
4752 (gate_vect_slp): Ditto.
4753 (gate_increase_alignment): Ditto.
4754 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4755 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4756 (gate_tree_vectorize): Name change.
4757 (tree_vectorize): Ditto.
4758 (pass_vectorize::gate): Call new function.
4759 (pass_vectorize::execute): Ditto.
7a5911d3 4760 * opts.c: O3 default setting change.
ea0f3e87
XDL		 4761 (finish_options): Check new flag.
4762 * doc/invoke.texi: Document new flags.
4763 * common.opt: New flags.
4764
4f3da2e5
AS		 47652013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4766
4767 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4768 (Condition Code): Don't reference it.
4769 * doc/tm.texi: Regenerate.
4770
aa940f7c
VM		 47712013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4772
4773 PR middle-end/58418
4774 * lra-constraints.c (undo_optional_reloads): Consider all optional
4775 reload even if it did not get a hard reg.
4776
e645e942
TJ		 47772013-09-16 Teresa Johnson <tejohnson@google.com>
4778
4779 * dumpfile.c (dump_loc): Remove newline emission.
4780 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4781 emission to dump_printf_loc calls where missing.
4782 (vect_mark_for_runtime_alias_test): Ditto.
4783 (vect_analyze_data_ref_dependence): Ditto.
4784 (vect_analyze_data_ref_dependences): Ditto.
4785 (vect_slp_analyze_data_ref_dependence): Ditto.
4786 (vect_slp_analyze_data_ref_dependences): Ditto.
4787 (vect_compute_data_ref_alignment): Ditto.
4788 (vect_update_misalignment_for_peel): Ditto.
4789 (vect_verify_datarefs_alignment): Ditto.
4790 (vector_alignment_reachable_p): Ditto.
4791 (vect_get_data_access_cost): Ditto.
4792 (vect_enhance_data_refs_alignment): Ditto.
4793 (vect_find_same_alignment_drs): Ditto.
4794 (vect_analyze_data_refs_alignment): Ditto.
4795 (vect_analyze_group_access): Ditto.
4796 (vect_analyze_data_ref_access): Ditto.
4797 (vect_analyze_data_ref_accesses): Ditto.
4798 (vect_prune_runtime_alias_test_list): Ditto.
4799 (vect_analyze_data_refs): Ditto.
4800 (vect_create_addr_base_for_vector_ref): Ditto.
4801 (vect_create_data_ref_ptr): Ditto.
4802 (vect_grouped_store_supported): Ditto.
4803 (vect_grouped_load_supported): Ditto.
4804 * value-prof.c (check_counter): Ditto.
4805 (check_ic_target): Ditto.
4806 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4807 (vect_recog_widen_mult_pattern): Ditto.
4808 (vect_recog_widen_sum_pattern): Ditto.
4809 (vect_recog_over_widening_pattern): Ditto.
4810 (vect_recog_widen_shift_pattern): Ditto.
4811 (vect_recog_rotate_pattern): Ditto.
4812 (vect_recog_vector_vector_shift_pattern): Ditto.
4813 (vect_recog_divmod_pattern): Ditto.
4814 (vect_recog_mixed_size_cond_pattern): Ditto.
4815 (vect_recog_bool_pattern): Ditto.
4816 (vect_pattern_recog_1): Ditto.
4817 (vect_pattern_recog): Ditto.
4818 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4819 (vect_is_simple_iv_evolution): Ditto.
4820 (vect_analyze_scalar_cycles_1): Ditto.
4821 (vect_get_loop_niters): Ditto.
4822 (vect_analyze_loop_1): Ditto.
4823 (vect_analyze_loop_form): Ditto.
4824 (vect_analyze_loop_operations): Ditto.
4825 (vect_analyze_loop_2): Ditto.
4826 (vect_analyze_loop): Ditto.
4827 (report_vect_op): Ditto.
4828 (vect_is_slp_reduction): Ditto.
4829 (vect_is_simple_reduction_1): Ditto.
4830 (vect_get_known_peeling_cost): Ditto.
4831 (vect_estimate_min_profitable_iters): Ditto.
4832 (vect_model_reduction_cost): Ditto.
4833 (vect_model_induction_cost): Ditto.
4834 (get_initial_def_for_induction): Ditto.
4835 (vect_create_epilog_for_reduction): Ditto.
4836 (vectorizable_reduction): Ditto.
4837 (vectorizable_induction): Ditto.
4838 (vectorizable_live_operation): Ditto.
4839 (vect_loop_kill_debug_uses): Ditto.
4840 (vect_transform_loop): Ditto.
4841 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4842 (vect_stmt_relevant_p): Ditto.
4843 (process_use): Ditto.
4844 (vect_mark_stmts_to_be_vectorized): Ditto.
4845 (vect_model_simple_cost): Ditto.
4846 (vect_model_promotion_demotion_cost): Ditto.
4847 (vect_model_store_cost): Ditto.
4848 (vect_get_store_cost): Ditto.
4849 (vect_model_load_cost): Ditto.
4850 (vect_get_load_cost): Ditto.
4851 (vect_init_vector_1): Ditto.
4852 (vect_get_vec_def_for_operand): Ditto.
4853 (vect_finish_stmt_generation): Ditto.
4854 (vectorizable_call): Ditto.
4855 (vectorizable_conversion): Ditto.
4856 (vectorizable_assignment): Ditto.
4857 (vectorizable_shift): Ditto.
4858 (vectorizable_operation): Ditto.
4859 (vectorizable_store): Ditto.
4860 (vectorizable_load): Ditto.
4861 (vectorizable_condition): Ditto.
4862 (vect_analyze_stmt): Ditto.
4863 (vect_transform_stmt): Ditto.
4864 (vect_is_simple_use): Ditto.
4865 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4866 (vect_can_advance_ivs_p): Ditto.
4867 (vect_update_ivs_after_vectorizer): Ditto.
4868 (vect_do_peeling_for_loop_bound): Ditto.
4869 (vect_gen_niters_for_prolog_loop): Ditto.
4870 (vect_update_inits_of_drs): Ditto.
4871 (vect_create_cond_for_alias_checks): Ditto.
4872 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4873 (vect_build_slp_tree_1): Ditto.
4874 (vect_supported_load_permutation_p): Ditto.
4875 (vect_analyze_slp_instance): Ditto.
4876 (vect_analyze_slp): Ditto.
4877 (vect_make_slp_decision): Ditto.
4878 (vect_detect_hybrid_slp): Ditto.
4879 (vect_bb_vectorization_profitable_p): Ditto.
4880 (vect_slp_analyze_bb_1): Ditto.
4881 (vect_update_slp_costs_according_to_vf): Ditto.
4882 (vect_get_mask_element): Ditto.
4883 (vect_transform_slp_perm_load): Ditto.
4884 (vect_schedule_slp_instance): Ditto.
4885 (vect_schedule_slp): Ditto.
4886 (vect_slp_transform_bb): Ditto.
4887 * profile.c (read_profile_edge_counts): Ditto.
4888 (compute_branch_probabilities): Ditto.
4889 * coverage.c (get_coverage_counts): Ditto.
4890
46f851f3
DN		 48912013-09-16 Diego Novillo <dnovillo@google.com>
4892
7a5911d3 4893 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 4894
1632a686
JH		 48952013-09-16 Jan Hubicka <jh@suse.cz>
4896
4897 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 4898 abstract functions; for static functions check the presence of body.
1632a686 4899
828e70c1
JG		 49002013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4901
4902 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4903 * config/aarch64/aarch64-simd.md
4904 (aarch64_mla_elt<mode>): New.
4905 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4906 (aarch64_mls_elt<mode>): Likewise.
4907 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4908 (aarch64_fma4_elt<mode>): Likewise.
4909 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4910 (aarch64_fma4_elt_to_128v2df): Likewise.
4911 (aarch64_fma4_elt_to_64df): Likewise.
4912 (fnma<mode>4): Likewise.
4913 (aarch64_fnma4_elt<mode>): Likewise.
4914 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4915 (aarch64_fnma4_elt_to_128v2df): Likewise.
4916 (aarch64_fnma4_elt_to_64df): Likewise.
4917 * config/aarch64/iterators.md (VDQSF): New.
4918 * config/aarch64/arm_neon.h
4919 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4920 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4921
779aea46
JG		 49222013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4923
4924 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4925 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4926 (aarch64_mul3_elt_to_128df): Likewise.
4927 (aarch64_mul3_elt_to_64v2df): Likewise.
4928 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4929 (VMUL): New.
4930 (VMUL_CHANGE_NLANES) Likewise.
4931 (h_con): Likewise.
4932 (f): Likewise.
4933 * config/aarch64/arm_neon.h
4934 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4935
a407a750
JG		 49362013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4937
4938 * config/aarch64/arm_neon.h
4939 (vcvtx_high_f32_f64): Fix parameters.
4940
f77e2d2d 49412013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 4942 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 4943
a5bd71fa 4944 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 4945
36875e8f
RB		 49462013-09-16 Richard Biener <rguenther@suse.de>
4947
4948 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4949 (dot_rdg_1): Handle control_dd.
4950 (create_edge_for_control_dependence): New function.
4951 (create_rdg_edges): Add control dependences if asked for.
4952 (build_rdg): Likewise.
4953 (generate_loops_for_partition): If there are not necessary
4954 control stmts remove all their dependencies.
4955 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4956 (distribute_loop): Pass on control dependences.
4957 (tree_loop_distribution): Compute control dependences and remove
4958 restriction on number of loop nodes.
4959
96e24d49
JJ		 49602013-09-16 Jakub Jelinek <jakub@redhat.com>
4961
4962 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4963 for internal calls.
4964
40dbb05c
RS		 49652013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4966
4967 * cse.c (try_const_anchors): Punt on CC modes.
4968
97191ad0
JBG		 49692013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4970
4971 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4972
403a3fb7
JDA		 49732013-09-14 John David Anglin <danglin@gcc.gnu.org>
4974
4975 PR target/58382
4976 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4977 calls to word_mode.
4978
fb5610fb
IS		 49792013-09-14 Iain Sandoe <iain@codesourcery.com>
4980
4981 PR target/48094
4982 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG		 4983 seen.
4984 (darwin_objc1_section): Likewise.
fb5610fb
IS		 4985 (darwin_file_end): Emit Image Info section when required.
4986
afdec9bd
JH		 49872013-09-14 Jan Hubicka <jh@suse.cz>
4988
4989 * tree-into-ssa.c (gate_into_ssa): New.
4990 (pass_data_build_ssa): Use it.
4991 * cgraph.h (expand_thunk): Update prototype.
4992 * cgraphunit.c (analyze_function): Expand thunks early.
4993 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4994 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4995 set lowered flag; do not add new function.
4996 (assemble_thunks_and_aliases): Update.
4997 * tree-ssa.c (gate_init_datastructures): New gate.
4998 (pass_data_init_datastructures): Use it.
4999
492d1e0a
IS		 50002013-09-14 Iain Sandoe <iain@codesourcery.com>
5001
5002 PR target/58269
f77e2d2d 5003 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS		 5004 xmm register set described in the psABI.
5005
d13c6b70
EG		 50062013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5007
f77e2d2d
JBG		 5008 * dwarf2out.c (should_emit_struct_debug): Add check
5009 for type_decl variable is not NULL.
d13c6b70 5010
88b0e79e
JC		 50112013-09-13 Jacek Caban <jacek@codeweavers.com>
5012
5013 * config.gcc: Use new winnt-c.c target hooks
5014 * config/t-winnt: New file
5015 * config/winnt-c.c: New file
5016 * doc/tm.texi.in: Document new hook
5017 * doc/tm.texi: Regenerated
5018
19ba6aab
JH		 50192013-09-13 Jan Hubicka <jh@suse.cz>
5020
5021 PR middle-end/58094
5022 * ipa-inline.c (check_callers): New function.
5023 (check_caller_edge): Remove.
5024 (want_inline_function_to_all_callers_p): Also permit alises that are
5025 called dirrectly.
5026 (inline_to_all_callers): Terminate the walk when devirtualization
5027 introduce new calls.
5028
a93c18c8
JH		 50292013-09-13 Jan Hubicka <jh@suse.cz>
5030
5031 * ipa-inline-analysis.c (struct growth_data): Add node.
5032 (do_estimate_growth_1): Fix detection of recursion.
5033
f3b331d1
JJ		 50342013-09-13 Jakub Jelinek <jakub@redhat.com>
5035
5036 PR tree-optimization/58392
5037 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5038 to avoid shadowing of outer loop variable. If
5039 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5040 replace_by_duplicate_decl simduid of loops that have it set and
5041 set dest_cfun->has_simduid_loops and/or
5042 dest_cfun->has_force_vect_loops.
5043 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5044 instead of maybe_lookup_decl.
5045 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5046 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5047 call. Copy over force_vect and copy and remap simduid. Set
5048 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5049 (copy_cfg_body): Remove blocks_to_copy argument. Use
5050 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5051 caller. Don't set cfun->has_simduid_loops and/or
5052 cfun->has_force_vect_loops here.
5053 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5054 caller.
5055 (expand_call_inline, tree_function_versioning): Adjust copy_body
5056 callers.
5057
ae6d0907
MJ		 50582013-09-13 Martin Jambor <mjambor@suse.cz>
5059
5060 PR bootstrap/58388
5061 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5062 the assert if the edge was a speculative one.
5063
2fd5894f
RB		 50642013-09-13 Richard Biener <rguenther@suse.de>
5065
5066 * tree-data-ref.h (known_dependences_p): Move here ...
5067 * tree-loop-distribution.c (known_dependences_p): ... from here.
5068 (dump_rdg_component, debug_rdg_component): Remove.
5069 (dump_rdg): Adjust.
5070 (generate_loops_for_partition): Use gimple_uid instead of
5071 relying on matching stmt visit order.
5072 (rdg_build_partitions): Take starting stmt vector.
5073 (ldist_gen): Merge into ...
5074 (distribute_loop): ... this function. Do not compute starting
5075 vertices vector.
5076 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5077
86eb4bd7
KT		 50782013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5079
5080 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5081 Set type attribute correctly. Set predicable_short_it attribute.
5082 (cmpsi_shiftsi): Remove %? from output template.
5083
83a95546
RB		 50842013-09-13 Richard Biener <rguenther@suse.de>
5085
5086 * tree-loop-distribution.c (struct rdg_component,
5087 rdg_defs_used_in_other_loops_p, free_rdg_components,
5088 rdg_build_components): Remove.
5089 (stmts_from_loop): Do not record virtual PHIs.
5090 (generate_loops_for_partition): Skip virtual PHIs.
5091 (build_rdg_partition_for_component): Rename to ...
5092 (build_rdg_partition_for_vertex): ... this and adjust.
5093 (rdg_build_partitions): Take a vector of starting vertices
5094 instead of components. Remove unnecessary leftover handling.
5095 (ldist_gen): Do not build components or record other stores.
5096 (distribute_loop): Do not distribute loops containing stmts
5097 with side-effects.
5098
c0ecf7f1
CB		 50992013-09-13 Christian Bruel <christian.bruel@st.com>
5100
5101 PR target/58314
5102 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5103
34f56efc
KT		 51042013-09-13 Kai Tietz <ktietz@redhat.com>
5105
740f2bf3
KT		 5106 * config.gcc: Separate cases for mingw and cygwin targets,
5107 and add 64-bit cygwin target case.
5108
34f56efc
KT		 5109 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5110 dll-export inline-functions.
5111 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5112
5113
51142013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL		 5115
5116 PR middle-end/58387
5117 Revert:
5118 2013-09-06 Jeff Law <law@redhat.com>
5119
5120 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5121 edge implied equivalences into successor phis.
5122
aadb52ad
DD		 51232013-09-12 DJ Delorie <dj@redhat.com>
5124
5125 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5126 separators.
5127
90ae7019
BM		 51282013-09-12 Brooks Moses <bmoses@google.com>
5129
5130 PR driver/42955
5131 * Makefile.in: Do not install driver binaries in $(target)/bin.
5132
17b2557c
DD		 51332013-09-12 DJ Delorie <dj@redhat.com>
5134
c5215a9b
DD		 5135 * config/rl78/rl78.opt (mrelax): New.
5136 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5137 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5138
17b2557c
DD		 5139 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5140 between SP and FP.
5141 (rl78_expand_epilogue): Likewise.
5142
ecafcf05
VM		 51432013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5144
5145 PR middle-end/58335
84f6d9fa 5146 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM		 5147 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5148 pointer elimination with using remove_reg_equal_offset_note.
5149
8369f38a
DD		 51502013-09-12 DJ Delorie <dj@redhat.com>
5151
f6a83b4a
DD		 5152 * config/msp430/: New port.
5153 * config.gcc (msp430): Added.
5154 * doc/invoke.texi: Document MSP430 options.
5155 * doc/install.texi: Document msp430-elf
5156 * doc/md.texi: Document msp430-elf
5157 * doc/contrib.texi: Document msp430-elf
5158
8369f38a
DD		 5159 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5160 PSImode.
5161
a854f856
MJ		 51622013-09-12 Martin Jambor <mjambor@suse.cz>
5163
5164 PR ipa/58389
5165 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5166 reference descriptor is NULL.
5167 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5168 edge to NULL.
5169
b12ebd96
AM		 51702013-09-12 Andrew MacLeod <amacleod@redhat.com>
5171
5172 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5173 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5174 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5175 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5176 make_temp_ssa_name): move to tree-ssanames.h
5177 * tree-ssa-alias.h: Move prototype.
5178 * tree-ssa.h: Include tree-ssanames.h.
5179 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5180 * tree-ssanames.h: New. Move items from tree-flow*.h
5181 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5182
e1fb4ad3
RB		 51832013-09-12 Richard Biener <rguenther@suse.de>
5184
5185 PR tree-optimization/58404
5186 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5187 propagate non-invariant addresses into dereferences wrapped
5188 in component references.
5189
154c9188
RB		 51902013-09-12 Richard Biener <rguenther@suse.de>
5191
5192 PR tree-optimization/58402
5193 * passes.def: Move pass_late_warn_uninitialized later.
5194
7a300452
AM		 51952013-09-12 Andrew MacLeod <amacleod@redhat.com>
5196
5197 * tree-ssa.h: New. Move content from tree-flow.h and
5198 tree-flow-inline.h.
5199 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5200 Move prototypes belonging to tree-ssa.c.
f77e2d2d 5201 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM		 5202 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5203 tree-ssa.h.
5204 * gimple.h: Adjust prototypes.
f77e2d2d
JBG		 5205 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5206 to...
5207 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM		 5208 * tree.h: Move prototype to tree-ssa.h.
5209 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5210 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5211 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5212 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5213 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5214 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5215 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5216 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG		 5217 graphite-blocking.c, graphite-clast-to-gimple.c,
5218 graphite-dependences.c, graphite-interchange.c,
5219 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5220 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5221 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5222 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM		 5223 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5224 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5225 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5226 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5227 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5228 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5229 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5230 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5231 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5232 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5233 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5234 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5235 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5236 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5237 value-prof.c, var-tracking.c,
5238 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5239
97463b2b
RB		 52402013-09-12 Richard Biener <rguenther@suse.de>
5241
5242 PR tree-optimization/58396
5243 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5244 (build_rdg): Take a loop-nest parameter, fix memleaks.
5245 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5246
80ecd89d
YR		 52472013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5248
5249 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5250 for SLM.
5251
893e85fa
CM		 52522013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5253
5254 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5255 The FMA instruction names should have a 'v' prefix.
5256
174ec470
RB		 52572013-09-12 Richard Biener <rguenther@suse.de>
5258
5259 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5260 (dot_rdg): Use popen instead of system in optional code.
5261 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5262 (already_processed_vertex_p): Adjust.
5263 (has_anti_or_output_dependence, predecessor_has_mem_write,
5264 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5265 rdg_flag_uses): Remove.
5266 (rdg_flag_vertex): Simplify.
5267 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5268 remove recursion.
5269 (build_rdg_partition_for_component): Process the first vertex
5270 of a component only.
5271 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5272
38c89763
AM		 52732013-09-12 Alan Modra <amodra@gmail.com>
5274
5275 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5276
3e3d9d17
DD		 52772013-09-11 DJ Delorie <dj@redhat.com>
5278 Nick Clifton <nickc@redhat.com>
5279
5280 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5281 (rl78_stack_based_mem): New.
5282 * config/rl78/constraints.md (Iv08): New.
5283 (Iv16): New.
5284 (Iv24): New.
5285 (Is09): New.
5286 (Is17): New.
5287 (Is25): New.
5288 (ISsi): New.
5289 (IShi): New.
5290 (ISqi): New.
5291 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5292 (movhi): Likewise.
5293 (movsi): Change from expand to insn-and-split.
5294 (ashrsi3): Clobber AX.
5295 (lshrsi3): New.
5296 (ashlsi3): New.
5297 (cbranchsi4): New.
5298 * config/rl78/rl78.md (CC_REG): Fix.
5299 (addsi3): Allow memory and immediate operands.
5300 (addsi3_internal): Split into...
5301 (addsi3_internal_virt): ...new, and ...
5302 (addsi3_internal_real): ...new.
5303 (subsi): New.
5304 (subsi3_internal_virt): New.
5305 (subsi3_internal_real): New.
5306 (mulsi3): Add memory operand.
5307 (mulsi3_rl78): Likewise.
5308 (mulsi3_g13): Likewise.
5309 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5310 (cbranchqi4_real): Add more constraint options.
5311 (cbranchhi4_real): Expand pattern.
5312 (cbranchhi4_real_signed): New.
5313 (cbranchhi4_real_inverted): New.
5314 (cbranchsi4_real_lt): New.
5315 (cbranchsi4_real_ge): New.
5316 (cbranchsi4_real_signed): New.
5317 (cbranchsi4_real): New.
5318 (peephole2): New.
893e85fa
CM		 5319 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5320 constant shifts.
3e3d9d17
DD		 5321 (lshrsi3_virt): Likewise.
5322 (ashlsi3_virt): Likewise.
5323 (cbranchqi4_virt_signed): New.
5324 (cbranchhi4_virt_signed): New.
5325 (cbranchsi4_virt): New.
5326 * config/rl78/rl78.c: Whitespace fixes throughout.
5327 (move_elim_pass): New.
5328 (pass_data_rl78_move_elim): New.
5329 (pass_rl78_move_elim): New.
5330 (make_pass_rl78_move_elim): New.
5331 (rl78_devirt_info): Run devirt earlier.
5332 (rl78_move_elim_info): New.
5333 (rl78_asm_file_start): Register it.
5334 (rl78_split_movsi): New.
5335 (rl78_as_legitimate_address): Allow virtual base registers when
5336 appropriate.
5337 (rl78_addr_space_convert): Remove spurious debug stuff.
5338 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5339 (rl78_print_operand): More cases for not printing '#'.
5340 (rl78_expand_compare): Remove most of the logic.
5341 (content_memory): New.
5342 (clear_content_memory): New.
5343 (get_content_index): New.
5344 (get_content_name): New.
5345 (display_content_memory): New.
5346 (update_content): New.
5347 (record_content): New.
5348 (already_contains): New.
5349 (insn_ok_now): Re-recog insns with virtual registers.
5350 (add_postponed_content_update): New.
5351 (process_postponed_content_update): New.
5352 (gen_and_emit_move): New.
893e85fa
CM		 5353 (transcode_memory_rtx): Record new location content.
5354 Use gen_and_emit_move.
3e3d9d17
DD		 5355 (force_into_acc): New.
5356 (move_to_acc): Use gen_and_emit_move.
5357 (move_from_acc): Likewise.
5358 (move_acc_to_reg): Likewise.
5359 (move_to_x): Likewise.
5360 (move_to_hl): Likewise.
5361 (move_to_de): Likewise.
5362 (rl78_alloc_physical_registers_op1): Record location content.
5363 (has_constraint): New.
5364 (rl78_alloc_physical_registers_op2): Record location content.
5365 Optimize use of HL.
5366 (rl78_alloc_physical_registers_ro1): Likewise.
5367 (rl78_alloc_physical_registers_cmp): Likewise.
5368 (rl78_alloc_physical_registers_umul): Likewise.
5369 (rl78_alloc_address_registers_macax): New.
5370 (rl78_alloc_physical_registers): Initialize and set location
5371 content memory as needed.
5372 (rl78_reorg): Make sure split2 is called.
5373 (rl78_rtx_costs): New.
5374
088845a5
RS		 53752013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5376
5377 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5378 for (not (neg ...)) and (neg (not ...)) cases.
5379
5c2961cf
RB		 53802013-09-11 Richard Biener <rguenther@suse.de>
5381
5382 PR middle-end/58377
5383 * passes.def: Split critical edges before late uninit warning passes.
5384 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5385
427e6a14
JJ		 53862013-09-11 Jakub Jelinek <jakub@redhat.com>
5387
5388 PR tree-optimization/58385
5389 * fold-const.c (build_range_check): If both low and high are NULL,
5390 use omit_one_operand_loc to preserve exp side-effects.
5391
df45c0cc
KT		 53922013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5393
5394 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5395
80ab0b19
RB		 53962013-09-11 Richard Biener <rguenther@suse.de>
5397
5398 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5399 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5400 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5401 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5402 stmts_from_loop, known_dependences_p, build_empty_rdg,
5403 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5404 * tree-loop-distribution.c: ... here.
5405 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5406 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5407 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5408 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5409 * tree-loop-distribution.c: ... here.
5410 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5411 (struct partition_s): Add loops member.
5412 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5413 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5414 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5415
85a77221
AI		 54162013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5417 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5418 Sergey Lega <sergey.s.lega@intel.com>
5419 Anna Tikhonova <anna.tikhonova@intel.com>
5420 Ilya Tocar <ilya.tocar@intel.com>
5421 Andrey Turetskiy <andrey.turetskiy@intel.com>
5422 Ilya Verbin <ilya.verbin@intel.com>
5423 Kirill Yukhin <kirill.yukhin@intel.com>
5424 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5425
5426 * config/i386/constraints.md (k): New.
5427 (Yk): Ditto.
5428 * config/i386/i386.c (const regclass_map): Add new mask registers.
5429 (dbx_register_map): Ditto.
5430 (dbx64_register_map): Ditto.
5431 (svr4_dbx_register_map): Ditto.
5432 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5433 disabled.
5434 (ix86_preferred_reload_class): Disable constants for mask registers.
5435 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5436 (ix86_hard_regno_mode_ok): Support new mask registers.
5437 (x86_order_regs_for_local_alloc): Ditto.
5438 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5439 (FIXED_REGISTERS): Add new mask registers.
5440 (CALL_USED_REGISTERS): Ditto.
5441 (REG_ALLOC_ORDER): Ditto.
5442 (VALID_MASK_REG_MODE): New.
5443 (FIRST_MASK_REG): Ditto.
5444 (LAST_MASK_REG): Ditto.
5445 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5446 (MAYBE_MASK_CLASS_P): New.
5447 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5448 (REG_CLASS_CONTENTS): Ditto.
5449 (MASK_REGNO_P): New.
5450 (ANY_MASK_REG_P): Ditto.
5451 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM		 5452 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5453 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5454 mask registers.
85a77221
AI		 5455 (attribute "type"): Add mskmov, msklog.
5456 (attribute "length_immediate"): Support them.
5457 (attribute "memory"): Ditto.
5458 (attribute "prefix_0f"): Ditto.
5459 (*movhi_internal): Support new mask registers.
5460 (*movqi_internal): Ditto.
5461 (define_split): Split out clobber pattern is a logic
5462 insn on mask registers.
5463 (*k<logic><mode>): New.
5464 (*andhi_1): Extend to support mask regs.
5465 (*andqi_1): Extend to support mask regs.
5466 (kandn<mode>): New.
5467 (define_split): Split and-not to and and not if operands
5468 are not mask regs.
5469 (*<code><mode>_1): Separate HI mode to new pattern...
5470 (*<code>hi_1): This.
5471 (*<code>qi_1): Extend to support mask regs.
5472 (kxnor<mode>): New.
5473 (kortestzhi): Ditto.
5474 (kortestchi): Ditto.
5475 (kunpckhi): Ditto.
5476 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5477 (*one_cmplhi2_1): ...Here, now with mask registers support.
5478 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM		 5479 (HI/QImode arithmetics splitter): Don't split if mask registers
5480 are used.
85a77221
AI		 5481 (HI/QImode not splitter): Ditto.
5482 * config/i386/predicated.md (mask_reg_operand): New.
5483 (general_reg_operand): Ditto.
5484
7e570821
AI		 54852013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5486
5487 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5488 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5489
9e1376e9
JL		 54902013-09-10 Jeff Law <law@redhat.com>
5491
5492 PR tree-optimization/58380
5493 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5494 of threading through a buried loop header.
5495
5496 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5497 return value for single successor case.
5498
5bbcb888
JH		 54992013-09-10 Jan Hubicka <jh@suse.cz>
5500
5501 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5502
d742ff4b
RE		 55032013-09-10 Richard Earnshaw <rearnsha@arm.com>
5504
5505 PR target/58361
5506 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5507 support conditional execution.
5508 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5509
b0681c9e
VM		 55102013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5511
5512 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5513 constraint pass.
f77e2d2d
JBG		 5514 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5515 Check destination too to check move insn.
b0681c9e 5516 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG		 5517 changed its allocation and the optional reload was inherited on last
5518 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM		 5519 (lra_undo_inheritance): Add check that inherited pseudo still in
5520 memory.
5521
c9b1c957
JG		 55222013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5523
5524 * config/aarch64/aarch64.md (generic_sched): New.
5525 * config/aarch64/aarch64-generic.md (load): Make conditional
5526 on generic_sched attribute.
5527 (nonload): Likewise.
5528
82d618d3
JH		 55292013-09-10 Jan Hubicka <jh@suse.cz>
5530
5531 * lto-cgraph.c: Include ipa-utils.h.
5532 (compute_ltrans_boundary): Also add possible targets into the boundary.
5533
668b6894
JH		 55342013-09-10 Jan Hubicka <jh@suse.cz>
5535
5536 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5537 VAR_DECL of vtable rather than full expression.
5538
4f90d3e0
JH		 55392013-09-10 Jan Hubicka <jh@suse.cz>
5540 Paolo Carlini <paolo.carlini@oracle.com>
5541
5542 * cgraphunit.c (analyze_functions): Save input_location, set it
5543 to UNKNOWN_LOCATION and restore it at the end.
5544
39e87baf
MJ		 55452013-09-10 Martin Jambor <mjambor@suse.cz>
5546
5547 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5548 represented by a thunk.
5549
581aedec
JL		 55502013-09-10 Jeff Law <law@redhat.com>
5551
5552 PR tree-optimization/58343
5553 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5554 jump threading requests through loop headers buried in the middle
5555 of a jump threading path.
5556
5557 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5558 in return value/type.
5559
84cf4ab6
JJ		 55602013-09-10 Jakub Jelinek <jakub@redhat.com>
5561
5562 PR rtl-optimization/58365
5563 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5564 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5565 it differs.
5566
447dd906
RB		 55672013-09-10 Richard Biener <rguenther@suse.de>
5568
5569 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5570 * tree-data-ref.c (create_rdg_vertices): Collect all data
5571 references, signal failure to the caller, use data-ref API.
893e85fa
CM		 5572 (build_rdg): Compute data references only once. Maintain lifetime
5573 of data references and data dependences from within RDG.
447dd906
RB		 5574 (free_rdg): Free dependence relations.
5575 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5576 inventing extra dependences.
5577 (distribute_loop): Update for RDG API changes.
5578
f14a7324
KT		 55792013-09-10 Kai Tietz <ktietz@redhat.com>
5580
5581 * doc/invoke.texi (fms-extensions): Document changed
5582 behavior for ms-abi targets.
5583 * config/i386/i386.c (ix86_option_override_internal):
5584 Set default value of option -fms-extension for ms-abi targets.
5585
ff131f26
MZ		 55862013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5587
5588 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5589
33115a2c
AM		 55902013-09-10 Alan Modra <amodra@gmail.com>
5591
5592 PR target/58330
5593 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5594
b6872a13
AM		 55952013-09-10 Alan Modra <amodra@gmail.com>
5596
5597 * config/rs6000/predicates.md (add_cint_operand): New.
5598 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5599 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5600 using add_cint_operand.
5601 (largetoc_high_plus_aix): Likewise.
5602
28fd0ba2
JJ		 56032013-09-09 Jakub Jelinek <jakub@redhat.com>
5604
5605 PR tree-optimization/58364
5606 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5607 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5608 the current range can't be an unconditional true or false.
5609
552d99fe
JG		 56102013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5611
5612 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5613
f43245d1
UB		 56142013-09-09 Uros Bizjak <ubizjak@gmail.com>
5615
5616 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5617
c9b0866a
PC		 56182013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5619
5620 PR c++/43452
5621 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5622
78d8b9f0
IB		 56232013-09-09 Ian Bolton <ian.bolton@arm.com>
5624
5625 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5626 NO_REGS for immediate that can't be moved directly into FP_REGS.
5627
274b2532
KT		 56282013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5629
5630 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5631 comparison with negated operand.
c9b0866a
PC		 5632 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5633 RTL form.
274b2532 5634
467a3558
RB		 56352013-09-09 Richard Biener <rguenther@suse.de>
5636
5637 PR middle-end/58326
5638 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5639 of a subloop record all its block as affecting loop-closed SSA form.
5640
088c5368
RS		 56412013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5642
5643 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5644 of an rtx/bitpos pair.
5645 (store_fixed_bit_field): Update accordingly.
5646
69db2d57
RS		 56472013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5648
5649 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5650 GEN_INT.
5651 * builtins.c (expand_errno_check): Likewise.
5652 * dwarf2cfi.c (init_return_column_size): Likewise.
5653 * except.c (sjlj_mark_call_sites): Likewise.
5654 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5655 * lra-constraints.c (emit_inc): Likewise.
5656 * ree.c (combine_set_extension): Likewise.
5657 * regmove.c (fixup_match_2): Likewise.
5658 * reload1.c (inc_for_reload): Likewise.
5659
69a59f0f
RS		 56602013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5661
5662 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5663 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5664 Use gen_int_mode with the mode of the associated simplify_* call.
5665 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5666 * expmed.c (expand_shift_1): Likewise.
5667 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5668 * loop-iv.c (iv_number_of_iterations): Likewise.
5669 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5670 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5671
2f1cd2eb
RS		 56722013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5673
5674 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5675 of the associated expand_* call.
5676 (asan_emit_stack_protection): Likewise.
5677 * builtins.c (round_trampoline_addr): Likewise.
5678 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5679 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5680 (emit_store_flag): Likewise.
5681 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5682 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5683 Likewise.
5684 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5685 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5686 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5687 * modulo-sched.c (generate_prolog_epilog): Likewise.
5688 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5689 (expand_ctz, expand_ffs, expand_unop): Likewise.
5690
4789c0ce
RS		 56912013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5692
5693 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5694 of the associated gen_rtx_* call.
5695 * caller-save.c (init_caller_save): Likewise.
5696 * combine.c (find_split_point, make_extraction): Likewise.
5697 (make_compound_operation): Likewise.
5698 * dwarf2out.c (mem_loc_descriptor): Likewise.
5699 * explow.c (plus_constant, probe_stack_range): Likewise.
5700 * expmed.c (expand_mult_const): Likewise.
5701 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5702 * reload1.c (init_reload): Likewise.
5703 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5704 * var-tracking.c (adjust_mems): Likewise.
5705 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5706 rather than gen_rtx_fmt_ee.
5707
a811f7cb
JH		 57082013-09-09 Jan Hubicka <jh@suse.cz>
5709
5710 PR middle-end/58294
5711 * value-prof.c (gimple_ic): Copy also abnormal edges.
5712
dcad1dd3
RS		 57132013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5714
5715 * asan.c (asan_shadow_cst): Use gen_int_mode.
5716
daf5c770
JH		 57172013-09-08 Jan Hubicka <jh@suse.cz>
5718
5719 * ipa-profile.c: Add toplevel comment.
5720 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5721 (contains_hot_call_p): New function.
5722 (ipa_propagate_frequency): Set frequencies based on counts when
5723 profile is read.
5724 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5725 profile; do not tamper with profile after inlining if it is read.
5726
a12cd2db
JH		 57272013-09-08 Jan Hubicka <jh@suse.cz>
5728
5729 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5730 speculative edges.
5731
856a1588
JH		 57322013-09-08 Jan Hubicka <jh@suse.cz>
5733
5734 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5735 summary generation.
5736
4f4b0b73
JL		 57372013-09-08 Jeff Law <law@redhat.com>
5738
5739 PR bootstrap/58340
5740 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5741 of 'found'.
5742
cc4029ee
AK		 57432013-09-08 Andi Kleen <ak@linux.intel.com>
5744
5745 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5746
7b395ddd
JH		 57472013-09-08 Jan Hubicka <jh@suse.cz>
5748
5749 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5750
c34c46dd
RS		 57512013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5752
5753 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5754 for non-debug insns.
5755 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5756 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5757 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5758 accordingly.
5759
3462aa02
JH		 57602013-09-08 Jan Hubicka <jh@suse.cz>
5761
5762 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5763 targets and devirtualize to BUILT_IN_UNREACHABLE.
5764 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5765 * ipa.c (walk_polymorphic_call_targets): New function.
5766 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5767 functions; use the new timevar.
5768 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5769 was removed from the program.
5770 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5771 not consider it in the walk when its vtable is dead.
5772 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5773 record_binfo.
5774 (devirt_variable_node_removal_hook): New function.
5775 (possible_polymorphic_call_targets): Also register
5776 devirt_variable_node_removal_hook.
5777 (ipa_devirt): Do not do non-speculative devirtualization.
5778 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5779
26e5b0fd
JH		 57802013-09-08 Jan Hubicka <jh@suse.cz>
5781
5782 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5783 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5784 varpool_remove_variable_insertion_hook): Declare.
5785 * varpool.c (varpool_node_hook_list): New structure.
5786 (first_varpool_node_removal_hook,
5787 first_varpool_variable_insertion_hook): New variables.
5788 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5789 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5790 varpool_remove_variable_insertion_hook,
5791 varpool_call_variable_insertion_hooks): New functions.
5792 (varpool_remove_node): Use it.
5793
97f1f314 57942013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 5795
97f1f314
PC		 5796 PR c++/54941
5797 * diagnostic.c (diagnostic_build_prefix): When s.file is
5798 "<built-in>" don't output line and column numbers.
6df2ca09 5799
bcb650cb
JH		 58002013-09-06 Jan Hubicka <jh@suse.cz>
5801
5802 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5803 * lto-streamer-out.c: Stream thunks, too.
5804 * lto-streamer-in.c (input_function): Pop cfun here
5805 (lto_read_body): Instead of here.
5806
87f3fea8
CT		 58072013-09-06 Caroline Tice <cmtice@google.com>
5808
5809 * doc/install.texi: Add documentation for the --enable-vtable-verify
5810 and the --disable-libvtv configure options.
5811
5e8d9b45
JL		 58122013-09-06 Jeff Law <law@redhat.com>
5813
5814 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5815 edge implied equivalences into successor phis.
5816
3a001aff
JR		 58172013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5818
5819 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5820
58212013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 5822
5823 * resource.c (mark_target_live_regs): Compute resources taking
5824 into account if a call is predicated or not.
5825
2c641f8b
EB		 58262013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5827
5828 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5829 the compiler in the function names.
5830
a81b0a3d
JH		 58312013-09-06 Jan Hubicka <jh@suse.cz>
5832
5833 PR middle-end/58094
5834 * ipa-inline.c (has_caller_p): New function.
5835 (want_inline_function_to_all_callers_p): Use it.
5836 (sum_callers, inline_to_all_callers): Break out from ...
5837 (ipa_inline): ... here.
5838
68e4ca7e
JH		 58392013-09-06 Jan Hubicka <jh@suse.cz>
5840
2c641f8b
EB		 5841 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5842 only when AVX is enabled.
68e4ca7e 5843
3c73e2fe
JG		 58442013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5845
5846 * config/aarch64/aarch64.md
5847 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5848 is fpsimd_<load/store>2.
5849 (load_pair<mode>): Likewise.
5850 (store_pair<mode>): Likewise.
5851
8c48eecd
JG		 58522013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5853
5854 * config/arm/types.md (type): Add "mrs" type.
5855 * config/aarch64/aarch64.md
5856 (aarch64_load_tp_hard): Make type "mrs".
5857 * config/arm/arm.md
5858 (load_tp_hard): Make type "mrs".
5859 * config/arm/cortex-a15.md: Update with new attributes.
5860 * config/arm/cortex-a5.md: Update with new attributes.
5861 * config/arm/cortex-a53.md: Update with new attributes.
5862 * config/arm/cortex-a7.md: Update with new attributes.
5863 * config/arm/cortex-a8.md: Update with new attributes.
5864 * config/arm/cortex-a9.md: Update with new attributes.
5865 * config/arm/cortex-m4.md: Update with new attributes.
5866 * config/arm/cortex-r4.md: Update with new attributes.
5867 * config/arm/fa526.md: Update with new attributes.
5868 * config/arm/fa606te.md: Update with new attributes.
5869 * config/arm/fa626te.md: Update with new attributes.
5870 * config/arm/fa726te.md: Update with new attributes.
5871
9ea187a5
JG		 58722013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5873
5874 * config/aarch64/aarch64.md
5875 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5876 (*movtf_aarch64): Likewise.
5877 * config/arm/arm.md
5878 (thumb1_movdi_insn): Use "multiple" for type where more than one
5879 instruction is used for a move.
5880 (*arm32_movhf): Likewise.
5881 (*thumb_movdf_insn): Likewise.
5882
292b89b3
JG		 58832013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5884
5885 * config/arm/types.md (type): Rename fcpys to fmov.
5886 * config/arm/vfp.md
5887 (*arm_movsi_vfp): Rename type fcpys as fmov.
5888 (*thumb2_movsi_vfp): Likewise
5889 (*movhf_vfp_neon): Likewise
5890 (*movhf_vfp): Likewise
5891 (*movsf_vfp): Likewise
5892 (*thumb2_movsf_vfp): Likewise
5893 (*movsfcc_vfp): Likewise
5894 (*thumb2_movsfcc_vfp): Likewise
5895 * config/aarch64/aarch64-simd.md
5896 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5897 * config/aarch64/aarch64.md
5898 (*movsi_aarch64): Replace type mov_reg with fmovs.
5899 (*movdi_aarch64): Likewise
5900 (*movsf_aarch64): Likewise
5901 (*movdf_aarch64): Likewise
5902 * config/arm/arm.c
5903 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5904 * config/arm/iwmmxt.md
5905 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5906 * config/arm/arm1020e.md: Update with new attributes.
5907 * config/arm/cortex-a15-neon.md: Update with new attributes.
5908 * config/arm/cortex-a5.md: Update with new attributes.
5909 * config/arm/cortex-a53.md: Update with new attributes.
5910 * config/arm/cortex-a7.md: Update with new attributes.
5911 * config/arm/cortex-a8-neon.md: Update with new attributes.
5912 * config/arm/cortex-a9.md: Update with new attributes.
5913 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5914 * config/arm/cortex-r4f.md: Update with new attributes.
5915 * config/arm/marvell-pj4.md: Update with new attributes.
5916 * config/arm/vfp11.md: Update with new attributes.
5917
6a24a83e
JG		 59182013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5919
5920 * config/aarch64/aarch64.md
5921 (*madd<mode>): Fix type attribute.
5922 (*maddsi_uxtw): Likewise.
5923 (*msub<mode>): Likewise.
5924 (*msubsi_uxtw): Likewise.
5925 (<su_optab>maddsidi4): Likewise.
5926 (<su_optab>msubsidi4): Likewise.
5927
b86923f0
JG		 59282013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5929
5930 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5931 * config/arm/arm.md (core_cycles): Remove fdiv.
5932 * config/arm/vfp.md:
5933 (*sqrtsf2_vfp): Update for attribute changes.
5934 (*sqrtdf2_vfp): Likewise.
5935 * config/aarch64/aarch64.md:
5936 (sqrt<mode>2): Update for attribute changes.
5937 * config/arm/arm1020e.md: Update with new attributes.
5938 * config/arm/cortex-a15-neon.md: Update with new attributes.
5939 * config/arm/cortex-a5.md: Update with new attributes.
5940 * config/arm/cortex-a53.md: Update with new attributes.
5941 * config/arm/cortex-a7.md: Update with new attributes.
5942 * config/arm/cortex-a8-neon.md: Update with new attributes.
5943 * config/arm/cortex-a9.md: Update with new attributes.
5944 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5945 * config/arm/cortex-r4f.md: Update with new attributes.
5946 * config/arm/marvell-pj4.md: Update with new attributes.
5947 * config/arm/vfp11.md: Update with new attributes.
5948
7b49c9e1
JG		 59492013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5950
5951 * config/arm/types.md
5952 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5953 * config/aarch64/aarch64.md
5954 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5955 new attributes.
5956 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5957 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5958 (float<GPI:mode><GPF:mode>2): Likewise.
5959 * config/arm/vfp.md
5960 (*truncsisf2_vfp): Update with new attributes.
5961 (*truncsidf2_vfp): Likewise.
5962 (fixuns_truncsfsi2): Likewise.
5963 (fixuns_truncdfsi2): Likewise.
5964 (*floatsisf2_vfp): Likewise.
5965 (*floatsidf2_vfp): Likewise.
5966 (floatunssisf2): Likewise.
5967 (floatunssidf2): Likewise.
5968 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5969 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5970 * config/arm/arm1020e.md: Update with new attributes.
5971 * config/arm/cortex-a15-neon.md: Update with new attributes.
5972 * config/arm/cortex-a5.md: Update with new attributes.
5973 * config/arm/cortex-a53.md: Update with new attributes.
5974 * config/arm/cortex-a7.md: Update with new attributes.
5975 * config/arm/cortex-a8-neon.md: Update with new attributes.
5976 * config/arm/cortex-a9.md: Update with new attributes.
5977 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5978 * config/arm/cortex-r4f.md: Update with new attributes.
5979 * config/arm/marvell-pj4.md: Update with new attributes.
5980 * config/arm/vfp11.md: Update with new attributes.
5981
e7df8af8
JG		 59822013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5983
5984 * config/aarch64/arm_neon.h
5985 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5986 (vqtbx<1,2,3,4><q>_s8): Likewise.
5987
594726e4
JG		 59882013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5989
f43245d1
UB		 5990 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5991 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG		 5992 (add<mode>3): Add type attribute.
5993 (add<mode>3): Likewise.
5994 (usadd<mode>3): Likewise.
5995 (ssadd<mode>3): Likewise.
5996 (sub<mode>3): Likewise.
5997 (sub<mode>3): Likewise.
5998 (ussub<mode>3): Likewise.
5999 (sssub<mode>3): Likewise.
6000 (ssmulsa3): Likewise.
6001 (usmulusa3): Likewise.
6002 (arm_usatsihi): Likewise.
6003 * config/arm/vfp.md
6004 (*movdi_vfp): Add types for all instructions.
6005 (*movdi_vfp_cortexa8): Likewise.
6006 (*movhf_vfp_neon): Likewise.
6007 (*movhf_vfp): Likewise.
6008 (*movdf_vfp): Likewise.
6009 (*thumb2_movdf_vfp): Likewise.
6010 (*thumb2_movdfcc_vfp): Likewise.
6011 * config/arm/arm.md: Add type attribute to all insn patterns.
6012 (*thumb1_adddi3): Add type attribute.
6013 (*arm_adddi3): Likewise.
6014 (*adddi_sesidi_di): Likewise.
6015 (*adddi_zesidi_di): Likewise.
6016 (*thumb1_addsi3): Likewise.
6017 (addsi3_compare0): Likewise.
6018 (*addsi3_compare0_scratch): Likewise.
6019 (*compare_negsi_si): Likewise.
6020 (cmpsi2_addneg): Likewise.
6021 (*addsi3_carryin_<optab>): Likewise.
6022 (*addsi3_carryin_alt2_<optab>): Likewise.
6023 (*addsi3_carryin_clobercc_<optab>): Likewise.
6024 (*subsi3_carryin): Likewise.
6025 (*subsi3_carryin_const): Likewise.
6026 (*subsi3_carryin_compare): Likewise.
6027 (*subsi3_carryin_compare_const): Likewise.
6028 (*arm_subdi3): Likewise.
6029 (*thumb_subdi3): Likewise.
6030 (*subdi_di_zesidi): Likewise.
6031 (*subdi_di_sesidi): Likewise.
6032 (*subdi_zesidi_di): Likewise.
6033 (*subdi_sesidi_di): Likewise.
6034 (*subdi_zesidi_ze): Likewise.
6035 (thumb1_subsi3_insn): Likewise.
6036 (*arm_subsi3_insn): Likewise.
6037 (*anddi3_insn): Likewise.
6038 (*anddi_zesidi_di): Likewise.
6039 (*anddi_sesdi_di): Likewise.
6040 (*ne_zeroextracts): Likewise.
6041 (*ne_zeroextracts): Likewise.
6042 (*ite_ne_zeroextr): Likewise.
6043 (*ite_ne_zeroextr): Likewise.
6044 (*anddi_notdi_di): Likewise.
6045 (*anddi_notzesidi): Likewise.
6046 (*anddi_notsesidi): Likewise.
6047 (andsi_notsi_si): Likewise.
6048 (thumb1_bicsi3): Likewise.
6049 (*iordi3_insn): Likewise.
6050 (*iordi_zesidi_di): Likewise.
6051 (*iordi_sesidi_di): Likewise.
6052 (*thumb1_iorsi3_insn): Likewise.
6053 (*xordi3_insn): Likewise.
6054 (*xordi_zesidi_di): Likewise.
6055 (*xordi_sesidi_di): Likewise.
6056 (*arm_xorsi3): Likewise.
6057 (*andsi_iorsi3_no): Likewise.
6058 (*smax_0): Likewise.
6059 (*smax_m1): Likewise.
6060 (*arm_smax_insn): Likewise.
6061 (*smin_0): Likewise.
6062 (*arm_smin_insn): Likewise.
6063 (*arm_umaxsi3): Likewise.
6064 (*arm_uminsi3): Likewise.
6065 (*minmax_arithsi): Likewise.
6066 (*minmax_arithsi_): Likewise.
6067 (*satsi_<SAT:code>): Likewise.
6068 (arm_ashldi3_1bit): Likewise.
6069 (arm_ashrdi3_1bit): Likewise.
6070 (arm_lshrdi3_1bit): Likewise.
6071 (*arm_negdi2): Likewise.
6072 (*thumb1_negdi2): Likewise.
6073 (*arm_negsi2): Likewise.
6074 (*thumb1_negsi2): Likewise.
6075 (*negdi_extendsid): Likewise.
6076 (*negdi_zero_extend): Likewise.
6077 (*arm_abssi2): Likewise.
6078 (*thumb1_abssi2): Likewise.
6079 (*arm_neg_abssi2): Likewise.
6080 (*thumb1_neg_abss): Likewise.
6081 (one_cmpldi2): Likewise.
6082 (extend<mode>di2): Likewise.
6083 (*compareqi_eq0): Likewise.
6084 (*arm_extendhisi2addsi): Likewise.
6085 (*arm_movdi): Likewise.
6086 (*thumb1_movdi_insn): Likewise.
6087 (*arm_movt): Likewise.
6088 (*thumb1_movsi_insn): Likewise.
6089 (pic_add_dot_plus_four): Likewise.
6090 (pic_add_dot_plus_eight): Likewise.
6091 (tls_load_dot_plus_eight): Likewise.
6092 (*thumb1_movhi_insn): Likewise.
6093 (*thumb1_movsf_insn): Likewise.
6094 (*movdf_soft_insn): Likewise.
6095 (*thumb_movdf_insn): Likewise.
6096 (cbranchsi4_insn): Likewise.
6097 (cbranchsi4_scratch): Likewise.
6098 (*negated_cbranchsi4): Likewise.
6099 (*tbit_cbranch): Likewise.
6100 (*tlobits_cbranch): Likewise.
6101 (*tstsi3_cbranch): Likewise.
6102 (*cbranchne_decr1): Likewise.
6103 (*addsi3_cbranch): Likewise.
6104 (*addsi3_cbranch_scratch): Likewise.
6105 (*arm_cmpdi_insn): Likewise.
6106 (*arm_cmpdi_unsig): Likewise.
6107 (*arm_cmpdi_zero): Likewise.
6108 (*thumb_cmpdi_zero): Likewise.
6109 (*deleted_compare): Likewise.
6110 (*mov_scc): Likewise.
6111 (*mov_negscc): Likewise.
6112 (*mov_notscc): Likewise.
6113 (*cstoresi_eq0_thumb1_insn): Likewise.
6114 (cstoresi_nltu_thumb1): Likewise.
6115 (cstoresi_ltu_thu): Likewise.
6116 (thumb1_addsi3_addgeu): Likewise.
6117 (*arm_jump): Likewise.
6118 (*thumb_jump): Likewise.
6119 (*check_arch2): Likewise.
6120 (arm_casesi_internal): Likewise.
6121 (thumb1_casesi_dispatch): Likewise.
6122 (*arm_indirect_jump): Likewise.
6123 (*thumb1_indirect_jump): Likewise.
6124 (nop): Likewise.
6125 (*and_scc): Likewise.
6126 (*ior_scc): Likewise.
6127 (*compare_scc): Likewise.
6128 (*cond_move): Likewise.
6129 (*cond_arith): Likewise.
6130 (*cond_sub): Likewise.
6131 (*cmp_ite0): Likewise.
6132 (*cmp_ite1): Likewise.
6133 (*cmp_and): Likewise.
6134 (*cmp_ior): Likewise.
6135 (*ior_scc_scc): Likewise.
6136 (*ior_scc_scc_cmp): Likewise.
6137 (*and_scc_scc): Likewise.
6138 (*and_scc_scc_cmp): Likewise.
6139 (*and_scc_scc_nod): Likewise.
6140 (*negscc): Likewise.
6141 (movcond_addsi): Likewise.
6142 (movcond): Likewise.
6143 (*ifcompare_plus_move): Likewise.
6144 (*if_plus_move): Likewise.
6145 (*ifcompare_move_plus): Likewise.
6146 (*if_move_plus): Likewise.
6147 (*ifcompare_arith_arith): Likewise.
6148 (*if_arith_arith): Likewise.
6149 (*ifcompare_arith_move): Likewise.
6150 (*if_arith_move): Likewise.
6151 (*ifcompare_move_arith): Likewise.
6152 (*if_move_arith): Likewise.
6153 (*ifcompare_move_not): Likewise.
6154 (*if_move_not): Likewise.
6155 (*ifcompare_not_move): Likewise.
6156 (*if_not_move): Likewise.
6157 (*ifcompare_shift_move): Likewise.
6158 (*if_shift_move): Likewise.
6159 (*ifcompare_move_shift): Likewise.
6160 (*if_move_shift): Likewise.
6161 (*ifcompare_shift_shift): Likewise.
6162 (*ifcompare_not_arith): Likewise.
6163 (*ifcompare_arith_not): Likewise.
6164 (*if_arith_not): Likewise.
6165 (*ifcompare_neg_move): Likewise.
6166 (*if_neg_move): Likewise.
6167 (*ifcompare_move_neg): Likewise.
6168 (*if_move_neg): Likewise.
6169 (prologue_thumb1_interwork): Likewise.
6170 (*cond_move_not): Likewise.
6171 (*sign_extract_onebit): Likewise.
6172 (*not_signextract_onebit): Likewise.
6173 (stack_tie): Likewise.
6174 (align_4): Likewise.
6175 (align_8): Likewise.
6176 (consttable_end): Likewise.
6177 (consttable_1): Likewise.
6178 (consttable_2): Likewise.
6179 (consttable_4): Likewise.
6180 (consttable_8): Likewise.
6181 (consttable_16): Likewise.
6182 (*thumb1_tablejump): Likewise.
6183 (prefetch): Likewise.
6184 (force_register_use): Likewise.
6185 (thumb_eh_return): Likewise.
6186 (load_tp_hard): Likewise.
6187 (load_tp_soft): Likewise.
6188 (tlscall): Likewise.
6189 (*arm_movtas_ze): Likewise.
6190 (*arm_rev): Likewise.
6191 (*arm_revsh): Likewise.
6192 (*arm_rev16): Likewise.
6193 * config/arm/thumb2.md
6194 (*thumb2_smaxsi3): Likewise.
6195 (*thumb2_sminsi3): Likewise.
6196 (*thumb32_umaxsi3): Likewise.
6197 (*thumb2_uminsi3): Likewise.
6198 (*thumb2_negdi2): Likewise.
6199 (*thumb2_abssi2): Likewise.
6200 (*thumb2_neg_abss): Likewise.
6201 (*thumb2_movsi_insn): Likewise.
6202 (tls_load_dot_plus_four): Likewise.
6203 (*thumb2_movhi_insn): Likewise.
6204 (*thumb2_mov_scc): Likewise.
6205 (*thumb2_mov_negs): Likewise.
6206 (*thumb2_mov_negs): Likewise.
6207 (*thumb2_mov_nots): Likewise.
6208 (*thumb2_mov_nots): Likewise.
6209 (*thumb2_movsicc_): Likewise.
6210 (*thumb2_movsfcc_soft_insn): Likewise.
6211 (*thumb2_indirect_jump): Likewise.
6212 (*thumb2_and_scc): Likewise.
6213 (*thumb2_ior_scc): Likewise.
6214 (*thumb2_ior_scc_strict_it): Likewise.
6215 (*thumb2_cond_move): Likewise.
6216 (*thumb2_cond_arith): Likewise.
6217 (*thumb2_cond_ari): Likewise.
6218 (*thumb2_cond_sub): Likewise.
6219 (*thumb2_negscc): Likewise.
6220 (*thumb2_movcond): Likewise.
6221 (thumb2_casesi_internal): Likewise.
6222 (thumb2_casesi_internal_pic): Likewise.
6223 (*thumb2_alusi3_short): Likewise.
6224 (*thumb2_mov<mode>_shortim): Likewise.
6225 (*thumb2_addsi_short): Likewise.
6226 (*thumb2_subsi_short): Likewise.
6227 (thumb2_addsi3_compare0): Likewise.
6228 (*thumb2_cbz): Likewise.
6229 (*thumb2_cbnz): Likewise.
6230 (*thumb2_one_cmplsi2_short): Likewise.
6231 (*thumb2_negsi2_short): Likewise.
6232 (*orsi_notsi_si): Likewise.
6233 * config/arm/arm1020e.md: Update with new attributes.
6234 * config/arm/arm1026ejs.md: Update with new attributes.
6235 * config/arm/arm1136jfs.md: Update with new attributes.
6236 * config/arm/arm926ejs.md: Update with new attributes.
6237 * config/arm/cortex-a15.md: Update with new attributes.
6238 * config/arm/cortex-a5.md: Update with new attributes.
6239 * config/arm/cortex-a53.md: Update with new attributes.
6240 * config/arm/cortex-a7.md: Update with new attributes.
6241 * config/arm/cortex-a8.md: Update with new attributes.
6242 * config/arm/cortex-a9.md: Update with new attributes.
6243 * config/arm/cortex-m4.md: Update with new attributes.
6244 * config/arm/cortex-r4.md: Update with new attributes.
6245 * config/arm/fa526.md: Update with new attributes.
6246 * config/arm/fa606te.md: Update with new attributes.
6247 * config/arm/fa626te.md: Update with new attributes.
6248 * config/arm/fa726te.md: Update with new attributes.
6249
1c83b673
JG		 62502013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6251
6252 * config/aarch64/aarch64-simd.md
6253 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6254 <vwx> iterator to ensure correct register choice.
6255 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6256 (aarch64_sqdmull_n<mode>): Likewise.
6257 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6258 * config/aarch64/arm_neon.h
6259 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6260 (vml<as><q>_n_<su>16): Likewise.
6261 (vml<as>l_high_lane<q>_<su>16): Likewise.
6262 (vml<as>l_high_n_<su>16): Likewise.
6263 (vml<as>l_lane<q>_<su>16): Likewise.
6264 (vml<as>l_n_<su>16): Likewise.
6265 (vmul<q>_lane<q>_<su>16): Likewise.
6266 (vmul<q>_n_<su>16): Likewise.
6267 (vmull_lane<q>_<su>16): Likewise.
6268 (vmull_n_<su>16): Likewise.
6269 (vmull_high_lane<q>_<su>16): Likewise.
6270 (vmull_high_n_<su>16): Likewise.
6271 (vqrdmulh<q>_n_s16): Likewise.
6272
bb1ae543
TB		 62732013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6274
6275 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6276 have the correct lane parameter.
6277
f23c0742
RB		 62782013-09-06 Richard Biener <rguenther@suse.de>
6279
6280 * cfganal.c (control_dependences::~control_dependences):
6281 Properly free all of the vector.
6282
c1d49770
KY		 62832013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6284
6285 PR target/58269
6286 * config/i386/i386.c (ix86_conditional_register_usage):
6287 Proper initialize extended SSE registers.
6288
6da47f52
JH		 62892013-09-06 Jan Hubicka <jh@suse.cz>
6290
6291 PR tree-optimization/58311
6292 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6293
62942013-09-06 Jan Hubicka <jh@suse.cz>
6295
6296 * Makefile.in (tree-sra.o): Update dependencies.
6297 * tree-sra.c: Include ipa-utils.h
6298 (scan_function): Use recursive_call_p.
6299 (has_caller_p): New function.
6300 (cgraph_for_node_and_aliases): Count also callers of aliases.
6301
fc11f321
JH		 63022013-09-06 Jan Hubicka <jh@suse.cz>
6303
6304 PR middle-end/58094
6305 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6306 * tree-tailcall.c: Include ipa-utils.h.
6307 (find_tail_calls): Use it.
6308 * ipa-pure-const.c (check_call): Likewise.
6309 * ipa-utils.c (recursive_call_p): New function.
6310 * ipa-utils.h (recursive_call_p): Dclare.
6311 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6312 (symtab_semantically_equivalent_p): New function.
6313 * Makefile.in (tree-tailcall.o): Update dependencies.
6314
9a6af450
EB		 63152013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6316
6317 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6318 non-inlinable part.
6319
35f32ad4
RB		 63202013-09-06 Richard Biener <rguenther@suse.de>
6321
6322 * lto-streamer.h (lto_global_var_decls): Remove.
6323 * Makefile.in (OBJS): Remove lto-symtab.o.
6324 (lto-symtab.o): Remove.
6325 (GTFILES): Remove lto-symtab.c
6326 * lto-symtab.c: Move to lto/
6327
d12a76f3
AK		 63282013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6329
6330 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6331 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6332 (UNSPEC_FPINT_RINT): New constant definitions.
6333 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6334 definition with 2 attributes.
6335 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6336 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6337 definitions.
6338
63392013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6340
f43245d1 6341 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK		 6342 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6343 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6344 attribute to "z196_alone".
6345 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6346 "zEC12_simple".
6347
c8e9d8c3
RB		 63482013-09-06 Richard Biener <rguenther@suse.de>
6349
6350 * basic-block.h (class control_dependences): New.
6351 * tree-ssa-dce.c (control_dependence_map): Remove.
6352 (cd): New global.
6353 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6354 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6355 find_pdom, find_control_dependence, find_all_control_dependences):
6356 Move to cfganal.c.
f43245d1
UB		 6357 (mark_control_dependent_edges_necessary,
6358 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6359 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB		 6360 * cfganal.c (set_control_dependence_map_bit,
6361 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6362 find_all_control_dependences): Move from tree-ssa-dce.c and
6363 implement as methods of control_dependences class.
6364 (control_dependences::control_dependences): New.
6365 (control_dependences::~control_dependences): Likewise.
6366 (control_dependences::get_edges_dependent_on): Likewise.
6367 (control_dependences::get_edge): Likewise.
6368
7d2ba471
JH		 63692013-09-04 Jan Hubicka <jh@suse.cz>
6370
6371 * tree.c (types_same_for_odr): Drop overactive check.
6372 * ipa-devirt.c (hash_type_name): Likewise.
6373
e18412fc
JH		 63742013-09-04 Jan Hubicka <jh@suse.cz>
6375
6376 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6377 (analyze_functions): ... here.
6378
d352b245
JH		 63792013-09-04 Jan Hubicka <jh@suse.cz>
6380
6381 PR middle-end/58201
6382 * cgraphunit.c (analyze_functions): Clear AUX fields
6383 after processing; initialize assembler name has.
6384
770da076
JL		 63852013-09-05 Jeff Law <law@redhat.com>
6386
6387 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6388 from thread_around_empty_block. Record threading path into PATH.
6389 Recurse if threading through the initial block is successful.
6390 (thread_across_edge): Corresponding changes to slightly simplify.
6391
fe6f68e2
JG		 63922013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6393
6394 * config/aarch64/aarch64.md
6395 (type): Remove frecpe, frecps, frecpx.
6396 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6397 fix to be a TARGET_SIMD instruction.
6398 (aarch64_frecps): Remove.
6399 * config/aarch64/aarch64-simd.md
6400 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6401 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6402
6e4150e1
JG		 64032013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6404 Sofiane Naci <sofiane.naci@arm.com>
6405
f43245d1 6406 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6407 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB		 6408 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6409 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6410 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6411 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6412 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6413 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6414 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6415 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6416 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG		 6417 shift* types.
6418 * config/arm/arm-fixed.md: Update for attribute changes
6419 all occurrences of arlo_* types.
6420 * config/arm/thumb2.md: Update for attribute changes all occurrences
6421 of arlo_* types.
6422 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6423 (cortexa7_older_only): Likewise.
6424 (cortexa7_younger): Likewise.
6425 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6426 (1020alu_shift_op): Likewise.
6427 (1020alu_shift_reg_op): Likewise.
6428 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6429 (alu_shift_op): Likewise.
6430 (alu_shift_reg_op): Likewise.
f43245d1 6431 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG		 6432 (11_alu_shift_op): Likewise.
6433 (11_alu_shift_reg_op): Likewise.
6434 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6435 (9_alu_shift_reg_op): Likewise.
6436 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6437 attribute changes.
6438 (cortex_a15_alu_shift): Likewise.
6439 (cortex_a15_alu_shift_reg): Likewise.
6440 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6441 attribute changes.
6442 (cortex_a5_alu_shift): Likewise.
f43245d1
UB		 6443 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6444 attribute changes.
6e4150e1 6445 (cortex_a53_alu_shift): Likewise.
f43245d1
UB		 6446 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6447 attribute changes.
6e4150e1
JG		 6448 (cortex_a7_alu_reg): Likewise.
6449 (cortex_a7_alu_shift): Likewise.
f43245d1
UB		 6450 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6451 attribute changes.
6e4150e1
JG		 6452 (cortex_a8_alu_shift): Likewise.
6453 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6454 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6455 (cortex_a9_dp_shift): Likewise.
f43245d1
UB		 6456 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6457 attribute changes.
6e4150e1
JG		 6458 * config/arm/cortex-r4.md
6459 (cortex_r4_alu): Update for attribute changes.
6460 (cortex_r4_mov): Likewise.
6461 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6462 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6463 (526_alu_shift_op): Likewise.
f43245d1
UB		 6464 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6465 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6466 (626te_alu_shift_op): Likewise.
f43245d1 6467 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG		 6468 (726te_alu_shift_op): Likewise.
6469 (726te_alu_shift_reg_op): Likewise.
6470 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6471 (mp626_alu_shift_op): Likewise.
6472 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6473 (pj4_alu_conds): Likewise.
6474 (pj4_shift): Likewise.
6475 (pj4_shift_conds): Likewise.
6476 (pj4_alu_shift): Likewise.
6477 (pj4_alu_shift_conds): Likewise.
6478 * config/aarch64/aarch64.md: Update for attribute change
6479 all occurrences of arlo_* and shift* types.
6480
e6523306
MS		 64812013-09-05 Mike Stump <mikestump@comcast.net>
6482
6483 * tree.h: Move documentation for tree_function_decl to tree-core.h
6484 with the declaration.
6485
80a18161
PB		 64862013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6487
6488 PR target/58139
6489 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6490 looking for widest mode.
6491
6994430a
EB		 64922013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6493
6494 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6495
e67f39f7
RB		 64962013-09-05 Richard Biener <rguenther@suse.de>
6497
6498 PR tree-optimization/58137
6499 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6500 Do not create vectors of pointers.
6501 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6502 types for the components of the vector initializer.
6503 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6504 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6505
568cda29
MJ		 65062013-09-05 Martin Jambor <mjambor@suse.cz>
6507
6508 * ipa-prop.c (remove_described_reference): Accept missing references,
6509 return false if that hppens, otherwise return true.
6510 (cgraph_node_for_jfunc): New function.
6511 (try_decrement_rdesc_refcount): Likewise.
6512 (try_make_edge_direct_simple_call): Use them.
6513 (ipa_edge_removal_hook): Remove references from rdescs.
6514 (ipa_edge_duplication_hook): Clone rdescs and their references
6515 when the new edge has the same caller as the old one.
6516 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6517 reference before removing any edges.
6518
5a200acb
RE		 65192013-09-05 Richard Earnshaw <rearnsha@arm.com>
6520
6521 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6522 initial store.
6523 * thumb2.md (thumb2_storewb_parisi): New pattern.
6524
5922847b
YZ		 65252013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6526
6527 * config/aarch64/aarch64-option-extensions.def: Add
6528 AARCH64_OPT_EXTENSION of 'crc'.
6529 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6530 (AARCH64_ISA_CRC): Ditto.
6531 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6532 description of the CRC extension.
6533
996746aa
AI		 65342013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6535
6536 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6537 * config/rs6000/linux.h: Ditto.
6538 * alpha/linux.h: Ditto.
6539 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6540 no_c99_libc_has_function.
6541 * config/c6x/uclinux-elf.h: Ditto.
6542 * config/lm32/uclinux-elf.h: Ditto.
6543 * config/m68k/uclinux.h: Ditto.
6544 * config/moxie/uclinux.h: Ditto.
6545 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6546 (crisv32-*-linux*, cris-*-linux*): Ditto.
6547 * config/bfin/bfin.c: Include "tm_p.h".
6548
c370d99f
RB		 65492013-09-05 Richard Biener <rguenther@suse.de>
6550
6551 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6552 check for a definition without a basic-block.
6553
003bb7f3
JG		 65542013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6555 Sofiane Naci <sofiane.naci@arm.com>
6556
6557 * config/aarch64/aarch64.md
6558 (*movti_aarch64): Rename r_2_f and f_2_r.
6559 (*movsf_aarch64): Likewise.
6560 (*movdf_aarch64): Likewise.
6561 (*movtf_aarch64): Likewise.
6562 (aarch64_movdi_<mode>low): Likewise.
6563 (aarch64_movdi_<mode>high): Likewise.
6564 (aarch64_mov<mode>high_di): Likewise.
6565 (aarch64_mov<mode>low_di): Likewise.
6566 (aarch64_movtilow_tilow): Likewise.
6567 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6568 values to config/arm/types.md
6569 (attribute "conds"): Update for attribute change.
6570 (anddi3_insn): Likewise.
6571 (iordi3_insn): Likewise.
6572 (xordi3_insn): Likewise.
6573 (one_cmpldi2): Likewise.
6574 * config/arm/types.md (type): Add Neon types.
6575 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6576 use "type" attribute.
6577 (movmisalign<mode>_neon_store): Likewise.
6578 (movmisalign<mode>_neon_load): Likewise.
6579 (vec_set<mode>_internal): Likewise.
6580 (vec_setv2di_internal): Likewise.
6581 (vec_extract<mode>): Likewise.
6582 (vec_extractv2di): Likewise.
6583 (add<mode>3_neon): Likewise.
6584 (adddi3_neon): Likewise.
6585 (sub<mode>3_neon): Likewise.
6586 (subdi3_neon): Likewise.
6587 (mul<mode>3_neon): Likewise.
6588 (mul<mode>3add<mode>_neon): Likewise.
6589 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6590 (fma<VCVTF:mode>4)): Likewise.
6591 (fma<VCVTF:mode>4_intrinsic): Likewise.
6592 (fmsub<VCVTF:mode>4)): Likewise.
6593 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6594 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6595 (ior<mode>3): Likewise.
6596 (and<mode>3): Likewise.
6597 (anddi3_neon): Likewise.
6598 (orn<mode>3_neon): Likewise.
6599 (orndi3_neon): Likewise.
6600 (bic<mode>3_neon): Likewise.
6601 (bicdi3_neon): Likewise.
6602 (xor<mode>3): Likewise.
6603 (one_cmpl<mode>2): Likewise.
6604 (abs<mode>2): Likewise.
6605 (neg<mode>2): Likewise.
6606 (umin<mode>3_neon): Likewise.
6607 (umax<mode>3_neon): Likewise.
6608 (smin<mode>3_neon): Likewise.
6609 (smax<mode>3_neon): Likewise.
6610 (vashl<mode>3): Likewise.
6611 (vashr<mode>3_imm): Likewise.
6612 (vlshr<mode>3_imm): Likewise.
6613 (ashl<mode>3_signed): Likewise.
6614 (ashl<mode>3_unsigned): Likewise.
6615 (neon_load_count): Likewise.
6616 (ashldi3_neon_noclobber): Likewise.
6617 (signed_shift_di3_neon): Likewise.
6618 (unsigned_shift_di3_neon): Likewise.
6619 (ashrdi3_neon_imm_noclobber): Likewise.
6620 (lshrdi3_neon_imm_noclobber): Likewise.
6621 (widen_ssum<mode>3): Likewise.
6622 (widen_usum<mode>3): Likewise.
6623 (quad_halves_<code>v4si): Likewise.
6624 (quad_halves_<code>v4sf): Likewise.
6625 (quad_halves_<code>v8hi): Likewise.
6626 (quad_halves_<code>v16qi): Likewise.
6627 (reduc_splus_v2di): Likewise.
6628 (neon_vpadd_internal<mode>): Likewise.
6629 (neon_vpsmin<mode>): Likewise.
6630 (neon_vpsmax<mode>): Likewise.
6631 (neon_vpumin<mode>): Likewise.
6632 (neon_vpumax<mode>): Likewise.
6633 (ss_add<mode>_neon): Likewise.
6634 (us_add<mode>_neon): Likewise.
6635 (ss_sub<mode>_neon): Likewise.
6636 (us_sub<mode>_neon): Likewise.
6637 (neon_vadd<mode>_unspec): Likewise.
6638 (neon_vaddl<mode>): Likewise.
6639 (neon_vaddw<mode>): Likewise.
6640 (neon_vhadd<mode>): Likewise.
6641 (neon_vqadd<mode>): Likewise.
6642 (neon_vaddhn<mode>): Likewise.
6643 (neon_vmul<mode>): Likewise.
6644 (neon_vmla<mode>): Likewise.
6645 (neon_vmlal<mode>): Likewise.
6646 (neon_vmls<mode>): Likewise.
6647 (neon_vmlsl<mode>): Likewise.
6648 (neon_vqdmulh<mode>): Likewise.
6649 (neon_vqdmlal<mode>): Likewise.
6650 (neon_vqdmlsl<mode>): Likewise.
6651 (neon_vmull<mode>): Likewise.
6652 (neon_vqdmull<mode>): Likewise.
6653 (neon_vsub<mode>_unspec): Likewise.
6654 (neon_vsubl<mode>): Likewise.
6655 (neon_vsubw<mode>): Likewise.
6656 (neon_vqsub<mode>): Likewise.
6657 (neon_vhsub<mode>): Likewise.
6658 (neon_vsubhn<mode>): Likewise.
6659 (neon_vceq<mode>): Likewise.
6660 (neon_vcge<mode>): Likewise.
6661 (neon_vcgeu<mode>): Likewise.
6662 (neon_vcgt<mode>): Likewise.
6663 (neon_vcgtu<mode>): Likewise.
6664 (neon_vcle<mode>): Likewise.
6665 (neon_vclt<mode>): Likewise.
6666 (neon_vcage<mode>): Likewise.
6667 (neon_vcagt<mode>): Likewise.
6668 (neon_vtst<mode>): Likewise.
6669 (neon_vabd<mode>): Likewise.
6670 (neon_vabdl<mode>): Likewise.
6671 (neon_vaba<mode>): Likewise.
6672 (neon_vabal<mode>): Likewise.
6673 (neon_vmax<mode>): Likewise.
6674 (neon_vmin<mode>): Likewise.
6675 (neon_vpaddl<mode>): Likewise.
6676 (neon_vpadal<mode>): Likewise.
6677 (neon_vpmax<mode>): Likewise.
6678 (neon_vpmin<mode>): Likewise.
6679 (neon_vrecps<mode>): Likewise.
6680 (neon_vrsqrts<mode>): Likewise.
6681 (neon_vqabs<mode>): Likewise.
6682 (neon_vqneg<mode>): Likewise.
6683 (neon_vcls<mode>): Likewise.
6684 (clz<mode>2): Likewise.
6685 (popcount<mode>2): Likewise.
6686 (neon_vrecpe): Likewise.
6687 (neon_vrsqrte): Likewise.
6688 (neon_vget_lane<mode>_sext_internal): Likewise.
6689 (neon_vget_lane<mode>_zext_internal): Likewise.
6690 (neon_vdup_n<mode>): Likewise.
6691 (neon_vdup_nv2di): Likewise.
6692 (neon_vdpu_lane<mode>_internal): Likewise.
6693 (neon_vswp<mode>): Likewise.
6694 (float<mode><V_cvtto>2): Likewise.
6695 (floatuns<mode><V_cvtto>2): Likewise.
6696 (fix_trunc<mode><V_cvtto>)2): Likewise
6697 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6698 (neon_vcvt<mode>): Likewise.
6699 (neon_vcvtv4sfv4hf): Likewise.
6700 (neon_vcvtv4hfv4sf): Likewise.
6701 (neon_vcvt_n<mode>): Likewise.
6702 (neon_vmovn<mode>): Likewise.
6703 (neon_vqmovn<mode>): Likewise.
6704 (neon_vqmovun<mode>): Likewise.
6705 (neon_vmovl<mode>): Likewise.
6706 (neon_vmul_lane<mode>): Likewise.
6707 (neon_vmull_lane<mode>): Likewise.
6708 (neon_vqdmull_lane<mode>): Likewise.
6709 (neon_vqdmulh_lane<mode>): Likewise.
6710 (neon_vmla_lane<mode>): Likewise.
6711 (neon_vmlal_lane<mode>): Likewise.
6712 (neon_vqdmlal_lane<mode>): Likewise.
6713 (neon_vmls_lane<mode>): Likewise.
6714 (neon_vmlsl_lane<mode>): Likewise.
6715 (neon_vqdmlsl_lane<mode>): Likewise.
6716 (neon_vext<mode>): Likewise.
6717 (neon_vrev64<mode>): Likewise.
6718 (neon_vrev32<mode>): Likewise.
6719 (neon_vrev16<mode>): Likewise.
6720 (neon_vbsl<mode>_internal): Likewise.
6721 (neon_vshl<mode>): Likewise.
6722 (neon_vqshl<mode>): Likewise.
6723 (neon_vshr_n<mode>): Likewise.
6724 (neon_vshrn_n<mode>): Likewise.
6725 (neon_vqshrn_n<mode>): Likewise.
6726 (neon_vqshrun_n<mode>): Likewise.
6727 (neon_vshl_n<mode>): Likewise.
6728 (neon_vqshl_n<mode>): Likewise.
6729 (neon_vqshlu_n<mode>): Likewise.
6730 (neon_vshll_n<mode>): Likewise.
6731 (neon_vsra_n<mode>): Likewise.
6732 (neon_vsri_n<mode>): Likewise.
6733 (neon_vsli_n<mode>): Likewise.
6734 (neon_vtbl1v8qi): Likewise.
6735 (neon_vtbl2v8qi): Likewise.
6736 (neon_vtbl3v8qi): Likewise.
6737 (neon_vtbl4v8qi): Likewise.
6738 (neon_vtbx1v8qi): Likewise.
6739 (neon_vtbx2v8qi): Likewise.
6740 (neon_vtbx3v8qi): Likewise.
6741 (neon_vtbx4v8qi): Likewise.
6742 (neon_vtrn<mode>_internal): Likewise.
6743 (neon_vzip<mode>_internal): Likewise.
6744 (neon_vuzp<mode>_internal): Likewise.
6745 (neon_vld1<mode>): Likewise.
6746 (neon_vld1_lane<mode>): Likewise.
6747 (neon_vld1_dup<mode>): Likewise.
6748 (neon_vld1_dupv2di): Likewise.
6749 (neon_vst1<mode>): Likewise.
6750 (neon_vst1_lane<mode>): Likewise.
6751 (neon_vld2<mode>): Likewise.
6752 (neon_vld2_lane<mode>): Likewise.
6753 (neon_vld2_dup<mode>): Likewise.
6754 (neon_vst2<mode>): Likewise.
6755 (neon_vst2_lane<mode>): Likewise.
6756 (neon_vld3<mode>): Likewise.
6757 (neon_vld3qa<mode>): Likewise.
6758 (neon_vld3qb<mode>): Likewise.
6759 (neon_vld3_lane<mode>): Likewise.
6760 (neon_vld3_dup<mode>): Likewise.
6761 (neon_vst3<mode>): Likewise.
6762 (neon_vst3qa<mode>): Likewise.
6763 (neon_vst3qb<mode>): Likewise.
6764 (neon_vst3_lane<mode>): Likewise.
6765 (neon_vld4<mode>): Likewise.
6766 (neon_vld4qa<mode>): Likewise.
6767 (neon_vld4qb<mode>): Likewise.
6768 (neon_vld4_lane<mode>): Likewise.
6769 (neon_vld4_dup<mode>): Likewise.
6770 (neon_vst4<mode>): Likewise.
6771 (neon_vst4qa<mode>): Likewise.
6772 (neon_vst4qb<mode>): Likewise.
6773 (neon_vst4_lane<mode>): Likewise.
6774 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6775 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6776 (neon_vec_<US>mult_lo_<mode>): Likewise.
6777 (neon_vec_<US>mult_hi_<mode>): Likewise.
6778 (neon_vec_<US>shiftl_<mode>): Likewise.
6779 (neon_unpack<US>_<mode>): Likewise.
6780 (neon_vec_<US>mult_<mode>): Likewise.
6781 (vec_pack_trunc_<mode>): Likewise.
6782 (neon_vec_pack_trunk_<mode>): Likewise.
6783 (neon_vabd<mode>_2): Likewise.
6784 (neon_vabd<mode>_3): Likewise.
6785 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6786 (thumb2_movsi_vfp): Likewise.
6787 (movdi_vfp): Likewise.
6788 (movdi_vfp_cortexa8): Likewise.
6789 (movhf_vfp_neon): Likewise.
6790 (movhf_vfp): Likewiwse.
6791 (movsf_vfp): Likewiwse.
6792 (thumb2_movsf_vfp): Likewiwse.
6793 (movdf_vfp): Likewise.
6794 (thumb2_movdf_vfp): Likewise.
6795 (movsfcc_vfp): Likewise.
6796 (thumb2_movsfcc_vfp): Likewise.
6797 (movdfcc_vfp): Likewise.
6798 (thumb2_movdfcc_vfp): Likewise.
6799 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6800 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6801 (v10_v2c): Likewise.
6802 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6803 attribute change.
6804 (cortex_a15_neon_int_2): Likewise.
6805 (cortex_a15_neon_int_3): Likewise.
6806 (cortex_a15_neon_int_4): Likewise.
6807 (cortex_a15_neon_int_5): Likewise.
6808 (cortex_a15_neon_vqneg_vqabs): Likewise.
6809 (cortex_a15_neon_vmov): Likewise.
6810 (cortex_a15_neon_vaba): Likewise.
6811 (cortex_a15_neon_vaba_qqq): Likewise.
6812 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6813 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6814 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6815 scalar_64_32_long_scalar): Likewise.
6816 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6817 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6818 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6819 lotype_qdd_64_32_long): Likewise.
6820 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6821 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6822 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6823 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6824 (cortex_a15_neon_shift_1): Likewise.
6825 (cortex_a15_neon_shift_2): Likewise.
6826 (cortex_a15_neon_shift_3): Likewise.
6827 (cortex_a15_neon_vshl_ddd): Likewise.
6828 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6829 (cortex_a15_neon_vsra_vrsra): Likewise.
6830 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6831 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6832 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6833 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6834 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6835 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6836 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6837 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6838 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6839 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6840 (cortex_a15_neon_bp_simple): Likewise.
6841 (cortex_a15_neon_bp_2cycle): Likewise.
6842 (cortex_a15_neon_bp_3cycle): Likewise.
6843 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6844 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6845 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6846 (cortex_a15_neon_vld2_4_regs): Likewise.
6847 (cortex_a15_neon_vld3_vld4): Likewise.
6848 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6849 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6850 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6851 (cortex_a15_neon_vst3_vst4): Likewise.
6852 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6853 (cortex_a15_neon_vld3_vld4_lane" 10
6854 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6855 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6856 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6857 (cortex_a15_neon_ldm_2): Likewise.0
6858 (cortex_a15_neon_stm_2): Likewise.
6859 (cortex_a15_neon_mcr): Likewise.
6860 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6861 (cortex_a15_neon_mrc): Likewise.
6862 (cortex_a15_neon_mrrc): Likewise.
6863 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6864 change.
6865 (cortex_a15_alu_shift): Likewise.
6866 (cortex_a15_alu_shift_reg): Likewise.
6867 (cortex_a15_mult32): Likewise.
6868 (cortex_a15_mult64): Likewise.
6869 (cortex_a15_block): Likewise.
6870 (cortex_a15_branch): Likewise.
6871 (cortex_a15_load1): Likewise.
6872 (cortex_a15_load3): Likewise.
6873 (cortex_a15_store1): Likewise.
6874 (cortex_a15_store3): Likewise.
6875 (cortex_a15_call): Likewise.
f43245d1 6876 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG		 6877 (cortex_a5_f2r): Likewise.
6878 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6879 change.
6880 (cortex_a53_f2r): Likewise.
6881 * config/arm/cortex-a7.md
6882 (cortex_a7_branch): Update for attribute change.
6883 (cortex_a7_call): Likewise.
6884 (cortex_a7_alu_imm): Likewise.
6885 (cortex_a7_alu_reg): Likewise.
6886 (cortex_a7_alu_shift): Likewise.
6887 (cortex_a7_mul): Likewise.
6888 (cortex_a7_load1): Likewise.
6889 (cortex_a7_store1): Likewise.
6890 (cortex_a7_load2): Likewise.
6891 (cortex_a7_store2): Likewise.
6892 (cortex_a7_load3): Likewise.
6893 (cortex_a7_store3): Likewise.
6894 (cortex_a7_load4): Likewise.
6895 (cortex_a7_store4): Likewise.
6896 (cortex_a7_fpalu): Likewise.
6897 (cortex_a7_fconst): Likewise.
6898 (cortex_a7_fpmuls): Likewise.
6899 (cortex_a7_neon_mul): Likewise.
6900 (cortex_a7_fpmacs): Likewise.
6901 (cortex_a7_neon_mla: Likewise.
6902 (cortex_a7_fpmuld: Likewise.
6903 (cortex_a7_fpmacd: Likewise.
6904 (cortex_a7_fpfmad: Likewise.
6905 (cortex_a7_fdivs: Likewise.
6906 (cortex_a7_fdivd: Likewise.
6907 (cortex_a7_r2f: Likewise.
6908 (cortex_a7_f2r: Likewise.
6909 (cortex_a7_f_flags: Likewise.
6910 (cortex_a7_f_loads: Likewise.
6911 (cortex_a7_f_loadd: Likewise.
6912 (cortex_a7_f_stores: Likewise.
6913 (cortex_a7_f_stored: Likewise.
6914 (cortex_a7_neon): Likewise.
6915 * config/arm/cortex-a8-neon.md
6916 (cortex_a8_neon_mrc): Update for attribute change.
6917 (cortex_a8_neon_mrrc): Likewise.
6918 (cortex_a8_neon_int_1): Likewise.
6919 (cortex_a8_neon_int_2): Likewise.
6920 (cortex_a8_neon_int_3): Likewise.
6921 (cortex_a8_neon_int_4): Likewise.
6922 (cortex_a8_neon_int_5): Likewise.
6923 (cortex_a8_neon_vqneg_vqabs): Likewise.
6924 (cortex_a8_neon_vmov): Likewise.
6925 (cortex_a8_neon_vaba): Likewise.
6926 (cortex_a8_neon_vaba_qqq): Likewise.
6927 (cortex_a8_neon_vsma): Likewise.
6928 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6929 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 6930 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6931 long_scalar): Likewise.
003bb7f3
JG		 6932 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6933 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6934 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6935 long_scalar_qdd_64_32_long): Likewise.
6936 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6937 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6938 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6939 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6940 (cortex_a8_neon_shift_1): Likewise.
6941 (cortex_a8_neon_shift_2): Likewise.
6942 (cortex_a8_neon_shift_3): Likewise.
6943 (cortex_a8_neon_vshl_ddd): Likewise.
6944 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6945 (cortex_a8_neon_vsra_vrsra): Likewise.
6946 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6947 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6948 (cortex_a8_neon_fp_vsum): Likewise.
6949 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6950 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6951 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6952 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6953 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6954 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6955 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6956 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6957 (cortex_a8_neon_bp_simple): Likewise.
6958 (cortex_a8_neon_bp_2cycle): Likewise.
6959 (cortex_a8_neon_bp_3cycle): Likewise.
6960 (cortex_a8_neon_ldr): Likewise.
6961 (cortex_a8_neon_str): Likewise.
6962 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6963 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6964 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6965 (cortex_a8_neon_vld2_4_regs): Likewise.
6966 (cortex_a8_neon_vld3_vld4): Likewise.
6967 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6968 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6969 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6970 (cortex_a8_neon_vst3_vst4): Likewise.
6971 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6972 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6973 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6974 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6975 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6976 (cortex_a8_neon_mcr): Likewise.
6977 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 6978 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG		 6979 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6980 change.
6981 (ca9_neon_mrrc): Likewise.
6982 (cortex_a9_neon_int_1): Likewise.
6983 (cortex_a9_neon_int_2): Likewise.
6984 (cortex_a9_neon_int_3): Likewise.
6985 (cortex_a9_neon_int_4): Likewise.
6986 (cortex_a9_neon_int_5): Likewise.
6987 (cortex_a9_neon_vqneg_vqabs): Likewise.
6988 (cortex_a9_neon_vmov): Likewise.
6989 (cortex_a9_neon_vaba): Likewise.
6990 (cortex_a9_neon_vaba_qqq): Likewise.
6991 (cortex_a9_neon_vsma): Likewise.
6992 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6993 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 6994 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6995 long_scalar): Likewise.
003bb7f3
JG		 6996 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6997 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6998 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6999 long_scalar_qdd_64_32_long): Likewise.
7000 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7001 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7002 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7003 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7004 (cortex_a9_neon_shift_1): Likewise.
7005 (cortex_a9_neon_shift_2): Likewise.
7006 (cortex_a9_neon_shift_3): Likewise.
7007 (cortex_a9_neon_vshl_ddd): Likewise.
7008 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7009 (cortex_a9_neon_vsra_vrsra): Likewise.
7010 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7011 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7012 (cortex_a9_neon_fp_vsum): Likewise.
7013 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7014 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7015 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7016 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7017 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7018 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7019 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7020 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7021 (cortex_a9_neon_bp_simple): Likewise.
7022 (cortex_a9_neon_bp_2cycle): Likewise.
7023 (cortex_a9_neon_bp_3cycle): Likewise.
7024 (cortex_a9_neon_ldr): Likewise.
7025 (cortex_a9_neon_str): Likewise.
7026 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7027 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7028 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7029 (cortex_a9_neon_vld2_4_regs): Likewise.
7030 (cortex_a9_neon_vld3_vld4): Likewise.
7031 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7032 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7033 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7034 (cortex_a9_neon_vst3_vst4): Likewise.
7035 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7036 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7037 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7038 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7039 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7040 (cortex_a9_neon_mcr): Likewise.
7041 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7042 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7043 (cortex_a9_fps): Likewise.
7044 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7045 change.
7046 (cortex_m4_fmuls): Likewise.
7047 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7048 change.
7049 (cortex_r4_mrc): Likewise.
7050 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 7051 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 7052 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB		 7053 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7054 attribute change.
003bb7f3
JG		 7055 (pj4_core_to_vfp): Likewise.
7056 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7057 attribute change.
7058 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7059 (vfp_fstore): Likewise.
7060 * doc/md.texi: Change references to neon_type to refer to type.
7061
fbe0dc78
DS		 70622013-09-04 Dodji Seketeli <dodji@redhat.com>
7063
7064 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7065
c167bc5b
DE		 70662013-09-04 David Edelsohn <dje.gcc@gmail.com>
7067
7068 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7069 lglobl if not weak.
7070
035cb59f
ER		 70712013-09-04 Easwaran Raman <eraman@google.com>
7072
7073 PR middle-end/57370
7074 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7075 (build_and_add_sum): Use it.
7076 (appears_later_in_bb): Simplify code.
7077
71191083
TJ		 70782013-09-04 Teresa Johnson <tejohnson@google.com>
7079
7080 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7081
0c488213
JG		 70822013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7083
7084 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7085
61a74079
JH		 70862013-09-04 Jan Hubicka <jh@suse.cz>
7087
7088 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7089 * ipa-devirt.c: Include diganostic.h
7090 (odr_type_d): Add types and types_set.
7091 (hash_type_name): Work for types with vtables during LTO.
7092 (odr_hasher::remove): Fix comment; destroy types_set.
7093 (add_type_duplicate): New function,
7094 (get_odr_type): Use it.
7095 (dump_type_inheritance_graph): Dump type duplicates.
7096 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7097 graph.
7098 * tree.c (types_same_for_odr): Give exact answers on types with
7099 virtual tables.
7100
24ba8195
DS		 71012013-09-04 Dodji Seketeli <dodji@redhat.com>
7102
7103 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 7104 explaining their differences.
24ba8195 7105
0a04869e
SKS		 71062013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7107
7108 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7109
3b18bc42
JL		 71102013-09-03 Jeff Law <law@redhat.com>
7111
7112 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7113 when not threading through a joiner block. Pass joiner/no joiner
7114 state to register_jump_thread.
7115 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7116 state from argument rather than implying on path length.
7117 Dump the entire jump thread path into debugging dump.
7118 * tree-flow.h (register_jump_thread): Update prototype.
7119
ae84e151
XDL		 71202013-08-29 Xinliang David Li <davidxl@google.com>
7121
7122 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7123 Remove a trivial gcc_assert.
7124
c716e67f
XDL		 71252013-08-29 Xinliang David Li <davidxl@google.com>
7126
7127 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7128 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7129 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7130 Delay base decl alignment adjustment.
7131 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7132 * tree-vectorizer.h: New data structure.
7133 (set_dr_misalignment): New function.
7134 (dr_misalignment): Ditto.
7135 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7136 (vectorizable_load): Ditto.
7137 (ensure_base_align): New function.
7138 (vectorize_loops): Add dbg_cnt support.
7139 (execute_vect_slp): Ditto.
7140 * dbgcnt.def: New debug counter.
7141 * Makefile: New dependency.
7142
db8800bc
MI		 71432013-09-03 Meador Inge <meadori@codesourcery.com>
7144
7145 Revert:
7146
7147 2013-08-30 Meador Inge <meadori@codesourcery.com>
7148
7149 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7150
0d9b4f5c
DE		 71512013-09-03 David Edelsohn <dje.gcc@gmail.com>
7152
7153 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7154 function descriptor.
7155
29e810b9
RB		 71562013-09-03 Richard Biener <rguenther@suse.de>
7157
7158 * tree-affine.c (add_elt_to_tree): Fix association issue,
7159 avoid useless converts and make sure to always return a
7160 properly typed result.
7161
dbfc2894
RB		 71622013-09-03 Richard Biener <rguenther@suse.de>
7163
7164 PR middle-end/57656
7165 * fold-const.c (negate_expr_p): Fix division case.
7166 (negate_expr): Likewise.
7167
fa3bf4ce
RB		 71682013-09-03 Richard Biener <rguenther@suse.de>
7169
7170 PR lto/58285
7171 * tree-streamer-out.c: Include tm.h.
7172 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7173
a64bbb3f
JH		 71742013-09-03 Jan Hubicka <jh@suse.cz>
7175
7176 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7177
93ed07e2
AM		 71782013-09-03 Alan Modra <amodra@gmail.com>
7179
7180 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7181 multilibs to big-endian target and vice versa.
7182 * config/rs6000/t-linux64: Use := assignment on all vars.
7183 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7184 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7185 * config/rs6000/t-linux64le: New file.
7186 * config/rs6000/t-linux64bele: New file.
7187 * config/rs6000/t-linux64lebe: New file.
7188
07d6cd64
JH		 71892013-09-02 Jan Hubicka <jh@suse.cz>
7190
7191 * ipa-inline-transform.c (inline_transform): Do not
7192 optimize_inline_calls when not optimizing.
7193
8e4c9a10
JH		 71942013-09-02 Jan Hubicka <jh@suse.cz>
7195
7196 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7197 duplicated nodes for assembler names.
7198 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7199 hard registers from assembler name hash.
7200
c91061e6
JH		 72012013-09-02 Jan Hubicka <jh@suse.cz>
7202
7203 * ipa-split.c (execute_split_functions): Split externally visible
7204 functions called once.
7205
2fd0985c
MJ		 72062013-09-02 Martin Jambor <mjambor@suse.cz>
7207
7208 PR ipa/58106
7209 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7210 linked list. When finding the correct duplicate, also consider also
7211 the caller in additon to its inlined_to node.
7212
91bd4114
JG		 72132013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7214
7215 * config/aarch64/aarch64-simd-builtins.def
7216 (dup_lane_scalar): Remove.
7217 * config/aarch64/aarch64-simd.md
7218 (aarch64_simd_dup): Add 'w->w' alternative.
7219 (aarch64_dup_lane<mode>): Allow for VALL.
7220 (aarch64_dup_lane_scalar<mode>): Remove.
7221 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7222 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7223 (aarch64_get_lane_unsigned<mode>): Likewise.
7224 (aarch64_get_lane<mode>): Likewise.
7225 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7226 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7227 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7228 (VCON): Change container of V2SF.
7229 (vswap_width_name): Likewise.
7230 * config/aarch64/arm_neon.h
7231 (__aarch64_vdup_lane_any): New.
7232 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7233 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7234 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7235
d617d2d8
EB		 72362013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7237
7238 PR middle-end/56382
7239 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7240 the source or the destination is a single hard register.
7241
63fdb7be
RB		 72422013-09-02 Richard Biener <rguenther@suse.de>
7243
7244 PR middle-end/57511
7245 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7246 non-linear SCEVs.
7247
78de2333
RB		 72482013-09-02 Richard Biener <rguenther@suse.de>
7249
7250 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7251 arithmetic to sizetype.
7252
85ff4ec6
BC		 72532013-09-02 Bin Cheng <bin.cheng@arm.com>
7254
7255 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7256 Find auto-increment use both before and after candidate.
7257
576016fe
MP		 72582013-09-02 Marek Polacek <polacek@redhat.com>
7259
7260 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7261
08f835dc
JH		 72622013-09-01 Jan Hubicka <jh@suse.cz>
7263
7264 * Makefile.in: Add ipa-profile.o
7265 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7266 * cgraph.c (struct cgraph_propagate_frequency_data,
7267 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7268 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7269 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB		 7270 * ipa-inline-analysis.c: Include ipa-utils.h;
7271 drop duplicated cfgloop.h.
08f835dc
JH		 7272 (inline_update_callee_summaries): Update.
7273 * ipa-profile.c: New file.
7274 * ipa-utils.h (ipa_propagate_frequency): Declare.
7275 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 7276 data-streamer.h, value-prof.h.
08f835dc
JH		 7277 (symtab_remove_unreachable_nodes): Update profile.
7278 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7279 account_time_size, cmp_counts, dump_histogram,
7280 ipa_profile_generate_summary, ipa_profile_write_summary,
7281 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7282 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 7283 Move to ipa-profile.c.
08f835dc 7284
cec8583c
JDA		 72852013-09-01 John David Anglin <danglin@gcc.gnu.org>
7286
7287 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7288
bbc9396b
JH		 72892013-09-01 Jan Hubicka <jh@suse.cz>
7290
7291 * common.opt (fdevirtualize-speculatively): New function.
7292 * invoke.texi (fdevirtualize-speculatively): Document.
7293 * ipa-devirt.c: Include ipa-inline.h
7294 (likely_target_p): New function.
7295 (ipa_devirt): New function.
7296 (gate_ipa_devirt): New function.
7297 (pass_data_ipa_devirt): New static var.
7298 (pass_ipa_devirt): Likewise.
7299 (make_pass_ipa_devirt): New function.
7300 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7301 (common_handle_option): Disable devirtualization when
7302 value range profiling is available.
7303 * passes.def (pass_ipa_devirt): Add.
7304 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 7305 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 7306
0cea211e
IS		 73072013-09-01 Iain Sandoe <iain@codesourcery.com>
7308
7309 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7310 include sanitize(undefined).
7311
5e351e96
DN		 73122013-08-31 Diego Novillo <dnovillo@google.com>
7313
7314 * Makefile.in (TREE_CORE_H): Define.
7315 (TREE_H): Use.
7316 (GTFILES): Add tree-core.h.
7317 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7318 size the array.
7319 * tree-core.h: New file.
7320 Move all data structures, enum, typedefs, global
7321 declarations and constants from ...
7322 * tree.h: ... here.
7323
63bf9a90
JH		 73242013-08-31 Jan Hubicka <jh@suse.cz>
7325
7326 * bulitins.c (expand_builtin): Do not early exit for gcov
7327 instrumented functions.
7328
5c0abd6a
MP		 73292013-08-31 Marek Polacek <polacek@redhat.com>
7330
7331 * ubsan.c: Include tm_p.h.
7332
4843f032
JH		 73332013-08-31 Jan Hubicka <jh@suse.cz>
7334
8356c89c
JH		 7335 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7336 warning.
7337
4843f032
JH		 7338 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7339 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7340 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 7341 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7342 (ipa_merge_profiles): New function.
7343 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7344 (lto_input_function_body): Likewise.
5e581212
JH		 7345 * ipa-utils.h (ipa_merge_profiles): Declare.
7346 * lto-streamer.h (lto_input_function_body): Update prototype.
7347 (emit_label_in_global_context_p): Remove.
7348 * lto-symtab.c: Include ipa-utils.h
7349 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7350
57292ce9
JH		 73512013-08-31 Jan Hubicka <jh@suse.cz>
7352
7353 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7354
b9cbfeeb
JH		 73552013-08-31 Jan Hubicka <jh@suse.cz>
7356
7357 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7358
41dedebd
UB		 73592013-08-31 Uros Bizjak <ubizjak@gmail.com>
7360
7361 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7362 "cmp" RTX before signed_comparison_operator check to account
7363 for "code" changes.
7364
06d65050
JH		 73652013-08-30 Jan Hubicka <jh@suse.cz>
7366
41dedebd 7367 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 7368 (detect_type_change_1): Rename to ...
7369 (detect_type_change): ... this one; early return on non-polymorphic
7370 types.
41dedebd 7371 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 7372 use of detect_type_change.
7373 (compute_complex_assign_jump_func): Add param_type parameter;
7374 update use of detect_type_change_ssa.
7375 (compute_complex_ancestor_jump_func): Likewise.
7376 (ipa_get_callee_param_type): New function.
7377 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7378 update calls to the jump function computation functions.
7379
600b5b1d
TJ		 73802013-08-30 Teresa Johnson <tejohnson@google.com>
7381 Steven Bosscher <steven@gcc.gnu.org>
7382
7383 * cfgrtl.c (fixup_new_cold_bb): New routine.
7384 (commit_edge_insertions): Invoke fixup_partitions.
7385 (find_partition_fixes): New routine.
7386 (fixup_partitions): Ditto.
7387 (verify_hot_cold_block_grouping): Update comments.
7388 (rtl_verify_edges): Invoke find_partition_fixes.
7389 (rtl_verify_bb_pointers): Update comments.
7390 (rtl_verify_bb_layout): Ditto.
7391 * basic-block.h (probably_never_executed_edge_p): Declare.
7392 (fixup_partitions): Ditto.
7393 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7394 * bb-reorder.c (sanitize_hot_paths): New function.
7395 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7396 sanitize_hot_paths.
7397 * predict.c (probably_never_executed_edge_p): New routine.
7398 * cfg.c (check_bb_profile): Add partition insanity warnings.
7399
0a6ea5c9
MI		 74002013-08-30 Meador Inge <meadori@codesourcery.com>
7401
7402 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7403
de5a5fa1
MP		 74042013-08-30 Marek Polacek <polacek@redhat.com>
7405
7406 * Makefile.in (ubsan.o): Add.
7407 (c-family/c-ubsan.o): Add.
7408 (builtins.o): Add ubsan.h dependency.
7409 * ubsan.h: New file.
7410 * ubsan.c: New file.
7411 * common.opt: Add -fsanitize=undefined option.
7412 (flag_sanitize): Add variable.
7413 (fsanitize=): Add option. Add Driver.
7414 (fsanitize=thread): Remove option.
7415 (fsanitize=address): Likewise.
7416 (static-libubsan): New option.
7417 * doc/invoke.texi: Document the new flag and -static-libubsan.
7418 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7419 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7420 * builtin-attrs.def (ATTR_COLD): Define.
7421 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7422 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7423 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7424 * flag-types.h (sanitize_code): New enum.
7425 * opts.c (common_handle_option): Parse command line arguments
7426 of -fsanitize=. Add -fsanitize=unreachable option.
7427 * varasm.c (get_variable_section): Adjust.
7428 (assemble_noswitch_variable): Likewise.
7429 (assemble_variable): Likewise.
7430 (output_constant_def_contents): Likewise.
7431 (categorize_decl_for_section): Likewise.
7432 (place_block_symbol): Likewise.
7433 (output_object_block): Likewise.
7434 * builtins.def: Likewise.
7435 * toplev.c (compile_file): Likewise.
7436 (process_options): Likewise.
7437 * cppbuiltin.c: Likewise.
7438 * tsan.c (tsan_pass): Likewise.
7439 (tsan_gate): Likewise.
7440 (tsan_gate_O0): Likewise.
7441 * cfgexpand.c (partition_stack_vars): Likewise.
7442 (expand_stack_vars): Likewise.
7443 (defer_stack_allocation): Likewise.
7444 (expand_used_vars): Likewise.
7445 * cfgcleanup.c (old_insns_match_p): Likewise.
7446 * asan.c (asan_finish_file): Likewise.
7447 (asan_instrument): Likewise.
7448 (gate_asan): Likewise.
7449 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7450 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7451 (asan_global_struct): Use pointer_sized_int_node instead
7452 calling build_nonstandard_integer_type.
7453 (initialize_sanitizer_builtins): Likewise.
7454 (asan_finish_file): Likewise.
7455 * gcc.c: Document %{%:function(args):X}.
7456 (static_spec_functions): Add sanitize.
7457 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7458 store funcval != NULL there.
7459 (do_spec_1): Adjust handle_spec_function caller.
7460 (handle_braces): Allow %:function(args) as condition.
7461 (sanitize_spec_function): New function.
7462 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7463 (LIBUBSAN_SPEC): Likewise.
7464 (LIBUBSAN_EARLY_SPEC): Likewise.
7465 (SANITIZER_SPEC): Handle libubsan.
7466 (SANITIZER_EARLY_SPEC): Likewise.
7467 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7468 instead of fsanitize=address.
7469 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7470 instead of fsanitize=address*.
7471 * builtins.c: Include ubsan.h.
7472 (fold_builtin_0): Instrument __builtin_unreachable.
7473 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7474 instead of flag_asan.
7475 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7476 (pointer_sized_int_node): Define.
41dedebd 7477 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7478
f07f30cf
MS		 74792013-08-30 Mike Stump <mikestump@comcast.net>
7480
7481 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7482 with RE patterns.
7483
8a41354f
JH		 74842013-08-29 Jan Hubicka <jh@suse.cz>
7485
7486 * cgraph.c (cgraph_function_body_availability): Handle weakref
7487 correctly.
7488 * passes.def: Remove pass_fixup_cfg.
7489 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7490 track when we need to remove functions.
7491 (gate_ipa_inline): Execute inlining always; add comment why.
7492 (pass_data_ipa_inline): Remove TODO_remove_functions.
7493 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7494 do not produce summaries.
7495 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7496 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7497 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7498 constant pool and vtable.
7499
dc8d7a0f
TB		 75002013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7501
41dedebd
UB		 7502 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7503 New arm_neon.h's internal macros to specify 64-bit constants.
7504 Avoid using stdint.h's macros.
dc8d7a0f 7505
ac7eacd2
JR		 75062013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7507
7508 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7509 before applying REG_P.
7510
8b29fd4e
JJ		 75112013-08-30 Jakub Jelinek <jakub@redhat.com>
7512
7513 PR tree-optimization/58277
7514 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7515 after seeing too many stmts with vdef in between dombb and current
7516 bb, invalidate everything.
7517
26d75703
RB		 75182013-08-30 Richard Biener <rguenther@suse.de>
7519
7520 * fold-const.c (fold_single_bit_test): Fix overflow test.
7521
f7e088e7
EB		 75222013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7523
7524 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7525 and which can live in a register, always retrieve the value on entry.
7526 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7527 passed by invisible reference specially.
7528 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7529 (vt_add_function_parameter): Correctly deal with a parameter passed by
7530 invisible reference.
7531
75322013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 7533
7534 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7535
6e8dad05
RB		 75362013-08-30 Richard Biener <rguenther@suse.de>
7537
7538 PR tree-optimization/58228
7539 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7540 allow invariant loads in nested loop vectorization.
7541
062ef2c8
RB		 75422013-08-30 Richard Biener <rguenther@suse.de>
7543
7544 PR tree-optimization/58223
7545 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7546 (has_anti_or_output_dependence): ... this and adjust to also
7547 look for output dependences.
7548 (mark_nodes_having_upstream_mem_writes): Adjust.
7549 (rdg_flag_uses): Likewise.
7550
7a764c60
RB		 75512013-08-30 Richard Biener <rguenther@suse.de>
7552
7553 PR tree-optimization/58010
7554 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7555 assert that we have a loop-closed PHI.
7556
82e9d642
JH		 75572013-08-29 Jan Hubicka <jh@suse.cz>
7558
7559 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7560 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7561 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7562 (lto_free_function_in_decl_state_for_node): New function.
7563
9cc1fb4b
XDL		 75642013-08-29 Xinliang David Li <davidxl@google.com>
7565
41dedebd 7566 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 7567 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7568 Emit alignment peeling message with default -fopt-info.
7569 (vect_loop_versioning): Emit loop version info message.
41dedebd 7570 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 7571 (execute_vect_slp): Ditto.
7572
440a5082
EB		 75732013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7574
7575 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7576 of the clone from the DECL_NAME of the original function.
7577
3fa3690d
OE		 75782013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7579
7580 * passes.c (register_pass): Add overload.
7581 * tree-pass.h (register_pass): Forward declare it. Add comment.
7582
0170f33c
JH		 75832013-08-29 Jan Hubicka <jh@suse.cz>
7584
41dedebd
UB		 7585 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7586 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7587 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7588 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7589 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7590 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 7591 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7592 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7593
103ff0d6
TJ		 75942013-08-29 Teresa Johnson <tejohnson@google.com>
7595
7596 * dumpfile.c (dump_loc): Output column number.
7597 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7598 * doc/invoke.texi: Document optall -fopt-info flag.
7599 * profile.c (read_profile_edge_counts): Use new dump framework.
7600 (compute_branch_probabilities): Ditto.
7601 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7602 when pass not in any opt group.
7603 * pass_manager.h (pass_manager::get_pass_profile): New method.
7604 * value-prof.c (check_counter): Use new dump framework.
7605 (check_ic_target): Ditto.
7606 * coverage.c (get_coverage_counts): Ditto.
7607 (coverage_init): Setup new dump framework.
7608
301bbc16
RB		 76092013-08-29 Richard Biener <rguenther@suse.de>
7610
7611 PR tree-optimization/58246
7612 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7613 handle the dominance check inside a basic-block.
7614
1b275000
RB		 76152013-08-29 Richard Biener <rguenther@suse.de>
7616
7617 PR middle-end/57287
7618 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7619 of default defs that appear in abnormal PHI nodes.
7620
6b1184ba
RB		 76212013-08-29 Richard Biener <rguenther@suse.de>
7622
7623 PR tree-optimization/57685
7624 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7625 single-use operands to avoid exponential complexity.
7626
1b08b734
DC		 76272013-08-28 Dehao Chen <dehao@google.com>
7628
7629 * ipa-inline.c (edge_badness): Fix integer underflow.
7630
48a3fa69
UB		 76312013-08-28 Uros Bizjak <ubizjak@gmail.com>
7632
7633 * gtm-builtins.def (_ITM_free): Declare leaf.
7634
215f73e6
JJ		 76352013-08-28 Jakub Jelinek <jakub@redhat.com>
7636
7637 PR target/58067
7638 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7639 (*tls_local_dynamic_base_64_largepic): Likewise.
7640 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7641 Remove predicate from call operand.
7642 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7643 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7644
34554d1a
JL		 76452013-08-28 Jeff Law <law@redhat.com>
7646
7647 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7648 checks for the number of predecessors and successors allowed.
7649 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7650 which require copying a joiner block if there is a request which
7651 is a subpath that requires no joiner block copying.
7652
c01c111b
JH		 76532013-08-28 Jan Hubicka <jh@suse.cz>
7654
7655 * lto-streamer-out.c (DFS_write_tree_body): Drop
7656 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7657 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7658 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7659 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7660 Do not read DECL_ERROR_ISSUED.
7661 (unpack_ts_decl_with_vis_value_fields): Do not read
7662 DECL_DEFER_OUTPUT.
48a3fa69
UB		 7663 (lto_input_ts_binfo_tree_pointers): Do not read
7664 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 7665 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7666 write DECL_ERROR_ISSUED..
7667 (pack_ts_decl_with_vis_value_fields): Do not write
7668 DECL_DEFER_OUTPUT.
7669 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 7670 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 7671 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7672 * tree.h (tree_decl_common): Update comment.
7673 (DECL_ERROR_ISSUED): Remove.
7674
4b128ece
JJ		 76752013-08-28 Jakub Jelinek <jakub@redhat.com>
7676
7677 PR middle-end/58257
7678 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7679
12211b99 76802013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 7681
7682 * builtins.def (free): Declare leaf.
7683
8dce4dbc
DM		 76842013-08-27 David Malcolm <dmalcolm@redhat.com>
7685
7686 * gdbhooks.py: New.
7687 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7688 * configure: Regenerate.
7689
b8f6e610
MJ		 76902013-08-27 Martin Jambor <mjambor@suse.cz>
7691
7692 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7693 (ipa_ancestor_jf_data): Likewise.
7694 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7695 (ipa_get_jf_pass_through_type_preserved): New function.
7696 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7697 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 7698 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 7699 (ipa_get_jf_ancestor_result): Likewise.
7700 (propagate_vals_accross_pass_through): Use
7701 ipa_get_jf_pass_through_result to do all the value mappings.
7702 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7703 type_preserved flag.
7704 (ipa_set_jf_cst_copy): New function.
7705 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7706 (ipa_set_jf_arith_pass_through): Likewise.
7707 (ipa_set_ancestor_jf): Likewise.
7708 (compute_complex_assign_jump_func): Set type_preserved instead of
7709 punting.
7710 (ipa_compute_jump_functions_for_edge): Likewise.
7711 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7712 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 7713 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 7714 (ipa_write_jump_function): Stream the type_preserved flags.
7715 (ipa_read_jump_function): Likewise.
7716
74bf76ed
JJ		 77172013-08-27 Jakub Jelinek <jakub@redhat.com>
7718 Aldy Hernandez <aldyh@redhat.com>
7719
7720 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7721 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7722 * function.h (struct function): Add has_force_vect_loops and
7723 has_simduid_loops.
12211b99 7724 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 7725 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7726 handle it.
7727 * gimple.def: Update CLAUSES comments.
7728 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7729 (gimple_build_omp_for): Add argument to prototype.
7730 (gimple_omp_for_kind): New.
7731 (gimple_omp_for_set_kind): New.
7732 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7733 GOVD_DATA_SHARE_CLASS.
7734 (enum omp_region_type): Add ORT_SIMD.
7735 (gimple_add_tmp_var): Handle ORT_SIMD.
7736 (gimplify_var_or_parm_decl): Same.
7737 (is_gimple_stmt): Same.
7738 (omp_firstprivatize_variable): Same.
7739 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7740 (omp_notice_variable): Handle ORT_SIMD.
12211b99 7741 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 7742 (omp_check_private): Handle ORT_SIMD.
7743 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7744 OMP_CLAUSE_SAFELEN.
7745 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7746 Handle OMP_CLAUSE_LASTPRIVATE.
7747 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7748 OMP_CLAUSE_SAFELEN.
7749 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7750 (gimplify_expr): Handle OMP_SIMD.
7751 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7752 (expand_GOMP_SIMD_VF): New.
7753 (expand_GOMP_SIMD_LAST_LANE): New.
7754 * internal-fn.def (GOMP_SIMD_LANE): New.
7755 (GOMP_SIMD_VF): New.
7756 (GOMP_SIMD_LAST_LANE): New.
7757 * omp-low.c: Include target.h.
7758 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7759 OMP_CLAUSE_SAFELEN.
7760 (check_omp_nesting_restrictions): Same.
7761 (omp_max_vf): New.
7762 (lower_rec_simd_input_clauses): New.
7763 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7764 OMP_CLAUSE_LINEAR.
7765 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7766 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7767 (expand_omp_build_assign): New.
7768 (expand_omp_for_init_counts): New.
7769 (expand_omp_for_init_vars): New.
7770 (extract_omp_for_update_vars): New.
7771 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7772 and rewrite accordingly.
7773 (expand_omp_simd): New.
7774 (expand_omp_for): Use expand_omp_simd.
7775 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7776 (lower_omp_for): Do not lower the body.
7777 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7778 in their own loops.
7779 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 7780 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7781 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 7782 (gate_tree_if_conversion): Similarly.
7783 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7784 gimple_build_omp_for.
7785 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7786 * tree-parloops (create_parallel_loop): Pass kind argument to
7787 gimple_build_omp_for.
7788 * tree-pretty-print.c (dump_omp_clause): Add cases for
7789 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7790 OMP_CLAUSE__SIMDUID_.
7791 (dump_generic_node): Handle OMP_SIMD.
7792 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7793 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7794 unroll OMP_SIMD loops here.
12211b99 7795 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 7796 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 7797 loop->safelen.
74bf76ed
JJ		 7798 (vect_analyze_data_refs): Handle simd loops.
7799 * tree-vect-loop.c (vectorizable_live_operation): Handle
7800 IFN_GOMP_SIMD*.
12211b99 7801 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 7802 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7803 (vectorizable_load): Same.
12211b99 7804 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 7805 (struct simduid_to_vf): New.
7806 (simduid_to_vf::hash): New.
7807 (simduid_to-vf::equal): New.
7808 (struct simd_array_to_simduid): New.
7809 (simd_array_to_simduid::hash): New.
7810 (simd_array_to_simduid::equal): New.
7811 (adjust_simduid_builtins): New.
7812 (struct note_simd_array_uses_struct): New.
7813 (note_simd_array_uses_cb): New.
7814 (note_simd_array_uses): New.
7815 (vectorize_loops): Handle simd hints and adjust simd builtins
7816 accordingly.
7817 * tree-vectorizer.h (struct _stmt_vec_info): Add
7818 simd_lane_access_p field.
7819 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7820 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7821 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7822 (omp_clause_code_name): Same.
7823 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7824 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7825 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 7826 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7827 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 7828 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7829 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7830 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7831 (OMP_CLAUSE_LINEAR_STEP): New.
7832 (OMP_CLAUSE_SAFELEN_EXPR): New.
7833 (OMP_CLAUSE__SIMDUID__DECL): New.
7834 (find_omp_clause): New prototype.
7835
d0cf4e84
L		 78362013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7837
7838 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7839 Haswell processor detection.
7840
992592ec
CW		 78412013-08-27 Christian Widmer <shadow@umbrox.de>
7842
7843 PR target/57927
7844 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7845 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7846 AVX2 capable processors.
7847
9116eb22
TB		 78482013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7849
7850 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7851 of vget_low_* with implementations in terms of other intrinsics.
7852
1c5abb53
MG		 78532013-08-27 Marc Glisse <marc.glisse@inria.fr>
7854
7855 PR middle-end/57219
7856 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7857 values to -1, 0 and 1.
7858
2e100703
VP		 78592013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7860
7861 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7862 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7863 (<optab><mode>3_insn): Remove.
7864 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7865 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7866 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7867 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7868 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7869 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7870 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7871 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7872 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7873 (ror<mode>3_insn): Likewise.
7874 * config/aarch64/predicates.md (aarch64_simd_register): New.
7875
4ded8276
RB		 78762013-08-27 Richard Biener <rguenther@suse.de>
7877
7878 PR tree-optimization/57521
7879 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7880 one edge is non-critical.
7881 (find_phi_replacement_condition): Make sure to use a non-critical
7882 edge. Cleanup and remove old bug workarounds.
7883 (bb_postdominates_preds): Remove.
7884 (if_convertible_loop_p_1): Do not compute post-dominators.
7885 (combine_blocks): Do not free post-dominators.
7886 (main_tree_if_conversion): Likewise.
7887 (pass_data_if_conversion): Add TODO_verify_ssa.
7888
5aa11061
DD		 78892013-08-27 DJ Delorie <dj@redhat.com>
7890
7891 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7892
fd91cfe3
YZ		 78932013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7894
7895 * function.c (assign_parm_find_data_types): Set passed_mode and
7896 nominal_mode to the TYPE_MODE of nominal_type for the built
7897 pointer type in case of the struct-pass-by-reference.
7898
907555ce
JR		 78992013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7900
7901 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7902 (UINT16_TYPE): Change default to "unsigned int".
7903
67518c93
JR		 7904 * config/avr/avr.opt (mfract-convert-truncate): New option.
7905 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7906 is set, round negative fractional integers according to n1169
7907 when converting to integer types.
7908
d6d989fc
JH		 79092013-08-26 Jan Hubicka <jh@suse.cz>
7910
7911 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7912 methods can not be called indirectly when their address is not taken.
7913
2aa3da06
JH		 79142013-08-26 Jan Hubicka <jh@suse.cz>
7915
12211b99
UB		 7916 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7917 ctor_for_folding.
2aa3da06 7918
0987ffe7
JH		 79192013-08-26 Jan Hubicka <jh@suse.cz>
7920
7921 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7922 can be unshared.
7923
befe8647
JR		 79242013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7925
7926 * reload.c (find_valid_class): Allow classes that do not include
7927 FIRST_PSEUDO_REGISTER - 1.
7928
a21e735e
JH		 79292013-08-26 Jan Hubicka <jh@suse.cz>
7930
7931 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7932 fix edge count/frequency when speculation failed; fix type check
7933 for the direct call.
7934
e067bd43
JH		 79352013-08-26 Jan Hubicka <jh@suse.cz>
7936
7937 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7938
e8aec975
JH		 79392013-08-26 Jan Hubicka <jh@suse.cz>
7940
7941 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7942 to be changed by edge redirection.
7943
d0b66480
JH		 79442013-08-26 Jan Hubicka <jh@suse.cz>
7945
12211b99
UB		 7946 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7947 formating; add sanity check.
d0b66480
JH		 7948 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7949 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7950 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7951 (resolve_noninline_speculation): Update callee keys, too.
7952
0f9aaac7
JH		 79532013-08-26 Jan Hubicka <jh@suse.cz>
7954
7955 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7956 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7957
4b37444e
JR		 79582013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7959
7960 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7961 into proper place.
7962
5339fc0c
UB		 79632013-08-26 Uros Bizjak <ubizjak@gmail.com>
7964
7965 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7966 (x86_64_elf_select_section): Ditto.
7967 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7968 arguments.
7969 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 7970 (output_set_got): Ditto.
7971 (ix86_unary_operator_ok): Ditto.
7972 (ix86_expand_builtin): Ditto.
7973
d5c3d3ef
JH		 79742013-08-23 Jan Hubicka <jh@suse.cz>
7975
5339fc0c 7976 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 7977
5b1e7435
JH		 79782013-08-23 Jan Hubicka <jh@suse.cz>
7979
7980 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7981 (tree_decl_with_vis): Add FINAL field.
7982
ead69dac
JL		 79832013-08-23 Jeff Law <law@redhat.com>
7984
7985 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7986 the debugging dump when the expression is fully redundant.
7987
025311c4
GDR		 79882013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7989
7990 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7991 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7992 * pretty-print.c (pp_formatted_text_data): Likewise.
7993 (pp_write_text_to_stream): Likewise.
7994 (pp_write_text_as_dot_label_to_stream): Likewise.
7995 (pp_append_r): Likewise.
7996 (pp_format): Likewise.
7997 (pp_flush): Likewise.
7998 (pp_clear_output_area): Likewise.
7999 (pp_append_text): Likewise.
8000 (pp_formatted_text): Likewise.
8001 (pp_remaining_character_count_for_line): Likewise.
8002 (pp_newline): Likewise.
8003 (pp_character): Likewise.
8004 (output_buffer::~output_buffer): Define.
8005 (pretty_printer::~pretty_printer): Destruct output buffer.
8006 * pretty-print.h (output_buffer::~output_buffer): Declare.
8007 (pretty_printer::~pretty_printer): Declare virtual.
8008
0cadbfaa
MG		 80092013-08-24 Marc Glisse <marc.glisse@inria.fr>
8010
8011 PR other/57324
8012 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8013 HOST_WIDE_INT_M1U): New macros.
8014 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8015 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8016 unsigned -1 for lshift.
8017 * cse.c (cse_insn): Likewise.
8018 * double-int.c (rshift_double, lshift_double): Likewise.
8019 * builtins.c (fold_builtin_bitop): Likewise.
8020 * combine.c (force_to_mode): Likewise.
8021 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8022 * simplify-rtx.c (simplify_const_unary_operation,
8023 simplify_const_binary_operation): Likewise.
8024 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8025 check_va_list_escapes): Likewise.
8026 * rtlanal.c (nonzero_bits1): Likewise.
8027 * expmed.c (expand_smod_pow2): Likewise.
8028 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8029
009e5353
JH		 80302013-08-23 Jan Hubicka <jh@suse.cz>
8031
8032 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8033 as having address taken.
8034
c4be6568
JH		 80352013-08-23 Jan Hubicka <jh@suse.cz>
8036
64cbf23d
JH		 8037 * ipa-utils.h (method_class_type): Declare.
8038 * ipa-devirt.c (method_class_type): Export.
8039
c4be6568
JH		 8040 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8041 do not walk base classes of anonymous types.
8042
97aba8e9
KK		 80432013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8044
8045 PR rtl-optimization/58220
8046 PR regression/58221
8047 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8048 handle SEQUENCE insns properly.
8049
0fc80001
GDR		 80502013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8051
8052 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8053 definition.
8054 (pp_newline_and_indent): Likewise.
8055 (pp_separate_with): Likewise.
8056 * pretty-print.c (pp_newline_and_flush): Define.
8057 (pp_newline_and_indent): Likewise.
8058 (pp_separate_with): Likewise.
8059
520a5868
JJ		 80602013-08-23 Jakub Jelinek <jakub@redhat.com>
8061
8062 PR target/58218
8063 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8064 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8065
e500c62a
KY		 80662013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8067
f43245d1
UB		 8068 * config/i386/predicates.md (ext_sse_reg_operand): New.
8069 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY		 8070 predicate to determine if EVEX is needed.
8071 (*movsi_internal): Ditto.
8072 (*movdf_internal): Ditto.
8073 (*movsf_internal): Ditto.
f43245d1 8074 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 8075
6b00d7dd
JJ		 80762013-08-23 Jakub Jelinek <jakub@redhat.com>
8077
8078 PR tree-optimization/58209
8079 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8080 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8081 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8082 emit POINTER_PLUS_EXPR.
8083 (create_tailcall_accumulator): For pointer result type accumulate in
8084 sizetype type.
8085
4f2a9f90
PC		 80862013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8087
8088 * configure.ac: Add backslashes missing from the last change.
8089 * configure: Regenerate.
8090
7bf4274e
JH		 80912013-08-22 Jan Hubicka <jh@susue.cz>
8092
4f2a9f90
PC		 8093 * ipa.c (function_and_variable_visibility): First remember function
8094 was global and then make it local.
7bf4274e 8095
1bea243d
JB		 80962013-08-22 Julian Brown <julian@codesourcery.com>
8097
8098 * configure.ac: Add aarch64 to list of arches which use "nop" in
8099 debug_line test.
8100 * configure: Regenerate.
8101
bcba7ecc
AK		 81022013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8103
8104 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8105 gnu_libc_has_function.
8106 * config/s390/tpf.h: Likewise.
8107
4c495b0d
JH		 81082013-08-22 Jan Hubicka <jh@susue.cz>
8109
8110 * timevar.c (validate_phases): Add cast.
8111
5e302bca
JH		 81122013-08-22 Jan Hubicka <jh@susue.cz>
8113
8114 * timevar.c (validate_phases): Use size_t for memory.
8115 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8116
da6ca2b5
GDR		 81172013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8118
8119 * pretty-print.h (output_buffer::output_buffer): Declare.
8120 (pretty_printer::pretty_printer): Likewise.
8121 (pp_construct): Remove.
8122 * pretty-print.c (output_buffer::output_buffer): Define.
8123 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8124 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8125 (print_gimple_expr): Likewise.
8126 (print_gimple_seq): Likewise.
8127 (gimple_dump_bb): Likewise.
8128 * sched-vis.c (dump_value_slim): Likewise.
8129 (dump_insn_slim): Likewise.
8130 (dump_rtl_slim): Likewise.
8131 (str_pattern_slim): Likewise.
8132 * tree-mudflap.c (mf_varname_tree): Likewise.
8133 * graph.c (print_graph_cfg): Likewise.
8134 (start_graph_dump): Likewise.
8135 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8136 placement-new.
8137 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8138 pretty printer initialization.
8139 * coretypes.h (diagnostic_context): Remove superflous type alias
8140 declaration.
8141 (pretty_printer): Likewise. Declare directly as a class.
8142 (pretty_print_info): Remove declaration as class.
8143 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8144 and pp_clear_output_area.
8145 (asan_add_global): Likewise.
8146
0e1474e5
JH		 81472013-08-22 Jan Hubicka <jh@suse.cz>
8148
8149 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8150 * ipa-utils.h (update_type_inheritance_graph): Declare.
8151 (possible_polymorphic_call_target_p): Declare.
8152 (possible_polymorphic_call_target_p): New.
8153 * ipa-devirt.c: Update toplevel comments.
8154 (cached_polymorphic_call_targets): Move up.
8155 (odr_type_d): Move ID down.
8156 (polymorphic_type_binfo_p): Update comment.
8157 (odr_hasher::remove): Likewise;
8158 (get_odr_type): Set anonymous_namespace.
8159 (dump_odr_type): Dump it.
8160 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8161 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8162 (record_binfo): Add comment.
5339fc0c
UB		 8163 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8164 built.
0e1474e5
JH		 8165 (devirt_node_removal_hook): Do not iCE when cache is freed.
8166 (possible_polymorphic_call_target_p): New predicate.
8167 (update_type_inheritance_graph): New function.
8168
3f97cb0b
AI		 81692013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8170 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8171 Sergey Lega <sergey.s.lega@intel.com>
8172 Anna Tikhonova <anna.tikhonova@intel.com>
8173 Ilya Tocar <ilya.tocar@intel.com>
8174 Andrey Turetskiy <andrey.turetskiy@intel.com>
8175 Ilya Verbin <ilya.verbin@intel.com>
8176 Kirill Yukhin <kirill.yukhin@intel.com>
8177 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8178
8179 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8180 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8181 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8182 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8183 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8184 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8185 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8186 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8187 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8188 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8189 OPT_mavx512pf, OPT_mavx512er cases.
8190 * config/i386/constraints.md (v): New constraint.
8191 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8192 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8193 (bit_AVX512CD): New.
8194 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8195 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8196 * config/i386/i386-c.c (ix86_target_macros_internal):
8197 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8198 __AVX512PF__.
8199 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8200 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8201 * config/i386/i386.c (regclass_map, dbx_register_map)
8202 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8203 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8204 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8205 -mavx512pf options.
8206 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8207 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8208 -mavx512pf options. Fix formatting.
8209 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8210 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8211 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8212 -mavx512cd, -mavx512pf options.
8213 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8214 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8215 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8216 ALL_SSE_REGS.
8217 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8218 (ix86_set_reg_reg_cost): Ditto.
8219 (x86_order_regs_for_local_alloc): Ditto.
8220 (MAX_VECT_LEN): Extend to 64-byte.
8221 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8222 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8223 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8224 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8225 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8226 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8227 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8228 (SSE_REG_MODE_P): Support new modes.
8229 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8230 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8231 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8232 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8233 (REG_CLASS_CONTENTS): Add new registers.
8234 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8235 (EXT_REX_SSE_REGNO_P): New.
8236 (HI_REGISTER_NAMES): Add new registers.
8237 * config/i386/i386.md: Define constants for new registers.
8238 (mode): Add new 512-bit modes.
8239 (prefix): Support evex prefix.
8240 (isa): Support avx512f, noavx512f, fma_avx512f.
8241 (ssemodesuffix): Add new 512-bit modes.
8242 (movxi): New.
8243 (*movxi_internal_avx512f): Ditto.
8244 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8245 Support MODE_XI.
8246 (*movsi_internal): Likewise.
8247 (*movdf_internal): Likewise.
8248 (*movsf_internal): Likewise.
8249 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8250 (<code><mode>3): Likewise.
5339fc0c
UB		 8251 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8252 New.
3f97cb0b
AI		 8253 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8254 with the new constraint "v".
8255 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8256 modes.
8257 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8258 with the new constraint "v".
8259 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8260 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8261 (*<plusminus_insn><mode>3): Likewise.
8262 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8263 (*mul<mode>3): Likewise.
8264 (<sse>_vmmul<mode>3): Likewise.
8265 (<sse>_div<mode>3): Likewise.
8266 (<sse>_vmdiv<mode>3): Likewise.
8267 (<sse>_sqrt<mode>2): Likewise.
8268 (<sse>_vmsqrt<mode>2): Likewise.
8269 (*<code><mode>3_finite): Likewise.
8270 (*<code><mode>3) <smaxmin>: Likewise.
8271 (<sse>_vm<code><mode>3): Likewise.
8272 (*<code><mode>3) <any_logic>: Likewise.
8273 (*fma_fmadd_<mode>): Likewise.
8274 (*fma_fmsub_<mode>): Likewise.
8275 (*fma_fnmadd_<mode>): Likewise.
8276 (*fma_fnmsub_<mode>): Likewise.
8277 (*fma_fmaddsub_<mode>): Likewise.
8278 (*fma_fmsubadd_<mode>): Likewise.
8279 (*fmai_fmadd_<mode>): Likewise.
8280 (*fmai_fmsub_<mode>): Likewise.
8281 (*fmai_fnmadd_<mode>): Likewise.
8282 (*fmai_fnmsub_<mode>): Likewise.
8283 (sse_cvtsi2ss): Likewise.
8284 (sse_cvtsi2ssq): Likewise.
8285 (sse_cvtss2si): Likewise.
8286 (sse_cvtss2si_2): Likewise.
8287 (sse_cvtss2siq): Likewise.
8288 (sse_cvtss2siq_2): Likewise.
8289 (sse_cvttss2si): Likewise.
8290 (sse_cvtss2siq_2): Likewise.
8291 (float<sseintvecmodelower><mode>2): Likewise.
8292 (sse2_cvtsd2si_2): Likewise.
8293 (sse2_cvtsd2siq_2): Likewise.
8294 (*<plusminus_insn><mode>3): Likewise.
8295 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8296 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8297 (ashr<mode>3): Likewise.
8298 (<shift_insn><mode>3): Likewise.
8299 (avx2_<code><mode>3): Likewise.
8300 (*avx2_<code><mode>3): Likewise.
8301 (*andnot<mode>3): Likewise.
8302 (*<code><mode>3) <any_logic>: Likewise.
8303 (abs<mode>2): Likewise.
8304 (avx2_permvar<mode>): Likewise.
8305 (avx2_perm<mode>_1): Likewise.
8306 (*avx_vpermilp<mode>): Likewise.
8307 (avx_vpermilvar<mode>3): Likewise.
8308 (avx2_ashrv<mode>): Likewise.
8309 (avx2_<shift_insn>v<mode>): Likewise.
8310 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8311 -mavx512cd.
8312 * doc/rtl.texi: Document XImode.
8313
dd5e8423
JL		 83142013-08-21 Jeff Law <law@redhat.com>
8315
b9ebee5d
JL		 8316 * tree-flow.h (register_jump_thread): Pass vector of edges
8317 instead of each important edge.
8318 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8319 thread path into a vector and pass that to register_jump_thread.
8320 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8321 passed in edge vector to the current 3-edge form.
8322
dd5e8423
JL		 8323 Revert:
8324 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8325
9bb6628e 8326 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 8327 delete_unmarked_insns removed anything.
8328
445dc8df
JR		 83292013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8330
8331 * reload.h (struct reg_equivs): Rename to ..
8332 (struct reg_equivs_s): .. this.
8333
e0df53dd
ML		 83342013-08-20 Martin Liska <marxin.liska@gmail.com>
8335
8336 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8337
e7d1d3eb
RO		 83382013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8339
5339fc0c 8340 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8341
5562e26e
JL		 83422013-08-21 Jeff Law <law@redhat.com>
8343
8344 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8345 simplify assignments too. If the RHS collapses to a singleton
8346 range, then return the value for the range.
8347
c7ecdec6
KY		 83482013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8349
7a5911d3
UB		 8350 * config/i386/sse.md (V16): Rename to...
8351 (VMOVE): this.
8352 (mov<mode>): Update iterator name.
8353 (*mov<mode>_internal): Ditto.
8354 (push<mode>1): Ditto.
8355 (movmisalign<mode>): Ditto.
c7ecdec6 8356
bfa3b50a
JH		 83572013-08-20 Jan Hubicka <jh@suse.cz>
8358
8359 PR bootstrap/58186
8360 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8361 entry for direct edges.
8362 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8363
6cd4d135
DM		 83642013-08-20 David Malcolm <dmalcolm@redhat.com>
8365
8366 Revert my last two changes, r201865 and r201864:
8367
8368 Revert r201865:
8369 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8370
8371 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8372 instances can own GC refs.
8373
8374 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8375 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8376 (gcc::context::gt_pch_nx): Likewise.
8377 (gcc::context::gt_pch_nx): Likewise.
8378 * ggc.h (gt_ggc_mx <T>): New.
8379 (gt_pch_nx_with_op <T>): New.
8380 (gt_pch_nx <T>): New.
8381 * passes.c (opt_pass::gt_ggc_mx): New.
8382 (opt_pass::gt_pch_nx): New.
8383 (opt_pass::gt_pch_nx_with_op): New.
8384 (pass_manager::gt_ggc_mx): New.
8385 (pass_manager::gt_pch_nx): New.
8386 (pass_manager::gt_pch_nx_with_op): New.
8387 (pass_manager::operator new): Use
8388 ggc_internal_cleared_alloc_stat rather than xcalloc.
8389 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8390 (pass_manager::gt_ggc_mx): New.
8391 (pass_manager::gt_pch_nx): New.
8392 (pass_manager::gt_pch_nx_with_op): New.
8393 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8394 (opt_pass::operator new): New.
8395 (opt_pass::gt_ggc_mx): New.
8396 (opt_pass::gt_pch_nx): New.
8397 (opt_pass::gt_pch_nx_with_op): New.
8398
8399 Revert r201864:
8400 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8401
8402 * Makefile.in (GTFILES): Add context.h.
8403 * context.c (gcc::context::operator new): New.
8404 (gcc::context::gt_ggc_mx): New.
8405 (gcc::context::gt_pch_nx): New.
8406 (gcc::context::gt_pch_nx): New.
8407 * context.h (gcc::context): Add GTY((user)) marking.
8408 (gcc::context::operator new): New.
8409 (gcc::context::gt_ggc_mx): New.
8410 (gcc::context::gt_pch_nx): New.
8411 (gcc::context::gt_pch_nx): New.
8412 (g): Add GTY marking.
8413 (gt_ggc_mx (gcc::context *)): New.
8414 (gt_pch_nx (gcc::context *)): New.
8415 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8416 void *cookie)): New.
8417 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8418
a046136a
AM		 84192013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8420
8421 * dce.c (fini_dce): Call df_analyze again just in case
8422 delete_unmarked_insns removed anything.
8423
67598720
TJ		 84242013-08-20 Teresa Johnson <tejohnson@google.com>
8425
8426 PR rtl-optimizations/57451
8427 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8428 from crossing split section boundaries.
8429
f8693faf
MGD		 84302013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8431
8432 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8433 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8434 with MULTLIB_DEFAULTS.
8435
7d0b9a9c
NC		 84362013-08-20 Nick Clifton <nickc@redhat.com>
8437
8438 * target.def (narrow_volatile_bitfield): Note that the default
8439 value is false, not !TARGET_STRICT_ALIGN.
8440 * doc/tm.texi: Regenerate.
8441
8edb8dc8
PC		 84422013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8443
8444 Fix LIB_SPEC for systems without libpthread.
8445
8446 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8447 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8448 for Android.
8449 * config/i386/linux-common.h: Likewise.
8450 * config/mips/linux-common.h: Likewise.
8451
12211b99 84522013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 8453
8454 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8455 checks.
8456
e42c64cb
DM		 84572013-08-20 David Malcolm <dmalcolm@redhat.com>
8458
8459 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8460 instances can own GC refs.
8461
8462 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8463 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8464 (gcc::context::gt_pch_nx): Likewise.
8465 (gcc::context::gt_pch_nx): Likewise.
8466 * ggc.h (gt_ggc_mx <T>): New.
8467 (gt_pch_nx_with_op <T>): New.
8468 (gt_pch_nx <T>): New.
8469 * passes.c (opt_pass::gt_ggc_mx): New.
8470 (opt_pass::gt_pch_nx): New.
8471 (opt_pass::gt_pch_nx_with_op): New.
8472 (pass_manager::gt_ggc_mx): New.
8473 (pass_manager::gt_pch_nx): New.
8474 (pass_manager::gt_pch_nx_with_op): New.
8475 (pass_manager::operator new): Use
8476 ggc_internal_cleared_alloc_stat rather than xcalloc.
8477 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8478 (pass_manager::gt_ggc_mx): New.
8479 (pass_manager::gt_pch_nx): New.
8480 (pass_manager::gt_pch_nx_with_op): New.
8481 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8482 (opt_pass::operator new): New.
8483 (opt_pass::gt_ggc_mx): New.
8484 (opt_pass::gt_pch_nx): New.
8485 (opt_pass::gt_pch_nx_with_op): New.
8486
910c02a0
DM		 84872013-08-20 David Malcolm <dmalcolm@redhat.com>
8488
8489 * Makefile.in (GTFILES): Add context.h.
8490 * context.c (gcc::context::operator new): New.
8491 (gcc::context::gt_ggc_mx): New.
8492 (gcc::context::gt_pch_nx): New.
8493 (gcc::context::gt_pch_nx): New.
8494 * context.h (gcc::context): Add GTY((user)) marking.
8495 (gcc::context::operator new): New.
8496 (gcc::context::gt_ggc_mx): New.
8497 (gcc::context::gt_pch_nx): New.
8498 (gcc::context::gt_pch_nx): New.
8499 (g): Add GTY marking.
8500 (gt_ggc_mx (gcc::context *)): New.
8501 (gt_pch_nx (gcc::context *)): New.
8502 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8503 void *cookie)): New.
8504 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8505
af4b4236
AM		 85062013-08-20 Alan Modra <amodra@gmail.com>
8507
8508 PR target/57865
8509 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8510 (rs6000_emit_epilogue): Likewise.
8511
12211b99 85122013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 8513
8514 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8515
e2323f5b
PB		 85162013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8517 Jakub Jelinek <jakub@redhat.com>
8518
8519 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8520 (BUILT_IN_FABSD64): Likewise.
8521 (BUILT_IN_FABSD128): Likewise.
8522 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8523 (fold_builtin_1): Likewise.
5339fc0c
UB		 8524 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8525 destination and source operands.
e2323f5b
PB		 8526 (*abstd2_fpr): Likewise.
8527 (*nabstd2_fpr): Likewise.
8528
cfbf3ee8
RS		 85292013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8530
8531 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8532 JUMP_P and INSN_P.
8533
d3136aeb
AH		 85342013-08-19 Aldy Hernandez <aldyh@redhat.com>
8535
8536 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8537 incomplete.
8538
d33d9e47
AI		 85392013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8540
8541 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8542 * builtins.c (default_libc_has_function): New.
8543 (gnu_libc_has_function): Ditto.
8544 (no_c99_libc_has_function): Ditto.
8545 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8546 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8547 (fold_builtin_sincos): Likewise.
8548 (fold_builtin_cexp): Likewise.
8549 * builtins.def (DEF_C94_BUILTIN): Likewise.
8550 (DEF_C99_BUILTIN): Likewise.
8551 (DEF_C99_C90RES_BUILTIN): Likewise.
8552 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8553 definitions to using this define.
8554 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 8555 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 8556 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8557 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8558 * config/darwin.h: Ditto.
8559 * config/elfos.h: Ditto.
8560 * config/freebsd.h: Ditto.
8561 * config/i386/cygming.h: Ditto.
8562 * config/i386/djgpp.h: Ditto.
8563 * config/i386/i386-interix.h: Ditto.
8564 * config/microblaze/microblaze.h: Ditto.
8565 * config/mmix/mmix.h: Ditto.
8566 * config/gnu-user.h: Ditto.
8567 * config/ia64/hpux.h: Ditto.
8568 * config/pa/pa-hpux.h: Ditto.
8569 * config/pdp11/pdp11.h: Ditto.
8570 * config/picochip/picochip.h: Ditto.
8571 * config/linux.h: Ditto.
8572 * config/netbsd.h: Ditto.
8573 * config/openbsd.h: Ditto.
8574 * config/rs6000/aix43.h: Ditto.
8575 * config/rs6000/aix51.h: Ditto.
8576 * config/rs6000/aix52.h: Ditto.
8577 * config/rs6000/aix53.h: Ditto.
8578 * config/rs6000/aix61.h: Ditto.
8579 * config/rs6000/darwin.h: Ditto.
8580 * config/rs6000/linux.h: Ditto.
8581 * config/rs6000/linux64.h: Ditto.
8582 * config/s390/tpf.h: Ditto.
8583 * config/sol2-10.h: Ditto.
8584 * config/sol2.h: Ditto.
8585 * config/vms/vms.h: Ditto.
8586 * config/vxworks.h: Ditto.
8587 * config/linux-android.c (linux_android_libc_has_function):
8588 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8589 * config/linux-protos.h (linux_android_libc_has_function):
8590 New declaration.
8591 * config/i386/i386.c (ix86_libc_has_function): New.
8592 * config/i386/i386-protos.h
8593 (ix86_libc_has_function): New declaration.
8594 * config/i386/i386.md
8595 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8596 ("isinf<mode>2): Likewise.
8597 * convert.c (convert_to_integer): Using new target hook
8598 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8599 TARGET_C99_FUNCTIONS.
8600 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8601 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8602 * coretypes.h (function_class): New enum for different
8603 classes of functions.
8604 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8605 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8606 (TARGET_HAS_SINCOS): Likewise.
8607 (TARGET_LIBC_HAS_FUNCTION): New.
8608 * doc/tm.texi: Regenerated.
8609 * targhooks.h (default_libc_has_function): New declaration.
8610 (no_c99_libc_has_function): Ditto.
8611 (gnu_libc_has_function): Ditto.
8612 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8613 and TARGET_HAS_SINCOS.
8614
eefe9a99
JH		 86152013-08-18 Jan Hubicka <jh@suse.cz>
8616
8617 * Makeifle-in (ipa-devirt.o): New.
8618 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8619 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8620 (analyze_functions): Look into possible targets of polymorphic call.
8621 * dumpfile.c (dump_files): Add type-inheritance dump.
8622 * dumpfile.h (TDI_inheritance): New.
8623 * ipa-devirt.c: New file.
8624 * ipa-utils.h (odr_type_d): Forward declare.
8625 (odr_type): New type.
8626 (build_type_inheritance_graph): Declare.
8627 (possible_polymorphic_call_targets): Declare and introduce inline
8628 variant when only edge is pased.
8629 (dump_possible_polymorphic_call_targets): Likewise.
8630 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8631 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8632 (types_same_for_odr): ... here.
8633 * tree.h (type_in_anonymous_namespace_p): Declare.
8634
29b89442
JJ		 86352013-08-18 Jakub Jelinek <jakub@redhat.com>
8636
8637 PR tree-optimization/58006
8638 * tree-parloops.c (take_address_of): Don't ICE if get_name
8639 returns NULL.
8640 (eliminate_local_variables_stmt): Remove clobber stmts.
8641
4f219961
EB		 86422013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8643
8644 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8645 error message is issued for an alias to undefined symbol.
8646
1d5755ef
JH		 86472013-08-18 Jan Hubicka <jh@suse.cz>
8648
8649 * cgraph.c (cgraph_create_indirect_edge): Discover
8650 polymorphic calls and record basic info into indirect_info.
8651 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8652 devirtualization, ignore objc function calls.
8653 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8654 call with no parm index info.
8655 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8656 * tree.c (virtual_method_call_p): New function.
8657 * tree.h (virtual_method_call_p): Declare.
8658
5945bebf
JH		 86592013-08-16 Jan Hubicka <jh@suse.cz>
8660
8661 PR middle-end/58179
8662 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8663
47fa5d23
DE		 86642013-08-16 David Edelsohn <dje.gcc@gmail.com>
8665
8666 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8667 attribute.
8668
c74f54a0
DM		 86692013-08-16 David Malcolm <dmalcolm@redhat.com>
8670
8671 * gengtype.c (type_for_name): Add special-case support for
8672 locating types within the "gcc::" namespace.
8673 (open_base_files): Emit a "using namespace gcc" directive.
8674
f758f299
MM		 86752013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8676
8677 PR target/58160
8678 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8679 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8680
8681 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8682 array instead of each individual operand as a separate argument.
8683 (emit_fusion_gpr_load): Likewise.
8684 (expand_fusion_gpr_load): Add new function declaration.
8685
8686 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8687 signature to have the operands passed as an array, instead of as
8688 separate arguments. Allow ZERO_EXTEND to be in the memory
8689 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8690 depend on the register live/dead flags when peepholes are run.
8691 (expand_fusion_gpr_load): New function to be called from the
8692 peephole2 pass, to change the register that addis sets to be the
8693 target register.
8694 (emit_fusion_gpr_load): Change the calling signature to have the
8695 operands passed as an array, instead of as separate arguments.
8696 Allow ZERO_EXTEND to be in the memory address, and also
8697 SIGN_EXTEND if -mpower8-fusion-sign.
8698
8699 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8700 unspec enumeration.
8701 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8702 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 8703 depend on the register live/dead state when the peephole pass is done.
f758f299 8704
158f4e4f
DM		 87052013-08-16 David Malcolm <dmalcolm@redhat.com>
8706
8707 * gengtype.c (create_user_defined_type): Ensure that the kind
8708 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8709 declaration is seen before the GTY((user)) marking.
8710
22f8cea5
BE		 87112013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8712
8713 PR target/58105
8714 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8715
c49bdb2e
JH		 87162013-08-16 Jan Hubicka <jh@suse.cz>
8717
8718 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8719 arugment expected_type.
8720 (gimple_fold_call): Use it.
8721 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8722 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 8723 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 8724 (try_make_edge_direct_virtual_call): Likewise.
8725 * tree.c (obj_type_ref_class): New.
8726 * tree.h (obj_type_ref_class): Use it.
8727
4042dca9
GDR		 87282013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8729
8730 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8731 (rtl_slim_pp): Likewise.
8732 (init_rtl_slim_pretty_print): Likewise.
8733 (dump_value_slim): Don't call it. Use local pretty printer.
8734 (dump_insn_slim): Likewise.
8735 (dump_rtl_slim): Likewise.
8736 (str_pattern_slim): Likewise.
8737 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8738 Simplify.
8739
7a460594
JJ		 87402013-08-16 Jakub Jelinek <jakub@redhat.com>
8741
639dc669
JJ		 8742 PR tree-optimization/58164
8743 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8744 walk gimple_goto_dest of GIMPLE_GOTO.
8745
7a460594
JJ		 8746 PR tree-optimization/58165
8747 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8748 bi_call must be the last stmt in a bb, don't split_block, instead
8749 use fallthru edge from it and give up if there is none.
8750 Release conds vector when returning early.
8751
e3c02cc7
XDL		 87522013-08-14 Xinliang David Li <davidxl@google.com>
8753
8754 * config/i386/i386.c (ix86_option_override_internal):
8755 Remove unused variable and field.
8756
6469da2c
BS		 87572013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8758
8759 PR target/57949
22f8cea5 8760 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 8761 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8762 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8763 and Linux, correct BLKmode alignment when 128-bit alignment is
8764 required and compatibility flag is not set.
22f8cea5
BE		 8765 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8766 for zero-size arguments when compatibility flag is not set.
6469da2c 8767
4ca890e2
JJ		 87682013-08-14 Jakub Jelinek <jakub@redhat.com>
8769
8770 PR tree-optimization/58145
8771 * tree-sra.c (build_ref_for_offset): If prev_base has
8772 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8773
daec1759 87742013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 8775
daec1759 8776 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 8777 Fix uninitialized variable error.
daec1759 8778
3ad20bd4
XDL		 87792013-08-14 Xinliang David Li <davidxl@google.com>
8780
8781 * config/i386/i386.opt: Define two new options.
8782 * config/i386/x86-tune.def: Add arch selector field in macros.
8783 * config/i386/i386.h: Adjust macro definition.
8784 * config/i386/i386.c (ix86_option_override_internal):
8785 Refactor the code.
8786 (parse_mtune_ctrl_str): New function.
8787 (set_ix86_tune_features): New function.
8788 (ix86_function_specific_restore): Call the new helper function.
8789
e839e2a9
AB		 87902013-08-14 Andrey Belevantsev <abel@ispras.ru>
8791
8792 PR rtl-optimization/57662
8793 * sel-sched.c (code_motion_process_successors): When the current insn
8794 is removed after the recursive traversal, break from the loop.
8795 Add comments and debug printouts.
8796
27ed665f
JJ		 87972013-08-14 Jakub Jelinek <jakub@redhat.com>
8798 Alexandre Oliva <aoliva@redhat.com>
8799
8800 PR target/58067
8801 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8802 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8803 there also UNSPEC_PLTOFF.
8804
61c1a609
MP		 88052013-08-14 Marek Polacek <polacek@redhat.com>
8806
8807 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8808 NUM_CONDITIONS bit positions.
8809
0e901c67
CC		 88102013-08-13 Cary Coutant <ccoutant@google.com>
8811
8812 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8813 (attr_checksum): Hash vector contents instead of pointer.
8814 (attr_checksum_ordered): Likewise.
8815
061eff6d
UB		 88162013-08-13 Uros Bizjak <ubizjak@gmail.com>
8817
8818 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8819 when Pmode != word_mode. Add length_address attribute.
8820 (sse3_monitor_<mode>): Merge from sse3_monitor and
8821 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8822 Pmode != word_mode. Update insn length attribute.
8823 * config/i386/i386.c (ix86_option_override_internal): Update
8824 ix86_gen_monitor selection for merged sse3_monitor insn.
8825
8ec9249b
JB		 88262013-08-13 Julian Brown <julian@codesourcery.com>
8827
8828 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8829 perform invalid legitimization on greater-than-word-size modes for
8830 TARGET_E500_DOUBLE.
8831
761a8eb7
VM		 88322013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8833
8834 * ira.c (setup_class_translate_array): Use aclass instead of cl
8835 for classes not fully covered by allocno classes.
8836
99c2bd54
JJ		 88372013-08-13 Jakub Jelinek <jakub@redhat.com>
8838
4029a5e0
JJ		 8839 PR tree-optimization/57661
8840 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8841 * tree-inline.c (tree_function_versioning): Initialize it.
8842 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8843 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8844 that is not being copied.
8845
99c2bd54
JJ		 8846 PR sanitizer/56417
8847 * asan.c (instrument_strlen_call): Fix typo in comment.
8848 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8849
b5f8f063
SE		 88502013-08-13 Steve Ellcey <sellcey@mips.com>
8851
8852 * config/mips/mips.md (prefetch): Use lw instead of ld on
8853 loongson in 32bit mode.
8854
4de80584
NC		 88552013-08-13 Nick Clifton <nickc@redhat.com>
8856
8857 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8858
09ce3660
JH		 88592013-08-13 Jan Hubicka <jh@suse.cz>
8860
8861 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8862 introduced edge; fix typo in sanity check.
8863 (cgraph_resolve_speculation): Export; improve diagnostic.
8864 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8865 speculation at type mismatch.
8866 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8867 (cgraph_resolve_speculation): Declare.
8868 (symtab_can_be_discarded): New function.
8869 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8870 * ipa-inline-transform.c (speculation_removed): New global var.
8871 (clone_inlined_nodes): See if speculation can be removed.
8872 (inline_call): If speculations was removed, we growths may not match.
8873 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8874 (speculation_useful_p): New function.
8875 (resolve_noninline_speculation): New function.
8876 (inline_small_functions): Resolve useless speculations.
8877 * ipa-inline.h (speculation_useful_p): Declare
8878 * ipa.c (can_replace_by_local_alias): Simplify.
8879 (ipa_profile): Produce speculative calls in non-lto, too;
8880 add simple cost model; produce local aliases.
8881
f971dc24
DM		 88822013-08-13 David Malcolm <dmalcolm@redhat.com>
8883
8884 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 8885 PASS_MANAGER_H.
f971dc24 8886
2e507b9a
PC		 88872013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8888
8889 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8890
c97fd4b8
UB		 88912013-08-12 Uros Bizjak <ubizjak@gmail.com>
8892
8893 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8894 iterator instead of X87MODEF.
8895
88962013-08-12 Perez Read <netfirewall@gmail.com>
8897
8898 PR target/58132
8899 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8900 operand 0 for intel asm alternative.
8901 (*movabs<mode>_2): Ditto for operand 1.
8902
f4f4204c
JG		 89032013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8904
8905 * config/aarch64/arm_none.h
8906 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8907
133b1a8e
NC		 89082013-08-12 Nick Clifton <nickc@redhat.com>
8909
8910 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8911
15dd8b3a
YR		 89122013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8913
8914 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8915 expand for QI/HImode operand to produce more effictive code for
8916 unsigned char(short) --> float(double) conversion.
8917
94109a6a
AM		 89182013-08-12 Alexander Monakov <amonakov@ispras.ru>
8919
8920 * doc/invoke.texi: Mention that -ftls-model does not force the final
8921 model.
8922
4e115102
MP		 89232013-08-12 Marek Polacek <polacek@redhat.com>
8924 Marc Glisse <marc.glisse@inria.fr>
8925
8926 PR tree-optimization/57980
8927 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8928 when creating -1 constant.
8929
748d71f3
JH		 89302013-08-10 Jan Hubicka <jh@suse.cz>
8931
1f97e193 8932 Workaround binutils PR14342.
748d71f3
JH		 8933 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8934 (gimple_init_edge_profiler): Likewise.
8935 (gimple_gen_ic_func_profiler): Likewise.
8936
5979aa54
JH		 89372013-08-09 Jan Hubicka <jh@suse.cz>
8938
8939 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8940
ad83025e
XDL		 89412013-08-09 Xinliang David Li <davidxl@google.com>
8942
8943 * config/i386/stringop.def: New file.
8944 * config/i386/stringop.opt: New file.
8945 * config/i386/i386-opts.h: Include stringopt.def.
8946 * config/i386/i386.opt: Include stringopt.opt.
8947 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 8948 Override default size based stringop inline strategies with options.
ad83025e
XDL		 8949 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8950 New function.
8951
e59df5fd
JH		 89522013-08-09 Jan Hubicka <jh@suse.cz>
8953
c97fd4b8 8954 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 8955
634ab819
JH		 89562013-08-09 Jan Hubicka <jh@suse.cz>
8957
8958 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8959 CGRAPH_FREQ_MAX.
8960 (dump_cgraph_node): Dump profile-id.
8961 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8962 and common_target_probability.
8963 * lto-cgraph.c (lto_output_edge): Stream common targets.
8964 (lto_output_node): Stream profile ids.
8965 (input_node): Stream profile ids.
8966 (input_edge): Stream common targets.
8967 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8968 * ipa.c: Include value-prof.h
8969 (ipa_profile_generate_summary): Turn indirect call statement histograms
8970 into common targets.
8971 (ipa_profile): Turn common targets into speculative edges.
8972
2fa3d31b
JH		 89732013-08-09 Jan Hubicka <jh@suse.cz>
8974
8975 * cgraph.h (cgraph_node): Add profile_id.
8976 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 8977 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 8978 (del_node_map): Update.
8979 (find_func_by_funcdef_no): Replace by ...
8980 (find_func_by_profile_id): ... this one.
8981 (gimple_ic_transform): Do not remove useful histograms when
8982 speculation is not done; dump info when indirect call removal
8983 can happen at LTO.
8984 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8985 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8986 (__gcov_indirect_call_profiler_v2): .. this one.
8987 * profile.h (init_node_map): Update.
8988 * coverage.c (coverage_compute_profile_id): New function.
8989 * coverage.h (coverage_compute_profile_id): Declare.
8990 * tree-profile.c (init_ic_make_global_vars): Make
8991 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8992 (gimple_init_edge_profiler): Update prototype of
8993 __gcov_indirect_call_profiler.
8994 (gimple_gen_ic_func_profiler): Simplify.
8995 (tree_profiling): Use init_node_map
8996
042ae7d2
JH		 89972013-08-09 Jan Hubicka <jh@suse.cz>
8998
c97fd4b8
UB		 8999 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9000 non-speculative refs.
042ae7d2
JH		 9001 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9002 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9003 (cgraph_set_call_stmt): Likewise.
9004 (cgraph_create_edge_1): Fix release checking compilatoin;
9005 clear lto_stmt_uid.
9006 (cgraph_free_edge): Free indirect info.
9007 (cgraph_turn_edge_to_speculative): New function.
9008 (cgraph_speculative_call_info): New function.
9009 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 9010 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 9011 (dump_cgraph_node): Dump speculation.
9012 (verify_edge_count_and_frequency): Accept speculative edges.
9013 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9014 (verify_cgraph_node): Handle speculation.
9015 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9016 (cgraph_set_call_stmt): Update prototype.
9017 (cgraph_make_edge_direct): Update prototype.
9018 (cgraph_speculative_call_info): Declare.
9019 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9020 to change; update call of ipa_find_references.
9021 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9022 lto_stmt_uid and speculative flags.
9023 (ipa_dump_references): Dump speculation.
9024 (ipa_clone_references): Clone speculative flag.
9025 (ipa_clone_referring): Likewise.
9026 (ipa_clone_ref): New function.
9027 (ipa_find_reference): Look into lto_stmt_uids
9028 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9029 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9030 (ipa_find_reference): Update declaration.
9031 (ipa_clone_ref): Declare.
9032 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9033 stream speculative flag.
9034 (lto_output_ref): Stream statements uids and speculation.
9035 (input_ref): Likewise.
9036 (input_edge): Stream speuclation.
9037 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9038 (cgraph_set_call_stmt_including_clones): Handle speculation.
9039 * ipa-inline.c (heap_edge_removal_hook): New function.
9040 (inline_small_functions): Register it.
9041 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9042 also initialize refs.
9043 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9044 edge to change.
9045 (try_make_edge_direct_simple_call): Likewise.
9046 (try_make_edge_direct_simple_call): Likewise.
9047 (update_indirect_edges_after_inlining): Likewise.
9048 (remove_described_reference): Look proper lto_stmt_uid.
9049 (propagate_controlled_uses): Likewise.
9050 (propagate_controlled_uses): Liekwise.
9051 * tree-inline.c (copy_bb): Copy speculative edges.
9052 (redirect_all_calls): New function.
c97fd4b8
UB		 9053 (copy_cfg_body): Do redirection after loop info is updated.
9054 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 9055
aa1e10cc
JH		 90562013-08-09 Jan Hubicka <jh@suse.cz>
9057
9058 * lto-streamer-out.c (output_function): Renumber PHIs.
9059 * lto-streamer-in.c (input_function): Likewise.
9060
66adb8eb
JG		 90612013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9062
9063 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9064 (get_lane_unsigned): Likewise.
9065 (dup_lane_scalar): Likewise.
9066 (get_lane): enable for VALL.
9067 * config/aarch64/aarch64-simd.md
9068 (aarch64_dup_lane_scalar<mode>): Remove.
9069 (aarch64_get_lane_signed<mode>): Likewise.
9070 (aarch64_get_lane_unsigned<mode>): Likewise.
9071 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9072 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9073 (aarch64_get_lane<mode>): Enable for all vector modes.
9074 (aarch64_get_lanedi): Remove misleading constraints.
9075 * config/aarch64/arm_neon.h
9076 (__aarch64_vget_lane_any): Define.
9077 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9078 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9079 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9080 * config/aarch64/iterators.md (VDQQH): New.
9081 (VDQQHS): Likewise.
9082 (vwcore): Likewise.
9083
1f65ae7a
EB		 90842013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9085
9086 * configure.ac: Add GAS check for LEON instructions on SPARC.
9087 * configure: Regenerate.
9088 * config.in: Likewise.
9089 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9090 sparc*-*-* block.
9091 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9092 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9093 for LEON or LEON3.
9094 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9095 (AS_LEON_FLAG): New macro.
9096 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9097 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9098 Deal with LEON and LEON3 for the memory model.
23a6cb78 9099 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 9100 (atomic_compare_and_swap<mode>_1): Likewise.
9101 (*atomic_compare_and_swap<mode>_1): Likewise.
9102
ff522f7f
ZC		 91032013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9104
9105 * config/arm/neon.md (vcond): Fix floating-point vector
9106 comparisons against 0.
9107
cb1cca12
VM		 91082013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9109
9110 * lra-constraints.c (emit_spill_move): Remove assert.
9111 (process_alt_operands): Add more debugging
9112 output. Increase reject for spilling into memory. Decrease
9113 reject for reloading scratch.
9114 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9115
a3719e31
SE		 91162013-08-08 Steve Ellcey <sellcey@mips.com>
9117
9118 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9119 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9120 micromips incompatible. Add nan2008.
9121 (MULTILIB_DIRNAMES): Add nan2008.
9122 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9123 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9124 and micromips incompatible. Add nan2008.
9125 (MULTILIB_DIRNAMES): Add nan2008.
9126 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9127
f40423e2
RS		 91282013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9129
9130 PR rtl-optimization/58079
9131 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9132 simplify_comparison has widened a comparison with an integer.
9133
869b9125
KT		 91342013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9135
9136 * config/arm/neon.md (movmisalign<mode>): Disable when we
9137 don't allow unaligned accesses.
9138 (*movmisalign<mode>_neon_store): Likewise.
9139 (*movmisalign<mode>_neon_load): Likewise.
9140 (*movmisalign<mode>_neon_store): Likewise.
9141 (*movmisalign<mode>_neon_load): Likewise.
9142
71cafea9
JH		 91432013-08-08 Jan Hubicka <jh@suse.cz>
9144
9145 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9146 (make_pass_rebuild_cgraph_edges): Also clear references.
9147 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9148 * ipa-inline-transform.c (inline_transform): Remove all references
9149 after inlining.
c97fd4b8
UB		 9150 * cgraphunit.c (expand_function): Remove all references after
9151 expansion.
71cafea9
JH		 9152 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9153 (ipa_find_reference): Rewrite to iterator.
9154 (remove_stmt_references): Likewise.
9155 (ipa_clear_stmts_in_references): New function.
9156 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 9157 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9158 clear references.
71cafea9
JH		 9159 * ipa-split.c (split_function): Remove references in split function.
9160
27d2e612
RE		 91612013-08-08 Richard Earnshaw <rearnsha@arm.com>
9162
9163 PR target/57431
1f65ae7a 9164 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 9165 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9166
5efc661c
RE		 91672013-08-08 Richard Earnshaw <rearnsha@arm.com>
9168
9169 PR target/56979
1f65ae7a 9170 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 9171 suggested mode for the assignment isn't compatible with the
9172 registers required.
9173
f276d31d
BE		 91742013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9175
9176 PR target/58065
9177 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9178
4b8bc035
XDL		 91792013-08-07 Xinliang David Li <davidxl@google.com>
9180
9181 * config/i386/i386.opt: New option -mtune-ctrl=.
9182 * config/i386/x86-tune.def: New file.
9183 * config/i386/i386.h: include x86-tune.def.
9184 * config/i386/i386.c (ix86_option_override_internal):
9185 Parsing -mtune-ctrl= option and set tune features.
9186
12211b99 91872013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 9188
9189 PR other/12081
03b0ee0a 9190 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 9191 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9192 to gen_2arg_fn_t.
9193
5cf6635b
EB		 91942013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9195
9196 * rtl.h (update_alignments): Declare.
9197 * final.c (grow_label_align): New function extracted from...
9198 (shorten_branches): ...here. Call it.
9199 (update_alignments): New function.
9200 * reorg.c (sibling_labels): New variable.
9201 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9202 the new label along with it onto the sibling_labels vector.
9203 (fill_simple_delay_slots): Adjust call to get_label_before.
9204 (fill_slots_from_thread): Likewise.
9205 (relax_delay_slots): Likewise.
9206 (make_return_insns): Likewise.
9207 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9208
3c8ca1ab
EB		 92092013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9210
9211 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9212 document its semantics.
9213 (diagnostic_report_diagnostic): Adjust accordingly.
9214
5ee5b32c
DM		 92152013-08-07 David Malcolm <dmalcolm@redhat.com>
9216
03b0ee0a 9217 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 9218 (sparc_option_override): ...and port to new C++ pass API.
9219 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9220
abd566fa
PB		 92212013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9222
9223 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9224
2077db1b
CT		 92252013-08-06 Caroline Tice <cmtice@google.com>
9226
9227 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9228 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9229 * tree-pass.h: Add pass_vtable_verify.
9230 * varasm.c (assemble_variable): Add code to properly set the comdat
9231 section and name for the .vtable_map_vars section.
9232 (assemble_vtyv_preinit_initializer): New function.
9233 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9234 LINK_ONCE flag.
9235 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9236 * vtable-verify.c: New file.
9237 * vtable-verify.h: New file.
9238 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9239 initialiation levels.
9240 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9241 * passes.def: Insert pass_vtable_verify.
9242 * aclocal.m4: Reorder includes.
9243 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9244 -fvtv-counts options.
9245 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9246 as appropriate, if -fvtable-verify=... is used.
9247 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9248 -fvtable-verify=... is used.
9249 * Makefile.in (OBJS): Add vtable-verify.o to list.
9250 (vtable-verify.o): Add new build rule.
9251 (GTFILES): Add vtable-verify.c to list.
9252 * common.opt (fvtable-verify=): New flag.
9253 (vtv_priority): Values for fvtable-verify= flag.
9254 (fvtv-counts): New flag.
9255 (fvtv-debug): New flag.
9256 * tree.h (save_vtable_map_decl): New extern function decl.
9257
03085d1c
DM		 92582013-08-07 David Malcolm <dmalcolm@redhat.com>
9259
9260 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9261 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9262 (pass_data_rl78_devirt): ...new pass_data instance and...
9263 (make_pass_rl78_devirt): ...new function.
9264 (rl78_asm_file_start): Port pass registration to new C++ API.
9265
05555c4a
DM		 92662013-08-07 David Malcolm <dmalcolm@redhat.com>
9267
9268 * coretypes.h (rtl_opt_pass): Add.
9269 (gcc::context): Add.
9270 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9271 (epiphany_init): Port to new C++ pass API.
9272 (epiphany_optimize_mode_switching): Likewise.
9273 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9274 (pass_manager::get_pass_mode_switching): New.
9275 (pass_manager::get_pass_peephole2): New.
9276 * mode-switching.c (pass_mode_switching): Add clone method.
9277 * recog.c (pass_peephole2): Add clone method.
9278 (pass_split_all_insns): Add clone method.
9279
c5a12e26
DM		 92802013-08-06 David Malcolm <dmalcolm@redhat.com>
9281
03b0ee0a
UB		 9282 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9283 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 9284
a2e2a668
JH		 92852013-08-06 Jan Hubicka <jh@suse.cz>
9286
9287 * cgraph.c (cgraph_get_body): New function based on lto.c
9288 implementation.
9289 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 9290 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9291 LTO paths.
a2e2a668
JH		 9292 * cgraphunit.c (expand_function): Get body prior expanding.
9293 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 9294 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9295 really need.
a2e2a668
JH		 9296 * passes.c (do_per_function_toporder): Get body.
9297 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9298 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9299
15682f24
MJ		 93002013-08-06 Martin Jambor <mjambor@suse.cz>
9301
9302 PR fortran/57987
9303 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9304 is not re-finalized. Rename second parameter to no_collect.
9305
78f6dd68
MJ		 93062013-08-06 Martin Jambor <mjambor@suse.cz>
9307
9308 PR middle-end/58041
9309 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9310 MEM_REF has proper alignment information.
9311
12211b99 93122013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 9313
9314 PR other/12081
9315 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9316 class insn_gen_fn.
9317 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9318 rtx (*) (rtx, ...) with insn_gen_fn.
9319 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9320 insn_gen_fn::stored_funcptr. Add initializer braces.
9321
8ac69a6c
DM		 93222013-08-05 David Malcolm <dmalcolm@redhat.com>
9323
9324 Rewrite how instances of passes are cloned to remove assumptions
9325 about their sizes (thus allowing pass subclasses to have
9326 additional data fields, albeit non-GC-managed ones at this point).
9327
9328 * passes.c (make_pass_instance): Now that passes have clone
9329 methods, rewrite this function to eliminate XNEW and memcpy
9330 calls that used hardcoded sizes. Since this function no longer
9331 creates pass instances, rename it to...
03b0ee0a
UB		 9332 (add_pass_instance): ...this. Document the old way that passes were
9333 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 9334 (next_pass_1): Add an initial_pass argument for use by
9335 add_pass_instance.
9336 (position_pass): When adding multiple instances of a pass, use
9337 the pass's clone method, rather than relying on the XNEW/memcpy
9338 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 9339 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9340 the initial instance of the current pass within the pass manager.
8ac69a6c 9341
27a4cd48
DM		 93422013-08-05 David Malcolm <dmalcolm@redhat.com>
9343
9344 This is the automated part of the conversion of passes from C
9345 structs to C++ classes.
9346
9347 Patch autogenerated by refactor_passes.py from
9348 https://github.com/davidmalcolm/gcc-refactoring-scripts
9349 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9350
9351 * asan.c (pass_asan): Convert from a global struct to a subclass of
9352 gimple_opt_pass along with...
9353 (pass_data_asan): ...new pass_data instance and...
9354 (make_pass_asan): ...new function.
9355 (pass_asan_O0): Convert from a global struct to a subclass of
9356 gimple_opt_pass along with...
9357 (pass_data_asan_O0): ...new pass_data instance and...
9358 (make_pass_asan_O0): ...new function.
9359 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9360 subclass of rtl_opt_pass along with...
9361 (pass_data_inc_dec): ...new pass_data instance and...
9362 (make_pass_inc_dec): ...new function.
9363 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9364 a subclass of rtl_opt_pass along with...
9365 (pass_data_reorder_blocks): ...new pass_data instance and...
9366 (make_pass_reorder_blocks): ...new function.
9367 (pass_duplicate_computed_gotos): Convert from a global struct to a
9368 subclass of rtl_opt_pass along with...
9369 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9370 (make_pass_duplicate_computed_gotos): ...new function.
9371 (pass_partition_blocks): Convert from a global struct to a subclass of
9372 rtl_opt_pass along with...
9373 (pass_data_partition_blocks): ...new pass_data instance and...
9374 (make_pass_partition_blocks): ...new function.
9375 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9376 struct to a subclass of rtl_opt_pass along with...
9377 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9378 and...
9379 (make_pass_branch_target_load_optimize1): ...new function.
9380 (pass_branch_target_load_optimize2): Convert from a global struct to a
9381 subclass of rtl_opt_pass along with...
9382 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9383 and...
9384 (make_pass_branch_target_load_optimize2): ...new function.
9385 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9386 of rtl_opt_pass along with...
9387 (pass_data_jump): ...new pass_data instance and...
9388 (make_pass_jump): ...new function.
9389 (pass_jump2): Convert from a global struct to a subclass of
9390 rtl_opt_pass along with...
9391 (pass_data_jump2): ...new pass_data instance and...
9392 (make_pass_jump2): ...new function.
9393 * cfgexpand.c (pass_expand): Convert from a global struct to a
9394 subclass of rtl_opt_pass along with...
9395 (pass_data_expand): ...new pass_data instance and...
9396 (make_pass_expand): ...new function.
9397 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9398 of rtl_opt_pass along with...
9399 (pass_data_free_cfg): ...new pass_data instance and...
9400 (make_pass_free_cfg): ...new function.
9401 (pass_into_cfg_layout_mode): Convert from a global struct to a
9402 subclass of rtl_opt_pass along with...
9403 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9404 (make_pass_into_cfg_layout_mode): ...new function.
9405 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9406 subclass of rtl_opt_pass along with...
9407 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9408 (make_pass_outof_cfg_layout_mode): ...new function.
9409 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9410 struct to a subclass of gimple_opt_pass along with...
9411 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9412 (make_pass_build_cgraph_edges): ...new function.
9413 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9414 subclass of gimple_opt_pass along with...
9415 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9416 (make_pass_rebuild_cgraph_edges): ...new function.
9417 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9418 subclass of gimple_opt_pass along with...
9419 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9420 and...
9421 (make_pass_remove_cgraph_callee_edges): ...new function.
9422 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9423 struct to a subclass of rtl_opt_pass along with...
9424 (pass_data_stack_adjustments): ...new pass_data instance and...
9425 (make_pass_stack_adjustments): ...new function.
9426 * combine.c (pass_combine): Convert from a global struct to a subclass
9427 of rtl_opt_pass along with...
9428 (pass_data_combine): ...new pass_data instance and...
9429 (make_pass_combine): ...new function.
9430 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9431 global struct to a subclass of rtl_opt_pass along with...
9432 (pass_data_compare_elim_after_reload): ...new pass_data instance
9433 and...
9434 (make_pass_compare_elim_after_reload): ...new function.
9435 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9436 of rtl_opt_pass along with...
9437 (pass_data_rtl_cprop): ...new pass_data instance and...
9438 (make_pass_rtl_cprop): ...new function.
9439 * cse.c (pass_cse): Convert from a global struct to a subclass of
9440 rtl_opt_pass along with...
9441 (pass_data_cse): ...new pass_data instance and...
9442 (make_pass_cse): ...new function.
9443 (pass_cse2): Convert from a global struct to a subclass of
9444 rtl_opt_pass along with...
9445 (pass_data_cse2): ...new pass_data instance and...
9446 (make_pass_cse2): ...new function.
9447 (pass_cse_after_global_opts): Convert from a global struct to a
9448 subclass of rtl_opt_pass along with...
9449 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9450 (make_pass_cse_after_global_opts): ...new function.
9451 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9452 of rtl_opt_pass along with...
9453 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9454 (make_pass_ud_rtl_dce): ...new function.
9455 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9456 rtl_opt_pass along with...
9457 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9458 (make_pass_fast_rtl_dce): ...new function.
9459 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9460 a subclass of rtl_opt_pass along with...
9461 (pass_data_df_initialize_opt): ...new pass_data instance and...
9462 (make_pass_df_initialize_opt): ...new function.
9463 (pass_df_initialize_no_opt): Convert from a global struct to a
9464 subclass of rtl_opt_pass along with...
9465 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9466 (make_pass_df_initialize_no_opt): ...new function.
9467 (pass_df_finish): Convert from a global struct to a subclass of
9468 rtl_opt_pass along with...
9469 (pass_data_df_finish): ...new pass_data instance and...
9470 (make_pass_df_finish): ...new function.
9471 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9472 rtl_opt_pass along with...
9473 (pass_data_rtl_dse1): ...new pass_data instance and...
9474 (make_pass_rtl_dse1): ...new function.
9475 (pass_rtl_dse2): Convert from a global struct to a subclass of
9476 rtl_opt_pass along with...
9477 (pass_data_rtl_dse2): ...new pass_data instance and...
9478 (make_pass_rtl_dse2): ...new function.
9479 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9480 subclass of rtl_opt_pass along with...
9481 (pass_data_dwarf2_frame): ...new pass_data instance and...
9482 (make_pass_dwarf2_frame): ...new function.
9483 * except.c (pass_set_nothrow_function_flags): Convert from a global
9484 struct to a subclass of rtl_opt_pass along with...
9485 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9486 and...
9487 (make_pass_set_nothrow_function_flags): ...new function.
9488 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9489 subclass of rtl_opt_pass along with...
9490 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9491 and...
9492 (make_pass_convert_to_eh_region_ranges): ...new function.
9493 * final.c (pass_compute_alignments): Convert from a global struct to a
9494 subclass of rtl_opt_pass along with...
9495 (pass_data_compute_alignments): ...new pass_data instance and...
9496 (make_pass_compute_alignments): ...new function.
9497 (pass_final): Convert from a global struct to a subclass of
9498 rtl_opt_pass along with...
9499 (pass_data_final): ...new pass_data instance and...
9500 (make_pass_final): ...new function.
9501 (pass_shorten_branches): Convert from a global struct to a subclass of
9502 rtl_opt_pass along with...
9503 (pass_data_shorten_branches): ...new pass_data instance and...
9504 (make_pass_shorten_branches): ...new function.
9505 (pass_clean_state): Convert from a global struct to a subclass of
9506 rtl_opt_pass along with...
9507 (pass_data_clean_state): ...new pass_data instance and...
9508 (make_pass_clean_state): ...new function.
9509 * function.c (pass_instantiate_virtual_regs): Convert from a global
9510 struct to a subclass of rtl_opt_pass along with...
9511 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9512 (make_pass_instantiate_virtual_regs): ...new function.
9513 (pass_leaf_regs): Convert from a global struct to a subclass of
9514 rtl_opt_pass along with...
9515 (pass_data_leaf_regs): ...new pass_data instance and...
9516 (make_pass_leaf_regs): ...new function.
9517 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9518 subclass of rtl_opt_pass along with...
9519 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9520 and...
9521 (make_pass_thread_prologue_and_epilogue): ...new function.
9522 (pass_match_asm_constraints): Convert from a global struct to a
9523 subclass of rtl_opt_pass along with...
9524 (pass_data_match_asm_constraints): ...new pass_data instance and...
9525 (make_pass_match_asm_constraints): ...new function.
9526 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9527 subclass of rtl_opt_pass along with...
9528 (pass_data_rtl_fwprop): ...new pass_data instance and...
9529 (make_pass_rtl_fwprop): ...new function.
9530 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9531 rtl_opt_pass along with...
9532 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9533 (make_pass_rtl_fwprop_addr): ...new function.
9534 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9535 rtl_opt_pass along with...
9536 (pass_data_rtl_pre): ...new pass_data instance and...
9537 (make_pass_rtl_pre): ...new function.
9538 (pass_rtl_hoist): Convert from a global struct to a subclass of
9539 rtl_opt_pass along with...
9540 (pass_data_rtl_hoist): ...new pass_data instance and...
9541 (make_pass_rtl_hoist): ...new function.
9542 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9543 subclass of gimple_opt_pass along with...
9544 (pass_data_lower_cf): ...new pass_data instance and...
9545 (make_pass_lower_cf): ...new function.
9546 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9547 from a global struct to a subclass of gimple_opt_pass along with...
9548 (pass_data_strength_reduction): ...new pass_data instance and...
9549 (make_pass_strength_reduction): ...new function.
9550 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9551 of rtl_opt_pass along with...
9552 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9553 (make_pass_rtl_ifcvt): ...new function.
9554 (pass_if_after_combine): Convert from a global struct to a subclass of
9555 rtl_opt_pass along with...
9556 (pass_data_if_after_combine): ...new pass_data instance and...
9557 (make_pass_if_after_combine): ...new function.
9558 (pass_if_after_reload): Convert from a global struct to a subclass of
9559 rtl_opt_pass along with...
9560 (pass_data_if_after_reload): ...new pass_data instance and...
9561 (make_pass_if_after_reload): ...new function.
9562 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9563 a subclass of rtl_opt_pass along with...
9564 (pass_data_initialize_regs): ...new pass_data instance and...
9565 (make_pass_initialize_regs): ...new function.
9566 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9567 of ipa_opt_pass_d along with...
9568 (pass_data_ipa_cp): ...new pass_data instance and...
9569 (make_pass_ipa_cp): ...new function.
9570 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9571 global struct to a subclass of gimple_opt_pass along with...
9572 (pass_data_inline_parameters): ...new pass_data instance and...
9573 (make_pass_inline_parameters): ...new function.
9574 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9575 subclass of gimple_opt_pass along with...
9576 (pass_data_early_inline): ...new pass_data instance and...
9577 (make_pass_early_inline): ...new function.
9578 (pass_ipa_inline): Convert from a global struct to a subclass of
9579 ipa_opt_pass_d along with...
9580 (pass_data_ipa_inline): ...new pass_data instance and...
9581 (make_pass_ipa_inline): ...new function.
9582 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9583 struct to a subclass of gimple_opt_pass along with...
9584 (pass_data_local_pure_const): ...new pass_data instance and...
9585 (make_pass_local_pure_const): ...new function.
9586 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9587 ipa_opt_pass_d along with...
9588 (pass_data_ipa_pure_const): ...new pass_data instance and...
9589 (make_pass_ipa_pure_const): ...new function.
9590 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9591 to a subclass of ipa_opt_pass_d along with...
9592 (pass_data_ipa_reference): ...new pass_data instance and...
9593 (make_pass_ipa_reference): ...new function.
9594 * ipa-split.c (pass_split_functions): Convert from a global struct to
9595 a subclass of gimple_opt_pass along with...
9596 (pass_data_split_functions): ...new pass_data instance and...
9597 (make_pass_split_functions): ...new function.
9598 (pass_feedback_split_functions): Convert from a global struct to a
9599 subclass of gimple_opt_pass along with...
9600 (pass_data_feedback_split_functions): ...new pass_data instance and...
9601 (make_pass_feedback_split_functions): ...new function.
9602 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9603 global struct to a subclass of simple_ipa_opt_pass along with...
9604 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9605 instance and...
9606 (make_pass_ipa_function_and_variable_visibility): ...new function.
9607 (pass_ipa_free_inline_summary): Convert from a global struct to a
9608 subclass of simple_ipa_opt_pass along with...
9609 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9610 (make_pass_ipa_free_inline_summary): ...new function.
9611 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9612 subclass of ipa_opt_pass_d along with...
9613 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9614 and...
9615 (make_pass_ipa_whole_program_visibility): ...new function.
9616 (pass_ipa_profile): Convert from a global struct to a subclass of
9617 ipa_opt_pass_d along with...
9618 (pass_data_ipa_profile): ...new pass_data instance and...
9619 (make_pass_ipa_profile): ...new function.
9620 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9621 ipa_opt_pass_d along with...
9622 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9623 (make_pass_ipa_cdtor_merge): ...new function.
9624 * ira.c (pass_ira): Convert from a global struct to a subclass of
9625 rtl_opt_pass along with...
9626 (pass_data_ira): ...new pass_data instance and...
9627 (make_pass_ira): ...new function.
9628 (pass_reload): Convert from a global struct to a subclass of
9629 rtl_opt_pass along with...
9630 (pass_data_reload): ...new pass_data instance and...
9631 (make_pass_reload): ...new function.
9632 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9633 subclass of rtl_opt_pass along with...
9634 (pass_data_cleanup_barriers): ...new pass_data instance and...
9635 (make_pass_cleanup_barriers): ...new function.
9636 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9637 of rtl_opt_pass along with...
9638 (pass_data_loop2): ...new pass_data instance and...
9639 (make_pass_loop2): ...new function.
9640 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9641 rtl_opt_pass along with...
9642 (pass_data_rtl_loop_init): ...new pass_data instance and...
9643 (make_pass_rtl_loop_init): ...new function.
9644 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9645 rtl_opt_pass along with...
9646 (pass_data_rtl_loop_done): ...new pass_data instance and...
9647 (make_pass_rtl_loop_done): ...new function.
9648 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9649 subclass of rtl_opt_pass along with...
9650 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9651 (make_pass_rtl_move_loop_invariants): ...new function.
9652 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9653 rtl_opt_pass along with...
9654 (pass_data_rtl_unswitch): ...new pass_data instance and...
9655 (make_pass_rtl_unswitch): ...new function.
9656 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9657 subclass of rtl_opt_pass along with...
9658 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9659 and...
9660 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9661 (pass_rtl_doloop): Convert from a global struct to a subclass of
9662 rtl_opt_pass along with...
9663 (pass_data_rtl_doloop): ...new pass_data instance and...
9664 (make_pass_rtl_doloop): ...new function.
9665 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9666 a subclass of rtl_opt_pass along with...
9667 (pass_data_lower_subreg): ...new pass_data instance and...
9668 (make_pass_lower_subreg): ...new function.
9669 (pass_lower_subreg2): Convert from a global struct to a subclass of
9670 rtl_opt_pass along with...
9671 (pass_data_lower_subreg2): ...new pass_data instance and...
9672 (make_pass_lower_subreg2): ...new function.
9673 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9674 struct to a subclass of ipa_opt_pass_d along with...
9675 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9676 (make_pass_ipa_lto_gimple_out): ...new function.
9677 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9678 of ipa_opt_pass_d along with...
9679 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9680 (make_pass_ipa_lto_finish_out): ...new function.
9681 * mode-switching.c (pass_mode_switching): Convert from a global struct
9682 to a subclass of rtl_opt_pass along with...
9683 (pass_data_mode_switching): ...new pass_data instance and...
9684 (make_pass_mode_switching): ...new function.
9685 * modulo-sched.c (pass_sms): Convert from a global struct to a
9686 subclass of rtl_opt_pass along with...
9687 (pass_data_sms): ...new pass_data instance and...
9688 (make_pass_sms): ...new function.
9689 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9690 subclass of gimple_opt_pass along with...
9691 (pass_data_expand_omp): ...new pass_data instance and...
9692 (make_pass_expand_omp): ...new function.
9693 (pass_lower_omp): Convert from a global struct to a subclass of
9694 gimple_opt_pass along with...
9695 (pass_data_lower_omp): ...new pass_data instance and...
9696 (make_pass_lower_omp): ...new function.
9697 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9698 of gimple_opt_pass along with...
9699 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9700 (make_pass_diagnose_omp_blocks): ...new function.
9701 * passes.c (pass_early_local_passes): Convert from a global struct to
9702 a subclass of simple_ipa_opt_pass along with...
9703 (pass_data_early_local_passes): ...new pass_data instance and...
9704 (make_pass_early_local_passes): ...new function.
9705 (pass_all_early_optimizations): Convert from a global struct to a
9706 subclass of gimple_opt_pass along with...
9707 (pass_data_all_early_optimizations): ...new pass_data instance and...
9708 (make_pass_all_early_optimizations): ...new function.
9709 (pass_all_optimizations): Convert from a global struct to a subclass
9710 of gimple_opt_pass along with...
9711 (pass_data_all_optimizations): ...new pass_data instance and...
9712 (make_pass_all_optimizations): ...new function.
9713 (pass_all_optimizations_g): Convert from a global struct to a subclass
9714 of gimple_opt_pass along with...
9715 (pass_data_all_optimizations_g): ...new pass_data instance and...
9716 (make_pass_all_optimizations_g): ...new function.
9717 (pass_rest_of_compilation): Convert from a global struct to a subclass
9718 of rtl_opt_pass along with...
9719 (pass_data_rest_of_compilation): ...new pass_data instance and...
9720 (make_pass_rest_of_compilation): ...new function.
9721 (pass_postreload): Convert from a global struct to a subclass of
9722 rtl_opt_pass along with...
9723 (pass_data_postreload): ...new pass_data instance and...
9724 (make_pass_postreload): ...new function.
9725 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9726 subclass of rtl_opt_pass along with...
9727 (pass_data_gcse2): ...new pass_data instance and...
9728 (make_pass_gcse2): ...new function.
9729 * postreload.c (pass_postreload_cse): Convert from a global struct to
9730 a subclass of rtl_opt_pass along with...
9731 (pass_data_postreload_cse): ...new pass_data instance and...
9732 (make_pass_postreload_cse): ...new function.
9733 * predict.c (pass_profile): Convert from a global struct to a subclass
9734 of gimple_opt_pass along with...
9735 (pass_data_profile): ...new pass_data instance and...
9736 (make_pass_profile): ...new function.
9737 (pass_strip_predict_hints): Convert from a global struct to a subclass
9738 of gimple_opt_pass along with...
9739 (pass_data_strip_predict_hints): ...new pass_data instance and...
9740 (make_pass_strip_predict_hints): ...new function.
9741 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9742 of rtl_opt_pass along with...
9743 (pass_data_peephole2): ...new pass_data instance and...
9744 (make_pass_peephole2): ...new function.
9745 (pass_split_all_insns): Convert from a global struct to a subclass of
9746 rtl_opt_pass along with...
9747 (pass_data_split_all_insns): ...new pass_data instance and...
9748 (make_pass_split_all_insns): ...new function.
9749 (pass_split_after_reload): Convert from a global struct to a subclass
9750 of rtl_opt_pass along with...
9751 (pass_data_split_after_reload): ...new pass_data instance and...
9752 (make_pass_split_after_reload): ...new function.
9753 (pass_split_before_regstack): Convert from a global struct to a
9754 subclass of rtl_opt_pass along with...
9755 (pass_data_split_before_regstack): ...new pass_data instance and...
9756 (make_pass_split_before_regstack): ...new function.
9757 (pass_split_before_sched2): Convert from a global struct to a subclass
9758 of rtl_opt_pass along with...
9759 (pass_data_split_before_sched2): ...new pass_data instance and...
9760 (make_pass_split_before_sched2): ...new function.
9761 (pass_split_for_shorten_branches): Convert from a global struct to a
9762 subclass of rtl_opt_pass along with...
9763 (pass_data_split_for_shorten_branches): ...new pass_data instance
9764 and...
9765 (make_pass_split_for_shorten_branches): ...new function.
9766 * ree.c (pass_ree): Convert from a global struct to a subclass of
9767 rtl_opt_pass along with...
9768 (pass_data_ree): ...new pass_data instance and...
9769 (make_pass_ree): ...new function.
9770 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9771 subclass of rtl_opt_pass along with...
9772 (pass_data_stack_regs): ...new pass_data instance and...
9773 (make_pass_stack_regs): ...new function.
9774 (pass_stack_regs_run): Convert from a global struct to a subclass of
9775 rtl_opt_pass along with...
9776 (pass_data_stack_regs_run): ...new pass_data instance and...
9777 (make_pass_stack_regs_run): ...new function.
9778 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9779 subclass of rtl_opt_pass along with...
9780 (pass_data_cprop_hardreg): ...new pass_data instance and...
9781 (make_pass_cprop_hardreg): ...new function.
9782 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9783 subclass of rtl_opt_pass along with...
9784 (pass_data_reginfo_init): ...new pass_data instance and...
9785 (make_pass_reginfo_init): ...new function.
9786 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9787 of rtl_opt_pass along with...
9788 (pass_data_regmove): ...new pass_data instance and...
9789 (make_pass_regmove): ...new function.
9790 * regrename.c (pass_regrename): Convert from a global struct to a
9791 subclass of rtl_opt_pass along with...
9792 (pass_data_regrename): ...new pass_data instance and...
9793 (make_pass_regrename): ...new function.
9794 * reorg.c (pass_delay_slots): Convert from a global struct to a
9795 subclass of rtl_opt_pass along with...
9796 (pass_data_delay_slots): ...new pass_data instance and...
9797 (make_pass_delay_slots): ...new function.
9798 (pass_machine_reorg): Convert from a global struct to a subclass of
9799 rtl_opt_pass along with...
9800 (pass_data_machine_reorg): ...new pass_data instance and...
9801 (make_pass_machine_reorg): ...new function.
9802 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9803 of rtl_opt_pass along with...
9804 (pass_data_sched): ...new pass_data instance and...
9805 (make_pass_sched): ...new function.
9806 (pass_sched2): Convert from a global struct to a subclass of
9807 rtl_opt_pass along with...
9808 (pass_data_sched2): ...new pass_data instance and...
9809 (make_pass_sched2): ...new function.
9810 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9811 to a subclass of rtl_opt_pass along with...
9812 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9813 (make_pass_stack_ptr_mod): ...new function.
9814 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9815 to a subclass of rtl_opt_pass along with...
9816 (pass_data_rtl_store_motion): ...new pass_data instance and...
9817 (make_pass_rtl_store_motion): ...new function.
9818 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9819 of gimple_opt_pass along with...
9820 (pass_data_tracer): ...new pass_data instance and...
9821 (make_pass_tracer): ...new function.
9822 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9823 to a subclass of gimple_opt_pass along with...
9824 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9825 (make_pass_diagnose_tm_blocks): ...new function.
9826 (pass_lower_tm): Convert from a global struct to a subclass of
9827 gimple_opt_pass along with...
9828 (pass_data_lower_tm): ...new pass_data instance and...
9829 (make_pass_lower_tm): ...new function.
9830 (pass_tm_init): Convert from a global struct to a subclass of
9831 gimple_opt_pass along with...
9832 (pass_data_tm_init): ...new pass_data instance and...
9833 (make_pass_tm_init): ...new function.
9834 (pass_tm_mark): Convert from a global struct to a subclass of
9835 gimple_opt_pass along with...
9836 (pass_data_tm_mark): ...new pass_data instance and...
9837 (make_pass_tm_mark): ...new function.
9838 (pass_tm_edges): Convert from a global struct to a subclass of
9839 gimple_opt_pass along with...
9840 (pass_data_tm_edges): ...new pass_data instance and...
9841 (make_pass_tm_edges): ...new function.
9842 (pass_tm_memopt): Convert from a global struct to a subclass of
9843 gimple_opt_pass along with...
9844 (pass_data_tm_memopt): ...new pass_data instance and...
9845 (make_pass_tm_memopt): ...new function.
9846 (pass_ipa_tm): Convert from a global struct to a subclass of
9847 simple_ipa_opt_pass along with...
9848 (pass_data_ipa_tm): ...new pass_data instance and...
9849 (make_pass_ipa_tm): ...new function.
9850 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9851 subclass of gimple_opt_pass along with...
9852 (pass_data_call_cdce): ...new pass_data instance and...
9853 (make_pass_call_cdce): ...new function.
9854 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9855 subclass of gimple_opt_pass along with...
9856 (pass_data_build_cfg): ...new pass_data instance and...
9857 (make_pass_build_cfg): ...new function.
9858 (pass_split_crit_edges): Convert from a global struct to a subclass of
9859 gimple_opt_pass along with...
9860 (pass_data_split_crit_edges): ...new pass_data instance and...
9861 (make_pass_split_crit_edges): ...new function.
9862 (pass_warn_function_return): Convert from a global struct to a
9863 subclass of gimple_opt_pass along with...
9864 (pass_data_warn_function_return): ...new pass_data instance and...
9865 (make_pass_warn_function_return): ...new function.
9866 (pass_warn_function_noreturn): Convert from a global struct to a
9867 subclass of gimple_opt_pass along with...
9868 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9869 (make_pass_warn_function_noreturn): ...new function.
9870 (pass_warn_unused_result): Convert from a global struct to a subclass
9871 of gimple_opt_pass along with...
9872 (pass_data_warn_unused_result): ...new pass_data instance and...
9873 (make_pass_warn_unused_result): ...new function.
9874 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9875 a subclass of gimple_opt_pass along with...
9876 (pass_data_merge_phi): ...new pass_data instance and...
9877 (make_pass_merge_phi): ...new function.
9878 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9879 a subclass of gimple_opt_pass along with...
9880 (pass_data_lower_complex): ...new pass_data instance and...
9881 (make_pass_lower_complex): ...new function.
9882 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9883 gimple_opt_pass along with...
9884 (pass_data_lower_complex_O0): ...new pass_data instance and...
9885 (make_pass_lower_complex_O0): ...new function.
9886 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9887 subclass of gimple_opt_pass along with...
9888 (pass_data_lower_eh): ...new pass_data instance and...
9889 (make_pass_lower_eh): ...new function.
9890 (pass_refactor_eh): Convert from a global struct to a subclass of
9891 gimple_opt_pass along with...
9892 (pass_data_refactor_eh): ...new pass_data instance and...
9893 (make_pass_refactor_eh): ...new function.
9894 (pass_lower_resx): Convert from a global struct to a subclass of
9895 gimple_opt_pass along with...
9896 (pass_data_lower_resx): ...new pass_data instance and...
9897 (make_pass_lower_resx): ...new function.
9898 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9899 of gimple_opt_pass along with...
9900 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9901 (make_pass_lower_eh_dispatch): ...new function.
9902 (pass_cleanup_eh): Convert from a global struct to a subclass of
9903 gimple_opt_pass along with...
9904 (pass_data_cleanup_eh): ...new pass_data instance and...
9905 (make_pass_cleanup_eh): ...new function.
9906 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9907 to a subclass of simple_ipa_opt_pass along with...
9908 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9909 (make_pass_ipa_lower_emutls): ...new function.
9910 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9911 a subclass of gimple_opt_pass along with...
9912 (pass_data_if_conversion): ...new pass_data instance and...
9913 (make_pass_if_conversion): ...new function.
9914 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9915 subclass of gimple_opt_pass along with...
9916 (pass_data_build_ssa): ...new pass_data instance and...
9917 (make_pass_build_ssa): ...new function.
9918 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9919 global struct to a subclass of gimple_opt_pass along with...
9920 (pass_data_loop_distribution): ...new pass_data instance and...
9921 (make_pass_loop_distribution): ...new function.
9922 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9923 subclass of gimple_opt_pass along with...
9924 (pass_data_mudflap_1): ...new pass_data instance and...
9925 (make_pass_mudflap_1): ...new function.
9926 (pass_mudflap_2): Convert from a global struct to a subclass of
9927 gimple_opt_pass along with...
9928 (pass_data_mudflap_2): ...new pass_data instance and...
9929 (make_pass_mudflap_2): ...new function.
9930 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9931 subclass of gimple_opt_pass along with...
9932 (pass_data_mudflap_1): ...new pass_data instance and...
9933 (make_pass_mudflap_1): ...new function.
9934 (pass_mudflap_2): Convert from a global struct to a subclass of
9935 gimple_opt_pass along with...
9936 (pass_data_mudflap_2): ...new pass_data instance and...
9937 (make_pass_mudflap_2): ...new function.
9938 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9939 gimple_opt_pass along with...
9940 (pass_data_nrv): ...new pass_data instance and...
9941 (make_pass_nrv): ...new function.
9942 (pass_return_slot): Convert from a global struct to a subclass of
9943 gimple_opt_pass along with...
9944 (pass_data_return_slot): ...new pass_data instance and...
9945 (make_pass_return_slot): ...new function.
9946 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9947 to a subclass of gimple_opt_pass along with...
9948 (pass_data_object_sizes): ...new pass_data instance and...
9949 (make_pass_object_sizes): ...new function.
9950 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9951 global struct to a subclass of gimple_opt_pass along with...
9952 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9953 and...
9954 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9955 (pass_fixup_cfg): Convert from a global struct to a subclass of
9956 gimple_opt_pass along with...
9957 (pass_data_fixup_cfg): ...new pass_data instance and...
9958 (make_pass_fixup_cfg): ...new function.
9959 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9960 (make_pass_mudflap_1): ...new function.
9961 (pass_mudflap_2): Replace declaration with that of...
9962 (make_pass_mudflap_2): ...new function.
9963 (pass_asan): Replace declaration with that of...
9964 (make_pass_asan): ...new function.
9965 (pass_asan_O0): Replace declaration with that of...
9966 (make_pass_asan_O0): ...new function.
9967 (pass_tsan): Replace declaration with that of...
9968 (make_pass_tsan): ...new function.
9969 (pass_tsan_O0): Replace declaration with that of...
9970 (make_pass_tsan_O0): ...new function.
9971 (pass_lower_cf): Replace declaration with that of...
9972 (make_pass_lower_cf): ...new function.
9973 (pass_refactor_eh): Replace declaration with that of...
9974 (make_pass_refactor_eh): ...new function.
9975 (pass_lower_eh): Replace declaration with that of...
9976 (make_pass_lower_eh): ...new function.
9977 (pass_lower_eh_dispatch): Replace declaration with that of...
9978 (make_pass_lower_eh_dispatch): ...new function.
9979 (pass_lower_resx): Replace declaration with that of...
9980 (make_pass_lower_resx): ...new function.
9981 (pass_build_cfg): Replace declaration with that of...
9982 (make_pass_build_cfg): ...new function.
9983 (pass_early_tree_profile): Replace declaration with that of...
9984 (make_pass_early_tree_profile): ...new function.
9985 (pass_cleanup_eh): Replace declaration with that of...
9986 (make_pass_cleanup_eh): ...new function.
9987 (pass_sra): Replace declaration with that of...
9988 (make_pass_sra): ...new function.
9989 (pass_sra_early): Replace declaration with that of...
9990 (make_pass_sra_early): ...new function.
9991 (pass_early_ipa_sra): Replace declaration with that of...
9992 (make_pass_early_ipa_sra): ...new function.
9993 (pass_tail_recursion): Replace declaration with that of...
9994 (make_pass_tail_recursion): ...new function.
9995 (pass_tail_calls): Replace declaration with that of...
9996 (make_pass_tail_calls): ...new function.
9997 (pass_tree_loop): Replace declaration with that of...
9998 (make_pass_tree_loop): ...new function.
9999 (pass_tree_loop_init): Replace declaration with that of...
10000 (make_pass_tree_loop_init): ...new function.
10001 (pass_lim): Replace declaration with that of...
10002 (make_pass_lim): ...new function.
10003 (pass_tree_unswitch): Replace declaration with that of...
10004 (make_pass_tree_unswitch): ...new function.
10005 (pass_predcom): Replace declaration with that of...
10006 (make_pass_predcom): ...new function.
10007 (pass_iv_canon): Replace declaration with that of...
10008 (make_pass_iv_canon): ...new function.
10009 (pass_scev_cprop): Replace declaration with that of...
10010 (make_pass_scev_cprop): ...new function.
10011 (pass_empty_loop): Replace declaration with that of...
10012 (make_pass_empty_loop): ...new function.
10013 (pass_record_bounds): Replace declaration with that of...
10014 (make_pass_record_bounds): ...new function.
10015 (pass_graphite): Replace declaration with that of...
10016 (make_pass_graphite): ...new function.
10017 (pass_graphite_transforms): Replace declaration with that of...
10018 (make_pass_graphite_transforms): ...new function.
10019 (pass_if_conversion): Replace declaration with that of...
10020 (make_pass_if_conversion): ...new function.
10021 (pass_loop_distribution): Replace declaration with that of...
10022 (make_pass_loop_distribution): ...new function.
10023 (pass_vectorize): Replace declaration with that of...
10024 (make_pass_vectorize): ...new function.
10025 (pass_slp_vectorize): Replace declaration with that of...
10026 (make_pass_slp_vectorize): ...new function.
10027 (pass_complete_unroll): Replace declaration with that of...
10028 (make_pass_complete_unroll): ...new function.
10029 (pass_complete_unrolli): Replace declaration with that of...
10030 (make_pass_complete_unrolli): ...new function.
10031 (pass_parallelize_loops): Replace declaration with that of...
10032 (make_pass_parallelize_loops): ...new function.
10033 (pass_loop_prefetch): Replace declaration with that of...
10034 (make_pass_loop_prefetch): ...new function.
10035 (pass_iv_optimize): Replace declaration with that of...
10036 (make_pass_iv_optimize): ...new function.
10037 (pass_tree_loop_done): Replace declaration with that of...
10038 (make_pass_tree_loop_done): ...new function.
10039 (pass_ch): Replace declaration with that of...
10040 (make_pass_ch): ...new function.
10041 (pass_ccp): Replace declaration with that of...
10042 (make_pass_ccp): ...new function.
10043 (pass_phi_only_cprop): Replace declaration with that of...
10044 (make_pass_phi_only_cprop): ...new function.
10045 (pass_build_ssa): Replace declaration with that of...
10046 (make_pass_build_ssa): ...new function.
10047 (pass_build_alias): Replace declaration with that of...
10048 (make_pass_build_alias): ...new function.
10049 (pass_build_ealias): Replace declaration with that of...
10050 (make_pass_build_ealias): ...new function.
10051 (pass_dominator): Replace declaration with that of...
10052 (make_pass_dominator): ...new function.
10053 (pass_dce): Replace declaration with that of...
10054 (make_pass_dce): ...new function.
10055 (pass_dce_loop): Replace declaration with that of...
10056 (make_pass_dce_loop): ...new function.
10057 (pass_cd_dce): Replace declaration with that of...
10058 (make_pass_cd_dce): ...new function.
10059 (pass_call_cdce): Replace declaration with that of...
10060 (make_pass_call_cdce): ...new function.
10061 (pass_merge_phi): Replace declaration with that of...
10062 (make_pass_merge_phi): ...new function.
10063 (pass_split_crit_edges): Replace declaration with that of...
10064 (make_pass_split_crit_edges): ...new function.
10065 (pass_pre): Replace declaration with that of...
10066 (make_pass_pre): ...new function.
10067 (pass_profile): Replace declaration with that of...
10068 (make_pass_profile): ...new function.
10069 (pass_strip_predict_hints): Replace declaration with that of...
10070 (make_pass_strip_predict_hints): ...new function.
10071 (pass_lower_complex_O0): Replace declaration with that of...
10072 (make_pass_lower_complex_O0): ...new function.
10073 (pass_lower_complex): Replace declaration with that of...
10074 (make_pass_lower_complex): ...new function.
10075 (pass_lower_vector): Replace declaration with that of...
10076 (make_pass_lower_vector): ...new function.
10077 (pass_lower_vector_ssa): Replace declaration with that of...
10078 (make_pass_lower_vector_ssa): ...new function.
10079 (pass_lower_omp): Replace declaration with that of...
10080 (make_pass_lower_omp): ...new function.
10081 (pass_diagnose_omp_blocks): Replace declaration with that of...
10082 (make_pass_diagnose_omp_blocks): ...new function.
10083 (pass_expand_omp): Replace declaration with that of...
10084 (make_pass_expand_omp): ...new function.
10085 (pass_expand_omp_ssa): Replace declaration with that of...
10086 (make_pass_expand_omp_ssa): ...new function.
10087 (pass_object_sizes): Replace declaration with that of...
10088 (make_pass_object_sizes): ...new function.
10089 (pass_strlen): Replace declaration with that of...
10090 (make_pass_strlen): ...new function.
10091 (pass_fold_builtins): Replace declaration with that of...
10092 (make_pass_fold_builtins): ...new function.
10093 (pass_stdarg): Replace declaration with that of...
10094 (make_pass_stdarg): ...new function.
10095 (pass_early_warn_uninitialized): Replace declaration with that of...
10096 (make_pass_early_warn_uninitialized): ...new function.
10097 (pass_late_warn_uninitialized): Replace declaration with that of...
10098 (make_pass_late_warn_uninitialized): ...new function.
10099 (pass_cse_reciprocals): Replace declaration with that of...
10100 (make_pass_cse_reciprocals): ...new function.
10101 (pass_cse_sincos): Replace declaration with that of...
10102 (make_pass_cse_sincos): ...new function.
10103 (pass_optimize_bswap): Replace declaration with that of...
10104 (make_pass_optimize_bswap): ...new function.
10105 (pass_optimize_widening_mul): Replace declaration with that of...
10106 (make_pass_optimize_widening_mul): ...new function.
10107 (pass_warn_function_return): Replace declaration with that of...
10108 (make_pass_warn_function_return): ...new function.
10109 (pass_warn_function_noreturn): Replace declaration with that of...
10110 (make_pass_warn_function_noreturn): ...new function.
10111 (pass_cselim): Replace declaration with that of...
10112 (make_pass_cselim): ...new function.
10113 (pass_phiopt): Replace declaration with that of...
10114 (make_pass_phiopt): ...new function.
10115 (pass_forwprop): Replace declaration with that of...
10116 (make_pass_forwprop): ...new function.
10117 (pass_phiprop): Replace declaration with that of...
10118 (make_pass_phiprop): ...new function.
10119 (pass_tree_ifcombine): Replace declaration with that of...
10120 (make_pass_tree_ifcombine): ...new function.
10121 (pass_dse): Replace declaration with that of...
10122 (make_pass_dse): ...new function.
10123 (pass_nrv): Replace declaration with that of...
10124 (make_pass_nrv): ...new function.
10125 (pass_rename_ssa_copies): Replace declaration with that of...
10126 (make_pass_rename_ssa_copies): ...new function.
10127 (pass_sink_code): Replace declaration with that of...
10128 (make_pass_sink_code): ...new function.
10129 (pass_fre): Replace declaration with that of...
10130 (make_pass_fre): ...new function.
10131 (pass_check_data_deps): Replace declaration with that of...
10132 (make_pass_check_data_deps): ...new function.
10133 (pass_copy_prop): Replace declaration with that of...
10134 (make_pass_copy_prop): ...new function.
10135 (pass_vrp): Replace declaration with that of...
10136 (make_pass_vrp): ...new function.
10137 (pass_uncprop): Replace declaration with that of...
10138 (make_pass_uncprop): ...new function.
10139 (pass_return_slot): Replace declaration with that of...
10140 (make_pass_return_slot): ...new function.
10141 (pass_reassoc): Replace declaration with that of...
10142 (make_pass_reassoc): ...new function.
10143 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10144 (make_pass_rebuild_cgraph_edges): ...new function.
10145 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10146 (make_pass_remove_cgraph_callee_edges): ...new function.
10147 (pass_build_cgraph_edges): Replace declaration with that of...
10148 (make_pass_build_cgraph_edges): ...new function.
10149 (pass_local_pure_const): Replace declaration with that of...
10150 (make_pass_local_pure_const): ...new function.
10151 (pass_tracer): Replace declaration with that of...
10152 (make_pass_tracer): ...new function.
10153 (pass_warn_unused_result): Replace declaration with that of...
10154 (make_pass_warn_unused_result): ...new function.
10155 (pass_diagnose_tm_blocks): Replace declaration with that of...
10156 (make_pass_diagnose_tm_blocks): ...new function.
10157 (pass_lower_tm): Replace declaration with that of...
10158 (make_pass_lower_tm): ...new function.
10159 (pass_tm_init): Replace declaration with that of...
10160 (make_pass_tm_init): ...new function.
10161 (pass_tm_mark): Replace declaration with that of...
10162 (make_pass_tm_mark): ...new function.
10163 (pass_tm_memopt): Replace declaration with that of...
10164 (make_pass_tm_memopt): ...new function.
10165 (pass_tm_edges): Replace declaration with that of...
10166 (make_pass_tm_edges): ...new function.
10167 (pass_split_functions): Replace declaration with that of...
10168 (make_pass_split_functions): ...new function.
10169 (pass_feedback_split_functions): Replace declaration with that of...
10170 (make_pass_feedback_split_functions): ...new function.
10171 (pass_strength_reduction): Replace declaration with that of...
10172 (make_pass_strength_reduction): ...new function.
10173 (pass_ipa_lower_emutls): Replace declaration with that of...
10174 (make_pass_ipa_lower_emutls): ...new function.
10175 (pass_ipa_function_and_variable_visibility): Replace declaration with
10176 that of...
10177 (make_pass_ipa_function_and_variable_visibility): ...new function.
10178 (pass_ipa_tree_profile): Replace declaration with that of...
10179 (make_pass_ipa_tree_profile): ...new function.
10180 (pass_early_local_passes): Replace declaration with that of...
10181 (make_pass_early_local_passes): ...new function.
10182 (pass_ipa_whole_program_visibility): Replace declaration with that
10183 of...
10184 (make_pass_ipa_whole_program_visibility): ...new function.
10185 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10186 (make_pass_ipa_lto_gimple_out): ...new function.
10187 (pass_ipa_increase_alignment): Replace declaration with that of...
10188 (make_pass_ipa_increase_alignment): ...new function.
10189 (pass_ipa_inline): Replace declaration with that of...
10190 (make_pass_ipa_inline): ...new function.
10191 (pass_ipa_free_lang_data): Replace declaration with that of...
10192 (make_pass_ipa_free_lang_data): ...new function.
10193 (pass_ipa_free_inline_summary): Replace declaration with that of...
10194 (make_pass_ipa_free_inline_summary): ...new function.
10195 (pass_ipa_cp): Replace declaration with that of...
10196 (make_pass_ipa_cp): ...new function.
10197 (pass_ipa_reference): Replace declaration with that of...
10198 (make_pass_ipa_reference): ...new function.
10199 (pass_ipa_pure_const): Replace declaration with that of...
10200 (make_pass_ipa_pure_const): ...new function.
10201 (pass_ipa_pta): Replace declaration with that of...
10202 (make_pass_ipa_pta): ...new function.
10203 (pass_ipa_lto_finish_out): Replace declaration with that of...
10204 (make_pass_ipa_lto_finish_out): ...new function.
10205 (pass_ipa_tm): Replace declaration with that of...
10206 (make_pass_ipa_tm): ...new function.
10207 (pass_ipa_profile): Replace declaration with that of...
10208 (make_pass_ipa_profile): ...new function.
10209 (pass_ipa_cdtor_merge): Replace declaration with that of...
10210 (make_pass_ipa_cdtor_merge): ...new function.
10211 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10212 of...
10213 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10214 (pass_init_datastructures): Replace declaration with that of...
10215 (make_pass_init_datastructures): ...new function.
10216 (pass_fixup_cfg): Replace declaration with that of...
10217 (make_pass_fixup_cfg): ...new function.
10218 (pass_expand): Replace declaration with that of...
10219 (make_pass_expand): ...new function.
10220 (pass_instantiate_virtual_regs): Replace declaration with that of...
10221 (make_pass_instantiate_virtual_regs): ...new function.
10222 (pass_rtl_fwprop): Replace declaration with that of...
10223 (make_pass_rtl_fwprop): ...new function.
10224 (pass_rtl_fwprop_addr): Replace declaration with that of...
10225 (make_pass_rtl_fwprop_addr): ...new function.
10226 (pass_jump): Replace declaration with that of...
10227 (make_pass_jump): ...new function.
10228 (pass_jump2): Replace declaration with that of...
10229 (make_pass_jump2): ...new function.
10230 (pass_lower_subreg): Replace declaration with that of...
10231 (make_pass_lower_subreg): ...new function.
10232 (pass_cse): Replace declaration with that of...
10233 (make_pass_cse): ...new function.
10234 (pass_fast_rtl_dce): Replace declaration with that of...
10235 (make_pass_fast_rtl_dce): ...new function.
10236 (pass_ud_rtl_dce): Replace declaration with that of...
10237 (make_pass_ud_rtl_dce): ...new function.
10238 (pass_rtl_dce): Replace declaration with that of...
10239 (make_pass_rtl_dce): ...new function.
10240 (pass_rtl_dse1): Replace declaration with that of...
10241 (make_pass_rtl_dse1): ...new function.
10242 (pass_rtl_dse2): Replace declaration with that of...
10243 (make_pass_rtl_dse2): ...new function.
10244 (pass_rtl_dse3): Replace declaration with that of...
10245 (make_pass_rtl_dse3): ...new function.
10246 (pass_rtl_cprop): Replace declaration with that of...
10247 (make_pass_rtl_cprop): ...new function.
10248 (pass_rtl_pre): Replace declaration with that of...
10249 (make_pass_rtl_pre): ...new function.
10250 (pass_rtl_hoist): Replace declaration with that of...
10251 (make_pass_rtl_hoist): ...new function.
10252 (pass_rtl_store_motion): Replace declaration with that of...
10253 (make_pass_rtl_store_motion): ...new function.
10254 (pass_cse_after_global_opts): Replace declaration with that of...
10255 (make_pass_cse_after_global_opts): ...new function.
10256 (pass_rtl_ifcvt): Replace declaration with that of...
10257 (make_pass_rtl_ifcvt): ...new function.
10258 (pass_into_cfg_layout_mode): Replace declaration with that of...
10259 (make_pass_into_cfg_layout_mode): ...new function.
10260 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10261 (make_pass_outof_cfg_layout_mode): ...new function.
10262 (pass_loop2): Replace declaration with that of...
10263 (make_pass_loop2): ...new function.
10264 (pass_rtl_loop_init): Replace declaration with that of...
10265 (make_pass_rtl_loop_init): ...new function.
10266 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10267 (make_pass_rtl_move_loop_invariants): ...new function.
10268 (pass_rtl_unswitch): Replace declaration with that of...
10269 (make_pass_rtl_unswitch): ...new function.
10270 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10271 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10272 (pass_rtl_doloop): Replace declaration with that of...
10273 (make_pass_rtl_doloop): ...new function.
10274 (pass_rtl_loop_done): Replace declaration with that of...
10275 (make_pass_rtl_loop_done): ...new function.
10276 (pass_web): Replace declaration with that of...
10277 (make_pass_web): ...new function.
10278 (pass_cse2): Replace declaration with that of...
10279 (make_pass_cse2): ...new function.
10280 (pass_df_initialize_opt): Replace declaration with that of...
10281 (make_pass_df_initialize_opt): ...new function.
10282 (pass_df_initialize_no_opt): Replace declaration with that of...
10283 (make_pass_df_initialize_no_opt): ...new function.
10284 (pass_reginfo_init): Replace declaration with that of...
10285 (make_pass_reginfo_init): ...new function.
10286 (pass_inc_dec): Replace declaration with that of...
10287 (make_pass_inc_dec): ...new function.
10288 (pass_stack_ptr_mod): Replace declaration with that of...
10289 (make_pass_stack_ptr_mod): ...new function.
10290 (pass_initialize_regs): Replace declaration with that of...
10291 (make_pass_initialize_regs): ...new function.
10292 (pass_combine): Replace declaration with that of...
10293 (make_pass_combine): ...new function.
10294 (pass_if_after_combine): Replace declaration with that of...
10295 (make_pass_if_after_combine): ...new function.
10296 (pass_ree): Replace declaration with that of...
10297 (make_pass_ree): ...new function.
10298 (pass_partition_blocks): Replace declaration with that of...
10299 (make_pass_partition_blocks): ...new function.
10300 (pass_match_asm_constraints): Replace declaration with that of...
10301 (make_pass_match_asm_constraints): ...new function.
10302 (pass_regmove): Replace declaration with that of...
10303 (make_pass_regmove): ...new function.
10304 (pass_split_all_insns): Replace declaration with that of...
10305 (make_pass_split_all_insns): ...new function.
10306 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10307 (make_pass_fast_rtl_byte_dce): ...new function.
10308 (pass_lower_subreg2): Replace declaration with that of...
10309 (make_pass_lower_subreg2): ...new function.
10310 (pass_mode_switching): Replace declaration with that of...
10311 (make_pass_mode_switching): ...new function.
10312 (pass_sms): Replace declaration with that of...
10313 (make_pass_sms): ...new function.
10314 (pass_sched): Replace declaration with that of...
10315 (make_pass_sched): ...new function.
10316 (pass_ira): Replace declaration with that of...
10317 (make_pass_ira): ...new function.
10318 (pass_reload): Replace declaration with that of...
10319 (make_pass_reload): ...new function.
10320 (pass_clean_state): Replace declaration with that of...
10321 (make_pass_clean_state): ...new function.
10322 (pass_branch_prob): Replace declaration with that of...
10323 (make_pass_branch_prob): ...new function.
10324 (pass_value_profile_transformations): Replace declaration with that
10325 of...
10326 (make_pass_value_profile_transformations): ...new function.
10327 (pass_postreload_cse): Replace declaration with that of...
10328 (make_pass_postreload_cse): ...new function.
10329 (pass_gcse2): Replace declaration with that of...
10330 (make_pass_gcse2): ...new function.
10331 (pass_split_after_reload): Replace declaration with that of...
10332 (make_pass_split_after_reload): ...new function.
10333 (pass_branch_target_load_optimize1): Replace declaration with that
10334 of...
10335 (make_pass_branch_target_load_optimize1): ...new function.
10336 (pass_thread_prologue_and_epilogue): Replace declaration with that
10337 of...
10338 (make_pass_thread_prologue_and_epilogue): ...new function.
10339 (pass_stack_adjustments): Replace declaration with that of...
10340 (make_pass_stack_adjustments): ...new function.
10341 (pass_peephole2): Replace declaration with that of...
10342 (make_pass_peephole2): ...new function.
10343 (pass_if_after_reload): Replace declaration with that of...
10344 (make_pass_if_after_reload): ...new function.
10345 (pass_regrename): Replace declaration with that of...
10346 (make_pass_regrename): ...new function.
10347 (pass_cprop_hardreg): Replace declaration with that of...
10348 (make_pass_cprop_hardreg): ...new function.
10349 (pass_reorder_blocks): Replace declaration with that of...
10350 (make_pass_reorder_blocks): ...new function.
10351 (pass_branch_target_load_optimize2): Replace declaration with that
10352 of...
10353 (make_pass_branch_target_load_optimize2): ...new function.
10354 (pass_leaf_regs): Replace declaration with that of...
10355 (make_pass_leaf_regs): ...new function.
10356 (pass_split_before_sched2): Replace declaration with that of...
10357 (make_pass_split_before_sched2): ...new function.
10358 (pass_compare_elim_after_reload): Replace declaration with that of...
10359 (make_pass_compare_elim_after_reload): ...new function.
10360 (pass_sched2): Replace declaration with that of...
10361 (make_pass_sched2): ...new function.
10362 (pass_stack_regs): Replace declaration with that of...
10363 (make_pass_stack_regs): ...new function.
10364 (pass_stack_regs_run): Replace declaration with that of...
10365 (make_pass_stack_regs_run): ...new function.
10366 (pass_df_finish): Replace declaration with that of...
10367 (make_pass_df_finish): ...new function.
10368 (pass_compute_alignments): Replace declaration with that of...
10369 (make_pass_compute_alignments): ...new function.
10370 (pass_duplicate_computed_gotos): Replace declaration with that of...
10371 (make_pass_duplicate_computed_gotos): ...new function.
10372 (pass_variable_tracking): Replace declaration with that of...
10373 (make_pass_variable_tracking): ...new function.
10374 (pass_free_cfg): Replace declaration with that of...
10375 (make_pass_free_cfg): ...new function.
10376 (pass_machine_reorg): Replace declaration with that of...
10377 (make_pass_machine_reorg): ...new function.
10378 (pass_cleanup_barriers): Replace declaration with that of...
10379 (make_pass_cleanup_barriers): ...new function.
10380 (pass_delay_slots): Replace declaration with that of...
10381 (make_pass_delay_slots): ...new function.
10382 (pass_split_for_shorten_branches): Replace declaration with that of...
10383 (make_pass_split_for_shorten_branches): ...new function.
10384 (pass_split_before_regstack): Replace declaration with that of...
10385 (make_pass_split_before_regstack): ...new function.
10386 (pass_convert_to_eh_region_ranges): Replace declaration with that
10387 of...
10388 (make_pass_convert_to_eh_region_ranges): ...new function.
10389 (pass_shorten_branches): Replace declaration with that of...
10390 (make_pass_shorten_branches): ...new function.
10391 (pass_set_nothrow_function_flags): Replace declaration with that of...
10392 (make_pass_set_nothrow_function_flags): ...new function.
10393 (pass_dwarf2_frame): Replace declaration with that of...
10394 (make_pass_dwarf2_frame): ...new function.
10395 (pass_final): Replace declaration with that of...
10396 (make_pass_final): ...new function.
10397 (pass_rtl_seqabstr): Replace declaration with that of...
10398 (make_pass_rtl_seqabstr): ...new function.
10399 (pass_release_ssa_names): Replace declaration with that of...
10400 (make_pass_release_ssa_names): ...new function.
10401 (pass_early_inline): Replace declaration with that of...
10402 (make_pass_early_inline): ...new function.
10403 (pass_inline_parameters): Replace declaration with that of...
10404 (make_pass_inline_parameters): ...new function.
10405 (pass_update_address_taken): Replace declaration with that of...
10406 (make_pass_update_address_taken): ...new function.
10407 (pass_convert_switch): Replace declaration with that of...
10408 (make_pass_convert_switch): ...new function.
10409 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10410 to a subclass of simple_ipa_opt_pass along with...
10411 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10412 (make_pass_ipa_tree_profile): ...new function.
10413 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10414 subclass of gimple_opt_pass along with...
10415 (pass_data_sra_early): ...new pass_data instance and...
10416 (make_pass_sra_early): ...new function.
10417 (pass_sra): Convert from a global struct to a subclass of
10418 gimple_opt_pass along with...
10419 (pass_data_sra): ...new pass_data instance and...
10420 (make_pass_sra): ...new function.
10421 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10422 gimple_opt_pass along with...
10423 (pass_data_early_ipa_sra): ...new pass_data instance and...
10424 (make_pass_early_ipa_sra): ...new function.
10425 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10426 subclass of gimple_opt_pass along with...
10427 (pass_data_ccp): ...new pass_data instance and...
10428 (make_pass_ccp): ...new function.
10429 (pass_fold_builtins): Convert from a global struct to a subclass of
10430 gimple_opt_pass along with...
10431 (pass_data_fold_builtins): ...new pass_data instance and...
10432 (make_pass_fold_builtins): ...new function.
10433 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10434 subclass of gimple_opt_pass along with...
10435 (pass_data_copy_prop): ...new pass_data instance and...
10436 (make_pass_copy_prop): ...new function.
10437 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10438 global struct to a subclass of gimple_opt_pass along with...
10439 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10440 (make_pass_rename_ssa_copies): ...new function.
10441 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10442 subclass of gimple_opt_pass along with...
10443 (pass_data_dce): ...new pass_data instance and...
10444 (make_pass_dce): ...new function.
10445 (pass_dce_loop): Convert from a global struct to a subclass of
10446 gimple_opt_pass along with...
10447 (pass_data_dce_loop): ...new pass_data instance and...
10448 (make_pass_dce_loop): ...new function.
10449 (pass_cd_dce): Convert from a global struct to a subclass of
10450 gimple_opt_pass along with...
10451 (pass_data_cd_dce): ...new pass_data instance and...
10452 (make_pass_cd_dce): ...new function.
10453 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10454 subclass of gimple_opt_pass along with...
10455 (pass_data_dominator): ...new pass_data instance and...
10456 (make_pass_dominator): ...new function.
10457 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10458 gimple_opt_pass along with...
10459 (pass_data_phi_only_cprop): ...new pass_data instance and...
10460 (make_pass_phi_only_cprop): ...new function.
10461 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10462 subclass of gimple_opt_pass along with...
10463 (pass_data_dse): ...new pass_data instance and...
10464 (make_pass_dse): ...new function.
10465 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10466 a subclass of gimple_opt_pass along with...
10467 (pass_data_forwprop): ...new pass_data instance and...
10468 (make_pass_forwprop): ...new function.
10469 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10470 struct to a subclass of gimple_opt_pass along with...
10471 (pass_data_tree_ifcombine): ...new pass_data instance and...
10472 (make_pass_tree_ifcombine): ...new function.
10473 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10474 subclass of gimple_opt_pass along with...
10475 (pass_data_ch): ...new pass_data instance and...
10476 (make_pass_ch): ...new function.
10477 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10478 subclass of gimple_opt_pass along with...
10479 (pass_data_tree_loop): ...new pass_data instance and...
10480 (make_pass_tree_loop): ...new function.
10481 (pass_tree_loop_init): Convert from a global struct to a subclass of
10482 gimple_opt_pass along with...
10483 (pass_data_tree_loop_init): ...new pass_data instance and...
10484 (make_pass_tree_loop_init): ...new function.
10485 (pass_lim): Convert from a global struct to a subclass of
10486 gimple_opt_pass along with...
10487 (pass_data_lim): ...new pass_data instance and...
10488 (make_pass_lim): ...new function.
10489 (pass_tree_unswitch): Convert from a global struct to a subclass of
10490 gimple_opt_pass along with...
10491 (pass_data_tree_unswitch): ...new pass_data instance and...
10492 (make_pass_tree_unswitch): ...new function.
10493 (pass_predcom): Convert from a global struct to a subclass of
10494 gimple_opt_pass along with...
10495 (pass_data_predcom): ...new pass_data instance and...
10496 (make_pass_predcom): ...new function.
10497 (pass_vectorize): Convert from a global struct to a subclass of
10498 gimple_opt_pass along with...
10499 (pass_data_vectorize): ...new pass_data instance and...
10500 (make_pass_vectorize): ...new function.
10501 (pass_graphite): Convert from a global struct to a subclass of
10502 gimple_opt_pass along with...
10503 (pass_data_graphite): ...new pass_data instance and...
10504 (make_pass_graphite): ...new function.
10505 (pass_graphite_transforms): Convert from a global struct to a subclass
10506 of gimple_opt_pass along with...
10507 (pass_data_graphite_transforms): ...new pass_data instance and...
10508 (make_pass_graphite_transforms): ...new function.
10509 (pass_check_data_deps): Convert from a global struct to a subclass of
10510 gimple_opt_pass along with...
10511 (pass_data_check_data_deps): ...new pass_data instance and...
10512 (make_pass_check_data_deps): ...new function.
10513 (pass_iv_canon): Convert from a global struct to a subclass of
10514 gimple_opt_pass along with...
10515 (pass_data_iv_canon): ...new pass_data instance and...
10516 (make_pass_iv_canon): ...new function.
10517 (pass_scev_cprop): Convert from a global struct to a subclass of
10518 gimple_opt_pass along with...
10519 (pass_data_scev_cprop): ...new pass_data instance and...
10520 (make_pass_scev_cprop): ...new function.
10521 (pass_record_bounds): Convert from a global struct to a subclass of
10522 gimple_opt_pass along with...
10523 (pass_data_record_bounds): ...new pass_data instance and...
10524 (make_pass_record_bounds): ...new function.
10525 (pass_complete_unroll): Convert from a global struct to a subclass of
10526 gimple_opt_pass along with...
10527 (pass_data_complete_unroll): ...new pass_data instance and...
10528 (make_pass_complete_unroll): ...new function.
10529 (pass_complete_unrolli): Convert from a global struct to a subclass of
10530 gimple_opt_pass along with...
10531 (pass_data_complete_unrolli): ...new pass_data instance and...
10532 (make_pass_complete_unrolli): ...new function.
10533 (pass_parallelize_loops): Convert from a global struct to a subclass
10534 of gimple_opt_pass along with...
10535 (pass_data_parallelize_loops): ...new pass_data instance and...
10536 (make_pass_parallelize_loops): ...new function.
10537 (pass_loop_prefetch): Convert from a global struct to a subclass of
10538 gimple_opt_pass along with...
10539 (pass_data_loop_prefetch): ...new pass_data instance and...
10540 (make_pass_loop_prefetch): ...new function.
10541 (pass_iv_optimize): Convert from a global struct to a subclass of
10542 gimple_opt_pass along with...
10543 (pass_data_iv_optimize): ...new pass_data instance and...
10544 (make_pass_iv_optimize): ...new function.
10545 (pass_tree_loop_done): Convert from a global struct to a subclass of
10546 gimple_opt_pass along with...
10547 (pass_data_tree_loop_done): ...new pass_data instance and...
10548 (make_pass_tree_loop_done): ...new function.
10549 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10550 struct to a subclass of gimple_opt_pass along with...
10551 (pass_data_cse_reciprocals): ...new pass_data instance and...
10552 (make_pass_cse_reciprocals): ...new function.
10553 (pass_cse_sincos): Convert from a global struct to a subclass of
10554 gimple_opt_pass along with...
10555 (pass_data_cse_sincos): ...new pass_data instance and...
10556 (make_pass_cse_sincos): ...new function.
10557 (pass_optimize_bswap): Convert from a global struct to a subclass of
10558 gimple_opt_pass along with...
10559 (pass_data_optimize_bswap): ...new pass_data instance and...
10560 (make_pass_optimize_bswap): ...new function.
10561 (pass_optimize_widening_mul): Convert from a global struct to a
10562 subclass of gimple_opt_pass along with...
10563 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10564 (make_pass_optimize_widening_mul): ...new function.
10565 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10566 subclass of gimple_opt_pass along with...
10567 (pass_data_phiopt): ...new pass_data instance and...
10568 (make_pass_phiopt): ...new function.
10569 (pass_cselim): Convert from a global struct to a subclass of
10570 gimple_opt_pass along with...
10571 (pass_data_cselim): ...new pass_data instance and...
10572 (make_pass_cselim): ...new function.
10573 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10574 subclass of gimple_opt_pass along with...
10575 (pass_data_phiprop): ...new pass_data instance and...
10576 (make_pass_phiprop): ...new function.
10577 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10578 subclass of gimple_opt_pass along with...
10579 (pass_data_pre): ...new pass_data instance and...
10580 (make_pass_pre): ...new function.
10581 (pass_fre): Convert from a global struct to a subclass of
10582 gimple_opt_pass along with...
10583 (pass_data_fre): ...new pass_data instance and...
10584 (make_pass_fre): ...new function.
10585 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10586 subclass of gimple_opt_pass along with...
10587 (pass_data_reassoc): ...new pass_data instance and...
10588 (make_pass_reassoc): ...new function.
10589 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10590 subclass of gimple_opt_pass along with...
10591 (pass_data_sink_code): ...new pass_data instance and...
10592 (make_pass_sink_code): ...new function.
10593 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10594 subclass of gimple_opt_pass along with...
10595 (pass_data_strlen): ...new pass_data instance and...
10596 (make_pass_strlen): ...new function.
10597 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10598 struct to a subclass of gimple_opt_pass along with...
10599 (pass_data_build_alias): ...new pass_data instance and...
10600 (make_pass_build_alias): ...new function.
10601 (pass_build_ealias): Convert from a global struct to a subclass of
10602 gimple_opt_pass along with...
10603 (pass_data_build_ealias): ...new pass_data instance and...
10604 (make_pass_build_ealias): ...new function.
10605 (pass_ipa_pta): Convert from a global struct to a subclass of
10606 simple_ipa_opt_pass along with...
10607 (pass_data_ipa_pta): ...new pass_data instance and...
10608 (make_pass_ipa_pta): ...new function.
10609 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10610 subclass of gimple_opt_pass along with...
10611 (pass_data_uncprop): ...new pass_data instance and...
10612 (make_pass_uncprop): ...new function.
10613 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10614 global struct to a subclass of gimple_opt_pass along with...
10615 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10616 (make_pass_late_warn_uninitialized): ...new function.
10617 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10618 to a subclass of gimple_opt_pass along with...
10619 (pass_data_init_datastructures): ...new pass_data instance and...
10620 (make_pass_init_datastructures): ...new function.
10621 (pass_early_warn_uninitialized): Convert from a global struct to a
10622 subclass of gimple_opt_pass along with...
10623 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10624 (make_pass_early_warn_uninitialized): ...new function.
10625 (pass_update_address_taken): Convert from a global struct to a
10626 subclass of gimple_opt_pass along with...
10627 (pass_data_update_address_taken): ...new pass_data instance and...
10628 (make_pass_update_address_taken): ...new function.
10629 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10630 struct to a subclass of gimple_opt_pass along with...
10631 (pass_data_release_ssa_names): ...new pass_data instance and...
10632 (make_pass_release_ssa_names): ...new function.
10633 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10634 subclass of gimple_opt_pass along with...
10635 (pass_data_stdarg): ...new pass_data instance and...
10636 (make_pass_stdarg): ...new function.
10637 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10638 global struct to a subclass of gimple_opt_pass along with...
10639 (pass_data_convert_switch): ...new pass_data instance and...
10640 (make_pass_convert_switch): ...new function.
10641 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10642 to a subclass of gimple_opt_pass along with...
10643 (pass_data_tail_recursion): ...new pass_data instance and...
10644 (make_pass_tail_recursion): ...new function.
10645 (pass_tail_calls): Convert from a global struct to a subclass of
10646 gimple_opt_pass along with...
10647 (pass_data_tail_calls): ...new pass_data instance and...
10648 (make_pass_tail_calls): ...new function.
10649 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10650 struct to a subclass of gimple_opt_pass along with...
10651 (pass_data_lower_vector): ...new pass_data instance and...
10652 (make_pass_lower_vector): ...new function.
10653 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10654 gimple_opt_pass along with...
10655 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10656 (make_pass_lower_vector_ssa): ...new function.
10657 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10658 to a subclass of gimple_opt_pass along with...
10659 (pass_data_slp_vectorize): ...new pass_data instance and...
10660 (make_pass_slp_vectorize): ...new function.
10661 (pass_ipa_increase_alignment): Convert from a global struct to a
10662 subclass of simple_ipa_opt_pass along with...
10663 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10664 (make_pass_ipa_increase_alignment): ...new function.
10665 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10666 gimple_opt_pass along with...
10667 (pass_data_vrp): ...new pass_data instance and...
10668 (make_pass_vrp): ...new function.
10669 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10670 subclass of simple_ipa_opt_pass along with...
10671 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10672 (make_pass_ipa_free_lang_data): ...new function.
10673 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10674 gimple_opt_pass along with...
10675 (pass_data_tsan): ...new pass_data instance and...
10676 (make_pass_tsan): ...new function.
10677 (pass_tsan_O0): Convert from a global struct to a subclass of
10678 gimple_opt_pass along with...
10679 (pass_data_tsan_O0): ...new pass_data instance and...
10680 (make_pass_tsan_O0): ...new function.
10681 * var-tracking.c (pass_variable_tracking): Convert from a global
10682 struct to a subclass of rtl_opt_pass along with...
10683 (pass_data_variable_tracking): ...new pass_data instance and...
10684 (make_pass_variable_tracking): ...new function.
10685 * web.c (pass_web): Convert from a global struct to a subclass of
10686 rtl_opt_pass along with...
10687 (pass_data_web): ...new pass_data instance and...
10688 (make_pass_web): ...new function.
10689 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10690 declaration with that of...
10691 (make_pass_mode_switch_use): ...new function.
10692 (pass_resolve_sw_modes): Replace declaration with that of...
10693 (make_pass_resolve_sw_modes): ...new function.
10694 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10695 from a global struct to a subclass of rtl_opt_pass along with...
10696 (pass_data_mode_switch_use): ...new pass_data instance and...
10697 (make_pass_mode_switch_use): ...new function.
10698 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10699 from a global struct to a subclass of rtl_opt_pass along with...
10700 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10701 (make_pass_resolve_sw_modes): ...new function.
10702 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10703 struct to a subclass of rtl_opt_pass along with...
10704 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10705 (make_pass_insert_vzeroupper): ...new function.
10706 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10707 global struct to a subclass of rtl_opt_pass along with...
10708 (pass_data_work_around_errata): ...new pass_data instance and...
10709 (make_pass_work_around_errata): ...new function.
10710 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10711 struct to a subclass of rtl_opt_pass along with...
10712 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10713 (make_pass_mips_machine_reorg2): ...new function.
10714
6a389ed5
DM		 107152013-08-05 David Malcolm <dmalcolm@redhat.com>
10716
10717 * passes.c (pass_manager::operator new): New.
10718
f7695dbf
DM		 107192013-08-05 David Malcolm <dmalcolm@redhat.com>
10720
10721 Handwritten part of conversion of passes to C++ classes.
10722
10723 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10724 (toplev.o): Add dep on PASS_MANAGER_H.
10725 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10726 of early local pases to reflect this moving from a global to a
10727 member of gcc::pass_manager.
10728 (cgraph_add_new_function): Likewise.
10729 * lto-cgraph.c (lto_output_node): Update for conversion of
10730 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10731 * passes.c (opt_pass::clone): New.
10732 (opt_pass::gate): New.
10733 (opt_pass::execute): New.
10734 (opt_pass::opt_pass): New.
10735 (pass_manager::execute_early_local_passes): New.
10736 (pass_manager::execute_pass_mode_switching): new.
10737 (finish_optimization_passes): Convert to...
10738 (pass_manager::finish_optimization_passes): ...this.
10739 (finish_optimization_passes): Update for conversion of passes to
10740 C++ classes.
10741 (register_dump_files_1): Use has_gate since we cannot portably
10742 check a vtable entry against NULL.
10743 (dump_one_pass): Likewise.
10744 (ipa_write_summaries_2): Likewise.
10745 (ipa_write_optimization_summaries_1): Likewise.
10746 (ipa_read_summaries_1): Likewise.
10747 (ipa_read_optimization_summaries_1): Likewise.
10748 (execute_ipa_stmt_fixups): Likewise.
10749 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10750 pass-creation functions rather than wiring up globals, and
10751 storing the results in fields of pass_manager generated using
10752 pass-instances.def.
10753 (pass_manager::dump_profile_report): Update for conversion of
10754 passes to C++ classes.
10755 (pass_manager::execute_ipa_summary_passes): Likewise.
10756 (execute_one_ipa_transform_pass): Likewise.
10757 (execute_one_pass): Use has_gate and has_execute since we cannot
10758 portably check a vtable entry against NULL.
10759 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10760 (pass_manager): Use pass-instances.def to add fields for the
10761 various pass instances.
10762 * toplev.c (finalize): Update for move of
10763 finish_optimization_passes to a method of gcc::pass_manager.
10764 * toplev.h (finish_optimization_passes): Move to method of class
10765 pass_manager.
10766 * tree-pass.h (struct pass_data): New.
03b0ee0a 10767 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 10768 (opt_pass::gate): Convert to virtual function.
10769 (opt_pass::~opt_pass): New.
10770 (opt_pass::clone): New.
10771 (opt_pass::execute): Convert to virtual function.
10772 (opt_pass::opt_pass): New.
10773 (opt_pass::ctxt_): new.
10774 (gimple_opt_pass): Convert to subclass of opt_pass.
10775 (gimple_opt_pass::gimple_opt_pass): New.
10776 (rtl_opt_pass): Convert to subclass of opt_pass.
10777 (rtl_opt_pass::rtl_opt_pass): New.
10778 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10779 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10780 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10781 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10782 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10783 invocation of pass_mode_switching to reflect this moving from a
10784 global to a member of gcc::pass_manager.
10785 (ix86_option_override): Rework how pass_insert_vzeroupper is
10786 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 10787 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 10788
d0b2f831
RE		 107892013-08-05 Richard Earnshaw <rearnsha@arm.com>
10790
10791 PR rtl-optimization/57708
10792 * recog.c (peep2_find_free_register): Validate all regs in a
10793 multi-reg mode.
10794
51a5c0c2
JH		 107952013-08-05 Jan Hubicka <jh@suse.cz>
10796
10797 PR lto/57602
03b0ee0a
UB		 10798 * cgraph.c (verify_cgraph_node): Accept local flags from other
10799 partitions.
51a5c0c2
JH		 10800 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10801 (function_and_variable_visibility): Likewise.
10802 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10803
65f0a120
GDR		 108042013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10805
10806 * graph.c (init_graph_slim_pretty_print): Remove.
10807 (print_graph_cfg): Do not call it. Use local pretty printer.
10808 (start_graph_dump): Likewise.
10809
b3f80694
GDR		 108102013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10811
10812 * gimple-pretty-print.c (buffer): Remove.
10813 (initialized): Likewise.
10814 (maybe_init_pretty_print): Likewise.
10815 (print_gimple_stmt): Do not call it. Use non-static local
10816 pretty_printer variable.
10817 (print_gimple_expr): Likewise.
10818 (print_gimple_seq): Likewise.
10819 (gimple_dump_bb): Likewise.
10820
11a877b3
GDR		 108212013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10822
10823 * asan.c (asan_pp): Remove.
10824 (asan_pp_initialized): Likewise.
10825 (asan_pp_initialize): Likewise.
10826 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10827 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10828 (asan_add_global): Likewise.
10829
b066401f
GDR		 108302013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10831
10832 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10833 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10834 * pretty-print.c (pp_formatted_text_data): Likewise.
10835 (pp_indent): Rename from pp_base_indent.
10836 (pp_format): Rename from pp_base_format.
10837 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10838 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10839 (pp_flush): Rename from pp_base_flush.
10840 (pp_set_line_maximum_length): Rename from
10841 pp_base_set_line_maximum_length.
10842 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10843 (pp_set_prefix): Rename from pp_base_set_prefix.
10844 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10845 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10846 (pp_append_text): Rename from pp_base_append_text.
10847 (pp_formatted_text): Rename from pp_base_formatted_text.
10848 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10849 (pp_remaining_character_count_for_line): Rename from
10850 pp_base_remaining_character_count_for_line.
10851 (pp_newline): Rename from pp_base_newline.
10852 (pp_character): Rename from pp_base_character.
10853 (pp_string): Rename from pp_base_string.
10854 (pp_maybe_space): Rename from pp_base_maybe_space.
10855 * asan.c (asan_pp_string): Adjust.
10856 (asan_emit_stack_protection): Likewise.
10857 (asan_add_global): Likewise.
10858 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10859 * tree-mudflap.c (mf_varname_tree): Likewise.
10860 * tree-pretty-print.c (pp_tree_identifier): Rename from
10861 pp_base_tree_identifier.
10862 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10863 Declare as function.
10864
137a1a27
GDR		 108652013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10866
10867 * pretty-print.h (pp_bar_bar): New.
10868 (pp_ampersand_ampersand): Likewise.
10869 (pp_less_equal): Likewise.
10870 (pp_greater_equal): Likewise.
10871 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10872 printer functions instead of pp_string or operators and punctuators.
10873 (dump_gimple_call): Likewise.
10874 (dump_gimple_omp_for): Likewise.
10875 (dump_gimple_transaction): Likewise.
10876 (dump_gimple_phi): Likewise.
10877 (pp_gimple_stmt_1): Likewise.
10878 * sched-vis.c (print_insn): Likewise.
10879 * tree-mudflap.c (mf_varname_tree): Likewise.
10880 * tree-pretty-print.c (dump_block_node): Likewise.
10881 (dump_generic_node): Likewise.
10882
815effe1
JH		 108832013-08-02 Jan Hubicka <jh@suse.cz>
10884
10885 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10886 boundaries.
10887 * lto-streamer-out.c (tree_is_indexable): Results decls and
10888 parm decls are not indexable.
10889 (DFS_write_tree_body): Do not follow args and results.
10890 (hash_tree): Likewise.
10891 (output_functions): Rearrange so struct function is needed
10892 only when real body is output; be able to also ouptut abstract
10893 functions; output DECL_ARGUMENTS and DECL_RESULT.
10894 (lto_output): When not in WPA, ale store abstract functions.
10895 (write_symbol): Do not care about RESULT_DECL.
10896 (output_symbol_p): Handle correctly sbtract decls.
10897 * lto-streamer-in.c (input_function): Rearrange so struct
10898 function can be NULL at entry; allow streaming of
10899 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10900 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10901 sanity check during LTO.
10902 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10903 RESULT_DECl and DECL_ARGUMENTS.
10904 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10905 Likewise.
10906
07838b13
GDR		 109072013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10908
10909 * pretty-print.h (pp_underscore): New.
10910 (pp_comma): Tidy.
10911 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10912 printer functions instead of pp_character.
10913 (dump_binary_rhs): Likewise.
10914 (dump_ternary_rhs): Likewise.
10915 (dump_gimple_call_args): Likewise.
10916 (pp_points_to_solution): Likewise.
10917 (dump_gimple_call): Likewise.
10918 (dump_gimple_switch): Likewise.
10919 (dump_gimple_cond): Likewise.
10920 (dump_gimple_bind): Likewise.
10921 (dump_gimple_try): Likewise.
10922 (dump_gimple_omp_for): Likewise.
10923 (dump_gimple_omp_continue): Likewise.
10924 (dump_gimple_omp_single): Likewise.
10925 (dump_gimple_omp_sections): Likewise.
10926 (dump_gimple_omp_block): Likewise.
10927 (dump_gimple_omp_critical): Likewise.
10928 (dump_gimple_transaction): Likewise.
10929 (dump_gimple_asm): Likewise.
10930 (dump_gimple_phi): Likewise.
10931 (dump_gimple_omp_parallel): Likewise.
10932 (dump_gimple_omp_task): Likewise.
10933 (dump_gimple_omp_atomic_load): Likewise.
10934 (dump_gimple_omp_atomic_store): Likewise.
10935 (dump_gimple_mem_ops): Likewise.
10936 (pp_gimple_stmt_1): Likewise.
10937 (pp_cfg_jump): Likewise.
10938 (dump_implicit_edges): Likewise.
10939 (gimple_dump_bb_for_graph): Likewise.
10940 * graph.c (draw_cfg_node): Likewise.
10941 * langhooks.c (lhd_print_error_function): Likewise.
10942 * sched-vis.c (print_exp): Likewise.
10943 (print_value): Likewise.
10944 (print_pattern): Likewise.
10945 (print_insn): Likewise.
10946 (rtl_dump_bb_for_graph): Likewise.
10947 * tree-pretty-print.c (dump_function_declaration): Likewise.
10948 (dump_array_domain): Likewise.
10949 (dump_omp_clause): Likewise.
10950 (dump_location): Likewise.
10951 (dump_generic_node): Likewise.
10952 (print_struct_decl): Likewise.
10953 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10954
bb0d2039
BS		 109552013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10956
10957 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10958 candidate table when replacing a candidate statement.
10959 (replace_rhs_if_not_dup): Likewise.
10960 (replace_one_candidate): Likewise.
10961
bc0ec027 109622013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 10963 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 10964
10965 * cgraphunit.c (add_new_function): Fix logic when adding from
10966 late IPA pass.
10967 (assemble_thunk): Rename to ...
03b0ee0a
UB		 10968 (expand_thunk); .. this one; export; get it working with
10969 general functions; make produced gimple valid.
bc0ec027
JH		 10970 * cgraph.h (expand_thunk): Declare.
10971
0e8853ee
JH		 109722013-08-02 Jan Hubicka <jh@suse.cz>
10973
03b0ee0a
UB		 10974 * ipa-cp.c (gather_context_independent_values): Use
10975 ipa_get_param_move_cost.
10976 (get_replacement_map): Remove PARAM; move parameter folding
10977 into tree-inline.c
0e8853ee
JH		 10978 (create_specialized_node): Update.
10979 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10980 assert that we have gimple body; update move_cost.
10981 (count_formal_params): Assert that we have gimple body.
10982 (ipa_dump_param): New function.
10983 (ipa_alloc_node_params): Break out from ...
10984 (ipa_initialize_node_params): ... here.
10985 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10986 (ipa_write_node_info): Stream move costs.
10987 (ipa_read_node_info): Read move costs.
10988 (ipa_update_after_lto_read): Do not recompute node params.
10989 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10990 (ipa_get_param): Check we are not in WPA.
10991 (ipa_get_param_move_cost): New.
10992 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10993 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10994 parm numbers to be present.
10995
ecee672b
VM		 109962013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10997
10998 PR rtl-optimization/58048
10999 * lra-constraints.c (process_alt_operands): Don't check asm
11000 operand on register.
11001
13ec0527
EB		 110022013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11003
11004 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11005 the implied StoreLoad barrier for atomic operations if before.
11006
af15184a
JH		 110072013-08-02 Jan Hubicka <jh@suse.cz>
11008 Martin Liska <marxin.liska@gmail.com>
11009
03b0ee0a
UB		 11010 * cgraph.c (cgraph_function_body_availability): Do not check
11011 cgraph flags.
af15184a
JH		 11012 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11013 symtab_node_availability): Declare.
11014 * ipa.c (can_replace_by_local_alias): New.
11015 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 11016 * symtab.c (symtab_for_node_and_aliases,
11017 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 11018
01e54ef8
VM		 110192013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11020
11021 PR rtl-optimization/57963
03b0ee0a 11022 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 11023 (lra_constraints): Use them.
11024
e2fd7ca7
SN		 110252013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11026
03b0ee0a
UB		 11027 * config/arm/types.md (define_attr "type"): Add "load_acq"
11028 and "store_rel".
e2fd7ca7
SN		 11029 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11030 changes.
11031 (cortex_a53_store1): Likewise.
11032
0dcc26c3
JH		 110332013-08-01 Jan Hubicka <jh@suse.cz>
11034
03b0ee0a
UB		 11035 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11036 partitions are not needed.
0dcc26c3 11037
0bd72901
UB		 110382013-08-01 Uros Bizjak <ubizjak@gmail.com>
11039
11040 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11041 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11042 MAYBE_NON_Q_CLASS_P where appropriate.
11043
12123452
JH		 110442013-08-01 Jan Hubicka <jh@suse.cz>
11045
11046 * cgraph.h (release_function_body): Declare.
11047 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11048 of unused delcarations.
11049
ce852f9c
KT		 110502013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11051
03b0ee0a
UB		 11052 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11053 RTL form when subtracting a constant.
ce852f9c 11054
c743b246
KT		 110552013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11056
11057 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11058 Generate canonical plus rtx with negated immediate instead of minus
11059 where appropriate.
11060 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11061
c0c123ef
JH		 110622013-08-01 Jan Hubicka <jh@suse.cz>
11063
11064 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11065 (cgraph_release_function_body): Likewise.
11066 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11067 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11068 to used_as_abstract_origin.
11069 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11070 symbols used as abstract origins.
11071 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 11072 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11073 used_as_abstract_origin.
c0c123ef 11074 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 11075 used_as_abstract_origin; be ready for DECL_RESULT and
11076 DECL_ARGUMENTS to be NULL.
c0c123ef 11077
0bd72901
UB		 11078 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11079 for abstract functions.
11080 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11081 real symbols.
c0c123ef 11082
1f029433
JH		 110832013-08-01 Jan Hubicka <jh@suse.cz>
11084
11085 * profile.c (compute_value_histograms): Fix thinko.
11086
c451f4d6
SN		 110872013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11088
11089 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11090 aarch-common-protos.h to extra_headers.
11091 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11092 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11093 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11094
34b512f5
SN		 110952013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11096
11097 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11098 Include "../arm/types.md". Define "type" attribute for all patterns.
11099 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11100 attribute changes.
11101
d86e633a
MM		 111022013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11103
11104 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11105 to support power8 load fusion.
11106 (fusion_gpr_mem_load): Likewise.
11107
11108 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11109
11110 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11111 declarations for power8 load fusion.
11112 (emit_fusion_gpr_load): Likewise.
11113
11114 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11115 tuning for power8, turn on fusion mode by default. Turn on sign
11116 extending fusion mode if normal fusion mode is on, and we are at
11117 -O2 or -O3.
11118 (fusion_gpr_load_p): New function, return true if we can fuse an
11119 addis instruction with a dependent load to a GPR.
11120 (emit_fusion_gpr_load): Emit the instructions for power8 load
11121 fusion to GPRs.
11122
0bd72901 11123 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 11124 (VSX load fusion peepholes): New peepholes to fuse together an
11125 addi instruction with a VSX load instruction.
11126
11127 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11128 peepholes to fuse an addis instruction with a load to a GPR base
11129 register. If we are supporting sign extending fusions, convert
11130 sign extending loads to zero extending loads and add an explicit
11131 sign extension.
11132
c3f35647
SN		 111332013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11134
11135 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11136 aarch-common-protos.h to extra_headers.
11137 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11138 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11139 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11140 (arm_no_early_alu_shift_dep): Likewise.
11141 (arm_no_early_alu_shift_value_dep): Likewise.
11142 (arm_no_early_mul_dep): Likewise.
11143 (arm_no_early_store_addr_dep): Likewise.
11144 (arm_mac_accumulator_is_mul_result): Likewise.
11145 (arm_mac_accumulator_is_result): Likewise.
c3f35647 11146 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 11147 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11148 here to ...
c3f35647 11149 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11150 (arm_no_early_alu_shift_dep): Likewise.
11151 (arm_no_early_alu_shift_value_dep): Likewise.
11152 (arm_no_early_mul_dep): Likewise.
11153 (arm_no_early_store_addr_dep): Likewise.
11154 (arm_mac_accumulator_is_mul_result): Likewise.
11155 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN		 11156 * config/arm/aarch-common-protos.h: ... here. New file.
11157 * config/arm/t-arm (aarch-common.o): Define.
11158
111592013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11160
11161 * config/arm/arm.md: Include new file "types.md".
11162 (define_attr "type"): Move from here to ...
11163 (define_attr "mul32"): Likewise.
11164 (define_attr "mul64"): Likewise.
11165 * config/arm/types.md: ... here. New file.
11166
228c1313
SH		 111672013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11168
11169 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11170 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11171
d6e9046d 111722013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 11173
11174 * gen-pass-instances.awk: Fix offset of substr().
11175
a167b052
DM		 111762013-07-31 David Malcolm <dmalcolm@redhat.com>
11177
11178 * Makefile.in (pass-instances.def): New.
11179 (passes.o): Replace dependency on passes.def with one on
11180 pass-instances.def
11181
11182 * gen-pass-instances.awk: New.
11183
11184 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11185 rather than passes.def, updating local definition of NEXT_PASS
11186 macro to add an extra NUM parameter (currently unused).
11187
315f8c0e
DM		 111882013-07-30 David Malcolm <dmalcolm@redhat.com>
11189
11190 * Makefile.in (PASS_MANAGER_H): New.
11191 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11192 (passes.o): Likewise.
11193 (statistics.o): Likewise.
11194 (cgraphunit.o): Likewise.
11195 (context.o): Depend on PASS_MANAGER_H.
11196
11197 * pass_manager.h: New.
11198
11199 * cgraphunit.c (cgraph_add_new_function): Update for moves
11200 of globals to fields of pass_manager.
11201 (analyze_function): Likewise.
11202 (expand_function): Likewise.
11203 (ipa_passes): Likewise.
11204 (compile): Likewise.
11205
11206 * context.c (context::context): New.
11207 * context.h (context::context): New.
11208 (context::get_passes): New.
11209 (context::passes_): New.
11210
11211 * lto-cgraph.c (input_node): Update for moves of globals to
11212 fields of pass_manager.
11213
11214 * passes.c (all_passes): Remove, in favor of a field of the
11215 same name within the new class pass_manager.
11216 (all_small_ipa_passes): Likewise.
11217 (all_lowering_passes): Likewise.
11218 (all_regular_ipa_passes): Likewise.
11219 (all_late_ipa_passes): Likewise.
11220 (all_lto_gen_passes): Likewise.
11221 (passes_by_id): Likewise.
11222 (passes_by_id_size): Likewise.
11223 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11224 the new class pass_manager.
11225 (set_pass_for_id): Convert to...
11226 (pass_manager::set_pass_for_id): ...method.
11227 (get_pass_for_id): Convert to...
11228 (pass_manager::get_pass_for_id): ...method.
11229 (register_one_dump_file): Move body of implementation into...
11230 (pass_manager::register_one_dump_file): ...here.
11231 (register_dump_files_1): Convert to...
11232 (pass_manager::register_dump_files_1): ...method.
11233 (register_dump_files): Convert to...
11234 (pass_manager::register_dump_files): ...method.
11235 (create_pass_tab): Update for moves of globals to fields of
11236 pass_manager.
11237 (dump_passes): Move body of implementation into...
11238 (pass_manager::dump_passes): ...here.
11239 (register_pass): Move body of implementation into...
11240 (pass_manager::register_pass): ...here.
11241 (init_optimization_passes): Convert into...
11242 (pass_manager::pass_manager): ...constructor for new
11243 pass_manager class, and initialize the pass_lists array.
11244 (check_profile_consistency): Update for moves of globals to
11245 fields of pass_manager.
11246 (dump_profile_report): Move body of implementation into...
11247 (pass_manager::dump_profile_report): ...here.
11248 (ipa_write_summaries_1): Update for moves of pass lists from
11249 being globals to fields of pass_manager.
11250 (ipa_write_optimization_summaries): Likewise.
11251 (ipa_read_summaries): Likewise.
11252 (ipa_read_optimization_summaries): Likewise.
11253 (execute_all_ipa_stmt_fixups): Likewise.
11254
11255 * statistics.c (statistics_fini): Update for moves of globals to
11256 fields of pass_manager.
11257
11258 * toplev.c (general_init): Replace call to
11259 init_optimization_passes with construction of the pass_manager
11260 instance.
11261
11262 * tree-pass.h (all_passes): Remove, in favor of a field of the
11263 same name within the new class pass_manager.
11264 (all_small_ipa_passes): Likewise.
11265 (all_lowering_passes): Likewise.
11266 (all_regular_ipa_passes): Likewise.
11267 (all_lto_gen_passes): Likewise.
11268 (all_late_ipa_passes): Likewise.
11269 (passes_by_id): Likewise.
11270 (passes_by_id_size): Likewise.
11271 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11272 the new class pass_manager.
11273 (get_pass_for_id): Remove.
11274
02cab1c1
RE		 112752013-07-30 Richard Earnshaw <rearnsha@arm.com>
11276
11277 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11278 configs.
11279
b2d3f886
RE		 112802013-07-30 Richard Earnshaw <rearnsha@arm.com>
11281
11282 * arm.md (mulhi3): New expand pattern.
11283
605e86fa
JH		 112842013-07-30 Jan Hubicka <jh@suse.cz>
11285 Martin Liska <marxin.liska@gmail.com>
11286
11287 * profile.c (compute_value_histograms): Do not ICE when
11288 there is mismatch only on some counters.
11289
14a87636
ZC		 112902013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11291
11292 PR rtl-optimization/57637
11293 * function.c (move_insn_for_shrink_wrap): Also check the
11294 GEN set of the LIVE problem for the liveness analysis
11295 if it exists, otherwise give up.
11296
0100cd3f
BS		 112972013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11298
11299 PR tree-optimization/57993
11300 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11301 replaced statement in the candidate table.
11302 (phi_add_costs): Return infinite cost when the hidden basis does
11303 not dominate all phis on which the candidate is dependent.
11304 (replace_one_candidate): Record replaced statement in the
11305 candidate table.
11306
f0d811f7
JR		 113072013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11308
11309 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11310 (ashlv2si3): New expander.
11311 (*ashlv2si3_i): New define_insn_and_split.
11312 * predicates.md (float_operation): Allow patterns with three
11313 basic sub-patterns.
11314
bba33211
JR		 11315 PR rtl-optimization/58021
11316 * mode-switching.c (create_pre_exit): Always split off preceding
11317 insns if we are not at the basic block head.
11318
ff3f3951
MR		 113192013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11320
11321 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11322 (UCLIBC_DYNAMIC_LINKER): New macro.
11323 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11324 `-mnan=2008'.
11325 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11326 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11327 `-mnan=2008'.
11328 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11329 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11330 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11331 for SF and DF modes. Use ieee_quad_format for TF mode.
11332 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11333 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11334 (mips_option_override): Handle `-mnan=legacy'.
11335 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11336 `-mabs=2008' and `-mnan=2008'.
11337 (OPTION_DEFAULT_SPECS): Add "nan" default.
11338 (ASM_SPEC): Handle `-mnan='.
11339 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11340 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11341 comment accordingly.
11342 (neg<mode>2): Likewise.
11343 * config/mips/mips.opt (mabs, mnan): New options.
11344 * doc/install.texi (Configuration): Document `--with-nan=' option.
11345 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11346 `-mnan=' options.
11347 (MIPS Options): Document them.
11348 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11349 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11350 * configure: Regenerate.
11351 * config.in: Regenerate.
11352
576e2f37
UB		 113532013-07-29 Uros Bizjak <ubizjak@gmail.com>
11354
11355 * config/i386/i386.md (float post-reload splitters): Do not check
11356 for subregs of SSE registers.
11357
113582013-07-29 Uros Bizjak <ubizjak@gmail.com>
11359 H.J. Lu <hongjiu.lu@intel.com>
11360
11361 PR target/57954
11362 PR target/57988
11363 * config/i386/i386.md (post-reload splitter
11364 to avoid partial SSE reg dependency stalls): New pattern.
11365
a71f0749
DV		 113662013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11367
11368 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11369
4c97f1cc
JR		 113702013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11371
11372 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11373 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11374
b0c733d8
CC		 113752013-07-26 Cary Coutant <ccoutant@google.com>
11376
11377 * dwarf2out.c (die_checksum_ordered): Don't include template
11378 instantiations in signature.
11379 (is_template_parameter): New function.
11380 (is_template_instantiation): New function.
11381 (generate_skeleton_bottom_up): Don't include template instantiations
11382 in type unit DIE.
11383 (generate_skeleton): Likewise.
11384 (break_out_comdat_types): Move recursive call to break out nested
11385 types earlier.
11386 (prune_unused_types_mark_generic_parms_dies): Call
11387 is_template_parameter.
11388
040d8a1c
IB		 113892013-07-26 Ian Bolton <ian.bolton@arm.com>
11390
11391 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11392 uses vector registers.
89fdc743 11393 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11394
85bd4ac6 113952013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11396 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 11397
11398 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11399 where appropriate.
85bd4ac6
KT		 11400 * config/arm/arm.md (movcond_addsi): New splitter.
11401
b6051207
SA		 114022013-07-25 Sterling Augustine <saugustine@google.com>
11403
11404 * dwarf2out.c (size_of_pubnames): Move code to...
11405 (include_pubname_in_output): ...here. New.
11406 (want_pubnames): Rearrange.
11407 (output_pubnames): Call include_pubname_in_output. Move assertion.
11408
41a7c215
CM		 114092013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11410
11411 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11412
114132013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11414
11415 PR target/38836
11416 * doc/extend.texi: Remove obsolete builtins. Fix
11417 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11418
49bde175
JH		 114192013-07-25 Jan Hubicka <jh@suse.cz>
11420
11421 * cgraph.c (release_function_body): Break out from ...
11422 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11423 and DECL_ARGUMENTS.
11424 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11425 old_tree in the map.
11426 (create_specialized_node): Update.
11427 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11428 into index.
41a7c215
CM		 11429 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11430 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 11431 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11432 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11433 DECL_RESULT.
11434
4870352d
KT		 114352013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11436
11437 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11438 addsi3_carryin_alt2_<optab>): Correct output template.
11439
1d6aee1c
KT		 114402013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11441
11442 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11443 Adjust for arm_restrict_it.
11444 Remove trailing whitespace.
11445
c7cafd75
MK		 114462013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11447
017d38f5
MK		 11448 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11449 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11450
c7cafd75
MK		 11451 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11452
00b7527b
VM		 114532013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11454
11455 PR rtl-optimization/57960
11456 * lra-constraints.c (process_alt_operands): Use the right mode
11457 when checking strict_low.
11458
4f63dfc6
JH		 114592013-07-25 Jan Hubicka <jh@suse.cz>
11460
11461 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11462 * cgraph.c (cgraph_remove_node): Do not release function body
11463 when in cgraph streaming.
11464 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 11465 in other partitions are not considered reachable; fix handling of
11466 clones.
4f63dfc6 11467
e6dcfa18
RR		 114682013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11469
11470 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11471
0b93d3b6
RR		 114722013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11473
11474 PR target/19599
11475 PR target/57731
bb80c2eb 11476 PR target/57837
e6dcfa18 11477 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11478 Ss with US. Adjust output for v5 and v4t.
41a7c215 11479 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11480
0b93d3b6
RR		 11481 * config/arm/constraints.md ("Ss"): Rename to US.
11482
690688b3
TG		 114832013-07-25 Terry Guo <terry.guo@arm.com>
11484
11485 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11486 shift_add/shift_sub0/shift_sub1 RTXs.
11487
a4ad093b 114882013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11489 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 11490
11491 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11492 (altivec_vpks<VI_char>ss): Likewise.
11493 (altivec_vpks<VI_char>us): Likewise.
11494 (altivec_vpku<VI_char>us): Likewise.
11495 (altivec_vpku<VI_char>um): Likewise.
11496
8c5005ce
DM		 114972013-07-24 David Malcolm <dmalcolm@redhat.com>
11498
11499 Introduce context class.
11500
11501 * Makefile.in (CONTEXT_H): New.
11502 (OBJS): Add context.o.
11503 (toplev.o): Add CONTEXT_H to dependencies.
11504 (context.o): New.
11505
41a7c215 11506 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 11507
11508 * context.c: New.
11509
11510 * context.h: New.
11511
ce4a9422
JR		 115122013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11513
11514 PR rtl-optimization/57968
11515 * mode-switching.c (create_pre_exit): Allow instructions that
11516 don't set a return register to need a non-exit mode.
11517
493f4c9e 115182013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11519 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11520
b5860fd3
WS		 11521 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11522 operands to vperm for little endian.
11523 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11524 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11525
73792b92 115262013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11527 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 11528
11529 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11530 two operands for little-endian.
11531
14c21302
SE		 115322013-07-23 Steve Ellcey <sellcey@mips.com>
11533
11534 * config/mips/mips.c (mips_case_values_threshold): New.
11535 (TARGET_CASE_VALUES_THRESHOLD): Define.
11536
a6056198 115372013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11538 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 11539
11540 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11541 selection of field for vector splat in little endian mode.
11542
dd7a40e1
MM		 115432013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11544
11545 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11546 expanders to rs6000.md.
11547 (ior<mode>3): Likewise.
11548 (and<mode>3): Likewise.
11549 (one_cmpl<mode>2): Likewise.
11550 (nor<mode>3): Likewise.
11551 (andc<mode>3): Likewise.
11552 (eqv<mode>3): Likewise.
11553 (nand<mode>3): Likewise.
11554 (orc<mode>3): Likewise.
11555
11556 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11557 declaration.
11558
11559 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11560 to split multi-word logical operations.
11561 (rs6000_split_logical_di): Likewise.
11562 (rs6000_split_logical): Likewise.
11563
11564 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11565 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11566 and allow TImode operations in 32-bit.
11567 (vsx_and<mode>3_64bit): Likewise.
11568 (vsx_ior<mode>3_32bit): Likewise.
11569 (vsx_ior<mode>3_64bit): Likewise.
11570 (vsx_xor<mode>3_32bit): Likewise.
11571 (vsx_xor<mode>3_64bit): Likewise.
11572 (vsx_one_cmpl<mode>2_32bit): Likewise.
11573 (vsx_one_cmpl<mode>2_64bit): Likewise.
11574 (vsx_nor<mode>3_32bit): Likewise.
11575 (vsx_nor<mode>3_64bit): Likewise.
11576 (vsx_andc<mode>3_32bit): Likewise.
11577 (vsx_andc<mode>3_64bit): Likewise.
11578 (vsx_eqv<mode>3_32bit): Likewise.
11579 (vsx_eqv<mode>3_64bit): Likewise.
11580 (vsx_nand<mode>3_32bit): Likewise.
11581 (vsx_nand<mode>3_64bit): Likewise.
11582 (vsx_orc<mode>3_32bit): Likewise.
11583 (vsx_orc<mode>3_64bit): Likewise.
11584
11585 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11586 logical types in GPRs.
11587
11588 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11589 logical insns to rs6000.md, and allow TImode operations in
11590 32-bit.
11591 (altivec_ior<mode>3): Likewise.
11592 (altivec_xor<mode>3): Likewise.
11593 (altivec_one_cmpl<mode>2): Likewise.
11594 (altivec_nor<mode>3): Likewise.
11595 (altivec_andc<mode>3): Likewise.
11596
11597 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11598 attributes for moving the 128-bit logical operations into
11599 rs6000.md.
11600 (BOOL_REGS_OUTPUT): Likewise.
11601 (BOOL_REGS_OP1): Likewise.
11602 (BOOL_REGS_OP2): Likewise.
11603 (BOOL_REGS_UNARY): Likewise.
11604 (BOOL_REGS_AND_CR0): Likewise.
11605 (one_cmpl<mode>2): Add support for DI logical operations on
11606 32-bit, splitting the operations to 32-bit.
11607 (anddi3): Likewise.
11608 (iordi3): Likewise.
11609 (xordi3): Likewise.
11610 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11611 changes to combine the 32/64-bit code, allow logical operations on
11612 TI mode in 32-bit, and to use similar match_operator patterns like
11613 scalar mode uses. Combine the Altivec and VSX code for logical
11614 operations, and move it here.
11615 (ior<mode>3, 128-bit types): Likewise.
11616 (xor<mode>3, 128-bit types): Likewise.
11617 (one_cmpl<mode>3, 128-bit types): Likewise.
11618 (nor<mode>3, 128-bit types): Likewise.
11619 (andc<mode>3, 128-bit types): Likewise.
11620 (eqv<mode>3, 128-bit types): Likewise.
11621 (nand<mode>3, 128-bit types): Likewise.
11622 (orc<mode>3, 128-bit types): Likewise.
11623 (and<mode>3_internal): Likewise.
11624 (bool<mode>3_internal): Likewise.
11625 (boolc<mode>3_internal1): Likewise.
11626 (boolc<mode>3_internal2): Likewise.
11627 (boolcc<mode>3_internal1): Likewise.
11628 (boolcc<mode>3_internal2): Likewise.
11629 (eqv<mode>3_internal1): Likewise.
11630 (eqv<mode>3_internal2): Likewise.
11631 (one_cmpl1<mode>3_internal): Likewise.
11632
12211b99 116332013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 11634
11635 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11636 Rename flag_stack_usage to flag_stack_usage_info.
11637
12211b99 116382013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 11639
a5f257fd
ME		 11640 * config/microblaze/sync.md: New file.
11641 * config/microblaze/microblaze.md: Include sync.md
11642 * config/microblaze/microblaze.c: Add print_operand 'y'.
11643 * config/microblaze/constraints.md: Add memory_contraint
11644 'Q' which is a single register.
450b0ebb 11645
49fba14f
EB		 116462013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11647
11648 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11649
c7b3b99f
PCC		 116502013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11651
11652 * reload.c (find_reloads): Exit loop once we find this operand
11653 cannot be reloaded somehow for this alternative.
11654
11655 * reload.c (find_reloads): Exit loop once we find a hard register.
11656
11657 * rtlanal.c (computed_jump_p): Exit loop once we find label
11658 reference is used.
11659
11660 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11661
11662 * cfgloopmanip.c (remove_path): Exit loop after setting
11663 irred_invalidated.
11664
11665 * gensupport.c (subst_dup): Avoid loop if code is not
11666 MATCH_DUP nor MATCH_OP_DUP.
11667
c8fbf1fa
NBJ		 116682013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11669
11670 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11671
9259db42
YZ		 116722013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11673
11674 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11675 true for SP_REGNUM if mode == ptr_mode.
11676 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11677 with value R0_REGNUM + 31.
11678
00edcfbe
YZ		 116792013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11680
11681 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11682 pad pointer-typed argument downward.
11683
43be9a95
YZ		 116842013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11685
11686 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11687 and __ILP32__ when the ILP32 model is in use.
11688
28514dda
YZ		 116892013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11690
11691 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11692 (aarch64_load_symref_appropriately): In the case of
11693 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11694 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11695 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11696 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11697 if 'mode' doesn't equal to 'ptr_mode'.
11698 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11699 'vcall_offset'; change to call aarch64_emit_move differently depending
11700 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11701 to calculate the upper bound of 'vcall_offset'.
11702 (aarch64_cannot_force_const_mem): Change to also return true if
11703 mode != ptr_mode.
11704 (aarch64_legitimize_reload_address): In the case of large
11705 displacements, add new local variable 'xmode' and an assertion
11706 based on it; change to use 'xmode' to generate the new rtx and
11707 reload.
11708 (aarch64_asm_trampoline_template): Change to generate the template
11709 differently depending on TARGET_ILP32 or not; change to use
11710 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11711 (aarch64_trampoline_size): Removed.
11712 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11713 and replace immediate literals with it. Change to use 'ptr_mode'
11714 instead of 'DImode' and call convert_memory_address if the mode
11715 of 'fnaddr' doesn't equal to 'ptr_mode'.
11716 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11717 to output symbol.
11718 (aarch64_elf_asm_destructor): Likewise.
11719 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11720 on TARGET_ILP32 instead of aarch64_trampoline_size.
11721 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11722 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11723 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11724 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11725 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11726 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11727 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11728 depending on the value of 'mode'.
11729 (add_losym_<mode>): New.
11730 (ldr_got_small_<mode>): New, based on ldr_got_small.
11731 (ldr_got_small): Remove.
11732 (ldr_got_small_sidi): New.
11733 * config/aarch64/iterators.md (P): New.
11734 (PTR): Change to 'ptr_mode' in the condition.
11735
17a819cb
YZ		 117362013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11737
11738 * config.gcc (aarch64*-*-*): Support --with-abi.
11739 (aarch64*-*-elf): Support --with-multilib-list.
11740 (aarch64*-*-linux*): Likewise.
11741 (supported_defaults): Add abi to aarch64*-*-*.
11742 * configure.ac: Mention AArch64 for --with-multilib-list.
11743 * configure: Re-generated.
11744 * config/aarch64/biarchilp32.h: New file.
11745 * config/aarch64/biarchlp64.h: New file.
11746 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11747 (ABI_SPEC): Ditto.
11748 (MULTILIB_DEFAULTS): Ditto.
11749 (DRIVER_SELF_SPECS): Ditto.
11750 (ASM_SPEC): Update to also substitute -mabi.
11751 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11752 file whose name depends on -mabi= and -mbig-endian.
11753 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11754 TARGET_ILP32.
11755 (POINTER_SIZE): New define.
11756 (POINTERS_EXTEND_UNSIGNED): Ditto.
11757 (enum aarch64_abi_type): New enumeration tag.
11758 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11759 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11760 (TARGET_ILP32): New define.
11761 * config/aarch64/aarch64.opt (mabi): New.
11762 (aarch64_abi): New.
11763 (ilp32, lp64): New values for -mabi.
11764 * config/aarch64/t-aarch64 (comma): New define.
11765 (MULTILIB_OPTIONS): Ditto.
11766 (MULTILIB_DIRNAMES): Ditto.
11767 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11768 * doc/invoke.texi: Document -mabi for AArch64.
11769
9f9cbdce
GJL		 117702013-07-23 Georg-Johann Lay <avr@gjlay.de>
11771
11772 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11773
7061977b 117742013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11775 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 11776
11777 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11778 endianness when selecting field to splat.
11779
75a41878
EC		 117802013-07-22 Eric Christopher <echristo@gmail.com>
11781
11782 * dwarf2out.c (die_odr_checksum): New function to use
11783 CHECKSUM_ macros and ULEB128 for DIE tag.
11784 (generate_type_signature): Use.
41a7c215 11785
38ae58ca
EB		 117862013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11787
11788 * config.gcc (sparc*-*-*): Accept leon3 processor.
11789 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11790 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11791 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11792 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11793 (mfix-ut699): Adjust comment.
11794 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11795 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11796 (CPP_CPU_SPEC): Likewise.
11797 (ASM_CPU_SPEC): Likewise.
11798 * config/sparc/sparc.c (leon3_cost): New constant.
11799 (sparc_option_override): Add leon3 support.
11800 (mem_ref): New function.
11801 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11802 (sparc_do_work_around_errata): Look into the instruction in the delay
11803 slot and adjust accordingly. Add fix for the data cache nullify issues
11804 of the UT699. Change insertion position for the NOP.
11805 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11806 (leon3_load): New reservation.
11807 (leon_store): Bump latency to 2.
11808 (grfpu): New automaton.
11809 (grfpu_alu): New unit.
11810 (grfpu_ds): Likewise.
11811 (leon_fp_alu): Adjust.
11812 (leon_fp_mult): Delete.
11813 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11814 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11815 * config/sparc/sparc.md (cpu): Add leon3.
11816 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11817 (swapsi): Likewise.
11818 (atomic_test_and_set): Likewise.
11819 (ldstub): Likewise.
11820
04dfc6df
JU		 118212013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11822
11823 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11824 default for R5900 targets.
11825 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11826 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11827 * config/mips/mips.c (mips_option_override): Report an error for
11828 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11829 for -march=r5900 -mhard-float.
11830
c7b3b99f 118312013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 11832
11833 * df-problems.c (can_move_insns_across): Exit loop once we
11834 find a non-fixed, non-global register.
11835
11836 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11837 setting can_throw.
11838
11839 * omega.c (omega_eliminate_red): Break after setting red_found.
11840 (omega_problem_has_red_equations): Similarly after setting found.
11841 (omega_query_variable): Similarly after setting coupled.
11842
9b6e6981
MP		 118432013-07-22 Marek Polacek <polacek@redhat.com>
11844
11845 * gimplify.c: Don't include gimple.h twice.
11846
5e5f7673
KT		 118472013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11848
11849 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11850 instead of TARGET_THUMB1.
11851 (Pz): New constraint.
11852 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11853 encodings.
11854 (compare_negsi_si): Likewise.
11855 (compare_addsi2_op0): Likewise.
11856 (compare_addsi2_op1): Likewise.
11857 (addsi3_carryin_<optab>): Likewise.
11858 (addsi3_carryin_alt2_<optab>): Likewise.
11859 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11860 for arm_restrict_it.
11861 (subsi3_carryin): Likewise.
11862 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11863 (minmax_arithsi): Disable for arm_restrict_it.
11864 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11865 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11866 (satsi_<SAT:code>_shift): Likewise.
11867 (arm_shiftsi3): Add alternative for 16-bit encoding.
11868 (arm32_movhf): Disable for arm_restrict_it.
11869 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11870 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11871
859abddd
SN		 118722013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11873
11874 * config/arm/arm.md (attribute "insn"): Delete.
11875 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11876 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11877 (not_shiftsi): Update for attribute change.
11878 (not_shiftsi_compare0): Likewise.
11879 (not_shiftsi_compare0_scratch): Likewise.
11880 (arm_one_cmplsi2): Likewise.
11881 (thumb1_one_cmplsi2): Likewise.
11882 (notsi_compare0): Likewise.
11883 (notsi_compare0_scratch): Likewise.
11884 (thumb1_movdi_insn): Likewise.
11885 (arm_movsi_insn): Likewise.
11886 (movhi_insn_arch4): Likewise.
11887 (movhi_bytes): Likewise.
11888 (arm_movqi_insn): Likewise.
11889 (thumb1_movqi_insn): Likewise.
11890 (arm32_movhf): Likewise.
11891 (thumb1_movhf): Likewise.
11892 (arm_movsf_soft_insn): Likewise.
11893 (thumb1_movsf_insn): Likewise.
11894 (thumb_movdf_insn): Likewise.
11895 (movsicc_insn): Likewise.
11896 (movsfcc_soft_insn): Likewise.
11897 (and_scc): Likewise.
11898 (cond_move): Likewise.
11899 (if_move_not): Likewise.
11900 (if_not_move): Likewise.
11901 (if_shift_move): Likewise.
11902 (if_move_shift): Likewise.
11903 (if_shift_shift): Likewise.
11904 (if_not_arith): Likewise.
11905 (if_arith_not): Likewise.
11906 (cond_move_not): Likewise.
11907 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11908 (neon_mov<mode>): Likewise.
11909 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11910 (thumb2_movsi_vfp): Likewise.
11911 (movsf_vfp): Likewise.
11912 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 11913 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11914 change.
859abddd
SN		 11915 (cortexa7_older_only): Likewise.
11916 (cortexa7_younger): Likewise.
11917 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11918 (1020alu_shift_op): Likewise.
11919 (1020alu_shift_reg_op): Likewise.
11920 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11921 (alu_shift_op): Likewise.
11922 (alu_shift_reg_op): Likewise.
11923 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11924 (11_alu_shift_op): Likewise.
11925 (11_alu_shift_reg_op): Likewise.
11926 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11927 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 11928 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11929 change.
859abddd
SN		 11930 (cortex_a15_alu_shift): Likewise.
11931 (cortex_a15_alu_shift_reg): Likewise.
11932 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11933 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 11934 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11935 change.
859abddd 11936 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 11937 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11938 change.
859abddd
SN		 11939 (cortex_a7_alu_reg): Likewise.
11940 (cortex_a7_alu_shift): Likewise.
11941 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11942 (cortex_a8_alu_shift): Likewise.
11943 (cortex_a8_alu_shift_reg): Likewise.
11944 (cortex_a8_mov): Likewise.
11945 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11946 (cortex_a9_dp_shift): Likewise.
11947 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11948 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11949 (cortex_r4_mov): Likewise.
11950 (cortex_r4_alu_shift): Likewise.
11951 (cortex_r4_alu_shift_reg): Likewise.
11952 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11953 (526_alu_shift_op): Likewise.
11954 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11955 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11956 (626te_alu_shift_op): Likewise.
11957 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11958 (726te_alu_op): Likewise.
11959 (726te_alu_shift_op): Likewise.
11960 (726te_alu_shift_reg_op): Likewise.
11961 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11962 (mp626_alu_shift_op): Likewise.
11963 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11964 (pj4_alu_e1_conds): Likewise.
11965 (pj4_alu): Likewise.
11966 (pj4_alu_conds): Likewise.
11967 (pj4_shift): Likewise.
11968 (pj4_shift_conds): Likewise.
11969 (pj4_alu_shift): Likewise.
11970 (pj4_alu_shift_conds): Likewise.
11971
95b97fac
KT		 119722013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11973
11974 * config/arm/predicates.md (shiftable_operator_strict_it):
11975 New predicate.
11976 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11977 Disable cond_exec version for arm_restrict_it.
11978 (thumb2_smaxsi3): Convert to generate cond_exec.
11979 (thumb2_sminsi3): Likewise.
11980 (thumb32_umaxsi3): Likewise.
11981 (thumb2_uminsi3): Likewise.
11982 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11983 (thumb2_neg_abssi2): Likewise.
11984 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11985 (thumb2_movsicc_insn): Adjust alternatives.
11986 (thumb2_mov_negscc): Disable for arm_restrict_it.
11987 (thumb2_mov_negscc_strict_it): New pattern.
11988 (thumb2_mov_notscc_strict_it): New pattern.
11989 (thumb2_mov_notscc): Disable for arm_restrict_it.
11990 (thumb2_ior_scc): Likewise.
11991 (thumb2_ior_scc_strict_it): New pattern.
11992 (thumb2_cond_move): Adjust for arm_restrict_it.
11993 (thumb2_cond_arith): Disable for arm_restrict_it.
11994 (thumb2_cond_arith_strict_it): New pattern.
11995 (thumb2_cond_sub): Adjust for arm_restrict_it.
11996 (thumb2_movcond): Likewise.
11997 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11998 (thumb2_zero_extendhisi2_v6): Likewise.
11999 (thumb2_zero_extendqisi2_v6): Likewise.
12000 (orsi_notsi_si): Likewise.
12001 (orsi_not_shiftsi_si): Likewise.
12002
d5a6ef82
GJL		 120032013-07-22 Georg-Johann Lay <avr@gjlay.de>
12004
12005 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12006 instruction sequence is 1 byte shorter.
12007
d9e71154
UB		 120082013-07-22 Uros Bizjak <ubizjak@gmail.com>
12009
12010 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12011 it is not needed after split.
12012
8c6381ec
IS		 120132013-07-20 Iain Sandoe <iain@codesourcery.com>
12014
12015 PR target/51784
12016 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12017 second label for nonlocal goto receivers. Don't output pic base labels
12018 unless we're producing PIC; mark that action unreachable().
12019 (ix86_save_reg): If the function contains a nonlocal label, save the
12020 PIC base reg.
12021 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 12022 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 12023 (update_pic_label_number_if_needed): New.
12024 (machopic_output_function_base_name): Adjust for nonlocal receiver
12025 case.
12026 (machopic_should_output_picbase_label): New.
12027 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12028 (nonlocal_goto_receiver): New insn and split.
12029
d05d0709
JG		 120302013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12031
12032 * config/aarch64/aarch64-builtins.c
12033 (aarch64_fold_builtin): Fold abs in all modes.
12034 * config/aarch64/aarch64-simd-builtins.def
12035 (abs): Enable for all modes.
12036 * config/aarch64/arm_neon.h
12037 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12038 (vabs_f64): Add missing intrinsic.
12039
ba590e63
IB		 120402013-07-19 Ian Bolton <ian.bolton@arm.com>
12041
12042 * config/aarch64/arm_neon.h (vabs_s64): New function
12043
02371798
GJL		 120442013-07-19 Georg-Johann Lay <avr@gjlay.de>
12045
12046 PR target/57516
12047 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12048 * config/avr/avr.md (adjust_len): Add `round'.
12049 * config/avr/avr-protos.h (avr_out_round): New prototype.
12050 (avr_out_plus): Add `out_label' argument.
12051 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12052 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12053 Handle the case where `insn' is just a pattern.
12054 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12055 (avr_out_round): New function.
12056 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12057
12211b99 120582013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 12059
12060 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12061 Add check for flag_stack_usage to handle -fstack-usage support
12062
12211b99 120632013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 12064
12065 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12066 interaction for new Power8 flags and VSX.
12067
df9dda2d
ST		 120682013-07-18 Sriraman Tallam <tmsriram@google.com>
12069
12070 PR middle-end/57698
12071 * tree-inline.c (expand_call_inline): Emit errors during
12072 early_inlining only if optimization is not turned on.
12073
0f9cdef4
DM		 120742013-07-18 David Malcolm <dmalcolm@redhat.com>
12075
12076 * passes.def: New.
12077
12078 * passes.c (init_optimization_passes): Move the construction of
12079 the pass hierarchy into a new passes.def file.
12080
12081 * Makefile.in (passes.o): Add dependency on passes.def.
12082
2efa4087
DM		 120832013-07-18 David Malcolm <dmalcolm@redhat.com>
12084
12085 * passes.c (init_optimization_passes): Introduce macros for
12086 constructing the tree of passes (INSERT_PASSES_AFTER,
12087 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12088 TERMINATE_PASS_LIST).
12089
bc404e1b
VM		 120902013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12091 Wei Mi <wmi@google.com>
12092
12093 PR rtl-optimization/57878
12094 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12095 top.
12096 (reload_pseudo_compare_func): Check nregs first for reload
12097 pseudos.
12098
d675843f
DM		 120992013-07-18 David Malcolm <dmalcolm@redhat.com>
12100
12101 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12102
c7b3b99f 121032013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 12104
12105 * read-rtl.c (validate_const_int): Once an invalid character is
12106 seen, quit the loop.
12107
12108 * gengtype.c (write_roots): Similarly once we find the "deletable"
12109 or "if_marked" option.
12110
52fceb44
SN		 121112013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12112
12113 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12114 "xtab" and "sat". Move value "clz" from here to ...
12115 (attriubte "type"): ... here.
12116 (satsi_<SAT:code>): Delete "insn" attribute.
12117 (satsi_<SAT:code>_shift): Likewise.
12118 (arm_zero_extendqisi2addsi): Likewise.
12119 (arm_extendqisi2addsi): Likewise.
12120 (clzsi2): Update for attribute changes.
12121 (rbitsi2): Likewise.
371e77e3
UB		 12122 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12123 attribute.
52fceb44
SN		 12124 (arm_usatsihi): Likewise.
12125 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12126
006bd006
SN		 121272013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12128
12129 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 12130 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12131 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 12132 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12133 in alphabetical order.
12134 (attribute "core_cycles"): Update for attribute changes.
12135 (arm_addsi3): Likewise.
12136 (addsi3_compare0): Likewise.
12137 (addsi3_compare0_scratch): Likewise.
12138 (addsi3_compare_op1): Likewise.
12139 (addsi3_compare_op2): Likewise.
12140 (compare_addsi2_op0): Likewise.
12141 (compare_addsi2_op1): Likewise.
12142 (addsi3_carryin_shift_<optab>): Likewise.
12143 (subsi3_carryin_shift): Likewise.
12144 (rsbsi3_carryin_shift): Likewise.
12145 (arm_subsi3_insn): Likewise.
12146 (subsi3_compare0): Likewise.
12147 (subsi3_compare): Likewise.
12148 (arm_andsi3_insn): Likewise.
12149 (thumb1_andsi3_insn): Likewise.
12150 (andsi3_compare0): Likewise.
12151 (andsi3_compare0_scratch): Likewise.
12152 (zeroextractsi_compare0_scratch
12153 (andsi_not_shiftsi_si): Likewise.
12154 (iorsi3_insn): Likewise.
12155 (iorsi3_compare0): Likewise.
12156 (iorsi3_compare0_scratch): Likewise.
12157 (arm_xorsi3): Likewise.
12158 (thumb1_xorsi3_insn): Likewise.
12159 (xorsi3_compare0): Likewise.
12160 (xorsi3_compare0_scratch): Likewise.
12161 (satsi_<SAT:code>_shift): Likewise.
12162 (rrx): Likewise.
12163 (arm_shiftsi3): Likewise.
12164 (shiftsi3_compare0): Likewise.
12165 (not_shiftsi): Likewise.
12166 (not_shiftsi_compare0): Likewise.
12167 (not_shiftsi_compare0_scratch): Likewise.
12168 (arm_one_cmplsi2): Likewise.
12169 (thumb_one_complsi2): Likewise.
12170 (notsi_compare0): Likewise.
12171 (notsi_compare0_scratch): Likewise.
12172 (thumb1_zero_extendhisi2): Likewise.
12173 (arm_zero_extendhisi2): Likewise.
12174 (arm_zero_extendhisi2_v6): Likewise.
12175 (arm_zero_extendhisi2addsi): Likewise.
12176 (thumb1_zero_extendqisi2): Likewise.
12177 (thumb1_zero_extendqisi2_v6): Likewise.
12178 (arm_zero_extendqisi2): Likewise.
12179 (arm_zero_extendqisi2_v6): Likewise.
12180 (arm_zero_extendqisi2addsi): Likewise.
12181 (thumb1_extendhisi2): Likewise.
12182 (arm_extendhisi2): Likewise.
12183 (arm_extendhisi2_v6): Likewise.
12184 (arm_extendqisi): Likewise.
12185 (arm_extendqisi_v6): Likewise.
12186 (arm_extendqisi2addsi): Likewise.
12187 (thumb1_extendqisi2): Likewise.
12188 (thumb1_movdi_insn): Likewise.
12189 (arm_movsi_insn): Likewise.
12190 (movsi_compare0): Likewise.
12191 (movhi_insn_arch4): Likewise.
12192 (movhi_bytes): Likewise.
12193 (arm_movqi_insn): Likewise.
12194 (thumb1_movqi_insn): Likewise.
12195 (arm32_movhf): Likewise.
12196 (thumb1_movhf): Likewise.
12197 (arm_movsf_soft_insn): Likewise.
12198 (thumb1_movsf_insn): Likewise.
12199 (movdf_soft_insn): Likewise.
12200 (thumb_movdf_insn): Likewise.
12201 (arm_cmpsi_insn): Likewise.
12202 (cmpsi_shiftsi): Likewise.
12203 (cmpsi_shiftsi_swp): Likewise.
12204 (arm_cmpsi_negshiftsi_si): Likewise.
12205 (movsicc_insn): Likewise.
12206 (movsfcc_soft_insn): Likewise.
12207 (arith_shiftsi): Likewise.
12208 (arith_shiftsi_compare0
12209 (arith_shiftsi_compare0_scratch
12210 (sub_shiftsi): Likewise.
12211 (sub_shiftsi_compare0
12212 (sub_shiftsi_compare0_scratch
12213 (and_scc): Likewise.
12214 (cond_move): Likewise.
12215 (if_plus_move): Likewise.
12216 (if_move_plus): Likewise.
12217 (if_move_not): Likewise.
12218 (if_not_move): Likewise.
12219 (if_shift_move): Likewise.
12220 (if_move_shift): Likewise.
12221 (if_shift_shift): Likewise.
12222 (if_not_arith): Likewise.
12223 (if_arith_not): Likewise.
12224 (cond_move_not): Likewise.
12225 (thumb1_ashlsi3): Set type attribute.
12226 (thumb1_ashrsi3): Likewise.
12227 (thumb1_lshrsi3): Likewise.
12228 (thumb1_rotrsi3): Likewise.
12229 (shiftsi3_compare0_scratch): Likewise.
12230 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12231 (neon_mov<mode>): Likewise.
371e77e3
UB		 12232 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12233 attribute changes.
006bd006
SN		 12234 (thumb2_movsi_insn): Likewise.
12235 (thumb2_cmpsi_neg_shiftsi): Likewise.
12236 (thumb2_extendqisi_v6): Likewise.
12237 (thumb2_zero_extendhisi2_v6): Likewise.
12238 (thumb2_zero_extendqisi2_v6): Likewise.
12239 (thumb2_shiftsi3_short): Likewise.
12240 (thumb2_addsi3_compare0_scratch): Likewise.
12241 (orsi_not_shiftsi_si): Likewise.
12242 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12243 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12244 changes.
12245 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12246 (1020alu_shift_op): Likewise.
12247 (1020alu_shift_reg_op): Likewise.
12248 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12249 (alu_shift_op): Likewise.
12250 (alu_shift_reg_op): Likewise.
12251 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12252 (11_alu_shift_op): Likewise.
12253 (11_alu_shift_reg_op): Likewise.
12254 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12255 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12256 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12257 changes.
006bd006
SN		 12258 (cortex_a15_alu_shift): Likewise.
12259 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 12260 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12261 changes.
006bd006
SN		 12262 (cortex_a5_alu_shift): Likewise.
12263 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12264 changes.
12265 (cortex_a53_alu_shift): Likewise.
12266 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12267 changes.
12268 (cortex_a7_alu_reg): Likewise.
12269 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 12270 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12271 changes.
006bd006
SN		 12272 (cortex_a8_alu_shift): Likewise.
12273 (cortex_a8_alu_shift_reg): Likewise.
12274 (cortex_a8_mov): Likewise.
12275 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12276 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 12277 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12278 changes.
12279 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12280 changes.
006bd006
SN		 12281 (cortex_r4_mov): Likewise.
12282 (cortex_r4_alu_shift): Likewise.
12283 (cortex_r4_alu_shift_reg): Likewise.
12284 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12285 (526_alu_shift_op): Likewise.
12286 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12287 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12288 (626te_alu_shift_op): Likewise.
12289 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12290 (726te_alu_op): Likewise.
12291 (726te_alu_shift_op): Likewise.
12292 (726te_alu_shift_reg_op): Likewise.
12293 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12294 (mp626_alu_shift_op): Likewise.
12295 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12296 (pj4_alu_e1_conds): Likewise.
12297 (pj4_alu): Likewise.
12298 (pj4_alu_conds): Likewise.
12299 (pj4_shift): Likewise.
12300 (pj4_shift_conds): Likewise.
12301 (pj4_alu_shift): Likewise.
12302 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 12303 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12304 changes.
006bd006
SN		 12305 (cortexa7_older_only): Likewise.
12306 (cortexa7_younger): Likewise.
12307
651df1b2
DM		 123082013-07-18 David Malcolm <dmalcolm@redhat.com>
12309
12310 * ipa-pure-const.c (generate_summary): Rename to...
12311 (pure_const_generate_summary): ... this.
12312
ea717bd7
IS		 123132013-07-17 Iain Sandoe <iain@codesourcery.com>
12314
12315 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12316
a1259a13
YR		 123172013-07-17 Yvan Roux <yvan.roux@linaro.org>
12318
12319 PR target/57909
12320 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12321 usage in HI mode.
12322
167f68ed
AK		 123232013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12324
12325 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12326 enabled without -march=zEC12.
12327 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12328 flags to be set.
12329
f900a982
MR		 123302013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12331
12332 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12333 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12334 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12335 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12336 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12337 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12338 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12339 ISA_HAS_FP4.
12340 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12341 and ISA_HAS_NMADD3_NMSUB3.
12342 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12343 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12344 (nmsub4<mode>, nmsub3<mode>): Likewise.
12345 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12346
d4ed27eb
MR		 123472013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12348
12349 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12350 TARGET_MIPS5400 checking.
12351
7a5add18
PB		 123522013-07-16 Jakub Jelinek <jakub@redhat.com>
12353 Peter Bergner <bergner@vnet.ibm.com>
12354
12355 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12356 registers in the comment.
12357 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12358 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12359 rather than FIRST_PSEUDO_REGISTERS.
12360
123612013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 12362
12363 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12364 enable extra ISA flags with TARGET_HTM.
12365
d2ab0929
MR		 123662013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12367
12368 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12369 Fix comment typos.
12370
839c74bc
CH		 123712013-07-15 Cong Hou <congh@google.com>
12372
12373 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12374 in compare function for sorting.
12375
7a5add18 123762013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 12377
12378 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12379 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12380 * config/rs6000/rs6000.opt: Add -mhtm option.
12381 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12382 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12383 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12384 __HTM__ if the HTM instructions are available.
12385 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12386 htm_spr_reg_operand): New define_predicates.
12387 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12388 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12389 Include htm.md.
12390 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12391 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12392 HTM builtin functions.
12393 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12394 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12395 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12396 (rs6000_builtin_mask_calculate): Likewise.
12397 (rs6000_option_override_internal): Likewise.
12398 (bdesc_htm): Add new HTM builtin support.
12399 (htm_spr_num): New function.
12400 (htm_spr_regno): Likewise.
12401 (rs6000_htm_spr_icode): Likewise.
12402 (htm_expand_builtin): Likewise.
12403 (htm_init_builtins): Likewise.
12404 (rs6000_expand_builtin): Add support for HTM builtin functions.
12405 (rs6000_init_builtins): Likewise.
371e77e3
UB		 12406 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12407 option.
0258b6e4
PB		 12408 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12409 (TARGET_HTM, MASK_HTM): Define macros.
12410 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12411 (FIXED_REGISTERS): Likewise.
12412 (CALL_USED_REGISTERS): Likewise.
12413 (CALL_REALLY_USED_REGISTERS): Likewise.
12414 (REG_ALLOC_ORDER): Likewise.
12415 (enum reg_class): Likewise.
12416 (REG_CLASS_NAMES): Likewise.
12417 (REG_CLASS_CONTENTS): Likewise.
12418 (REGISTER_NAMES): Likewise.
12419 (ADDITIONAL_REGISTER_NAMES): Likewise.
12420 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12421 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12422 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12423 * config/rs6000/htm.md: New file.
12424 * config/rs6000/htmintrin.h: New file.
12425 * config/rs6000/htmxlintrin.h: New file.
12426
87dd8ab0
MS		 124272013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12428
12429 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12430 Define SYMBOL_TINY_GOT, update comment.
12431 * config/aarch64/aarch64.c
12432 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12433 (aarch64_expand_mov_immediate): Likewise.
12434 (aarch64_print_operand): Likewise.
12435 (aarch64_classify_symbol): Likewise.
12436 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12437 (ldr_got_tiny): Define.
12438
52d676b6
TG		 124392013-07-13 Tobias Grosser <tobias@grosser.es>
12440
12441 PR tree-optimization/54094
12442 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12443 scheduling dimension for the parallelism check from the polyhedral
12444 information in the AST.
12445 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12446 in 2D + 1 form.
12447
286e8fc1
JM		 124482013-07-13 Jason Merrill <jason@redhat.com>
12449
12450 * print-tree.c (debug_vec_tree): Use debug_raw.
12451 (debug_raw (vec<tree, va_gc> &)): New.
12452 (debug_raw (vec<tree, va_gc> *)): New.
12453 * tree.h: Declare them.
12454
e43257e8
BC		 124552013-07-13 Bin Cheng <bin.cheng@arm.com>
12456
12457 * ifcvt.c (ifcvt_after_combine): New static variable.
12458 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12459 for size.
12460 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12461 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12462 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12463
da582d46
MR		 124642013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12465
12466 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12467
18b0ea8f
MM		 124682013-07-12 Michael Matz <matz@suse.de>
12469
12470 PR middle-end/55771
12471 * convert.c (convert_to_real): Reject non-float inner types.
12472
e4f0f84d
TB		 124732013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12474
12475 * config/aarch64/aarch64-protos.h
12476 (aarch64_simd_immediate_valid_for_move): Remove.
12477 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12478 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12479 cases.
12480 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12481
76500022
SE		 124822013-07-11 Steve Ellcey <sellcey@mips.com>
12483
12484 * config/mips/mips.c (mips_conditional_register_usage): Do not
12485 use t[0-7] registers in MIPS16 mode when optimizing for size.
12486
78039734
ST		 124872013-07-11 Sriraman Tallam <tmsriram@google.com>
12488
12489 * config/i386/i386.c (dispatch_function_versions): Fix array
12490 indexing of function_version_info to match actual_versions.
12491
7b24b675
TJ		 124922013-07-11 Teresa Johnson <tejohnson@google.com>
12493
12494 * vec.h (struct va_gc): Move release out-of-line.
12495 (va_gc::release): Call ggc_free on released vec.
12496
85d4cbb8
UW		 124972013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12498
12499 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12500 Require GOT register as additional operand in UNSPEC.
12501 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12502 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12503 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12504 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12505 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12506 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12507 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12508
52b3f9e2
GJL		 125092013-07-11 Georg-Johann Lay <avr@gjlay.de>
12510
12511 PR target/57631
12512 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12513 name seen by assembler/linker rather if available.
12514
3f3e266e
AS		 125152013-07-11 Andreas Schwab <schwab@suse.de>
12516
12517 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12518
21231ca6
VM		 125192013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12520
371e77e3 12521 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12522
d331d374
JM		 125232013-07-10 Joseph Myers <joseph@codesourcery.com>
12524
673c2f63
JM		 12525 * doc/tm.texi.in: Move hook documentation to ....
12526 * target.def: ... here.
12527
d331d374
JM		 12528 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12529 text on @hook line.
12530 * doc/tm.texi: Regenerate.
12531
dfeadaa0
PC		 125322013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12533
12534 PR c++/57869
12535 * doc/invoke.texi: Document -Wconditionally-supported.
12536
bb6f2bac
GJL		 125372013-07-10 Georg-Johann Lay <avr@gjlay.de>
12538
12539 PR target/57844
12540 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12541 of my_fp.
12542
5219b232
GJL		 125432013-07-10 Georg-Johann Lay <avr@gjlay.de>
12544
12545 PR target/57506
12546 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12547 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12548 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12549 Remove duplicate devices.
12550 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12551 * config/avr/t-multilib: Regenerate.
12552 * config/avr/avr-tables.opt: Regenerate.
12553 * doc/avr-mmcu.texi: Regenerate.
12554
283045c0
GJL		 125552013-07-10 Georg-Johann Lay <avr@gjlay.de>
12556
12557 PR target/56987
12558 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12559
ceb2d59f 125602013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 12561
ceb2d59f
GS		 12562 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12563 the cost of MULT when optimizing for size.
12564
055e0a99
JBG		 125652013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12566
12567 * config/cr16/cr16-protos.h: Don't include target.h.
12568
ef43e868
JM		 125692013-07-09 Joseph Myers <joseph@codesourcery.com>
12570
12571 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12572 adjust register size for TDmode and TFmode for VSX registers.
12573
4a283fdf
KT		 125742013-07-08 Kai Tietz <ktietz@redhat.com>
12575
12576 PR target/56892
12577 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12578 hook_bool_const_tree_true.
12579
2cf4c39e
AK		 125802013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12581
12582 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12583 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12584 * config/s390/s390.md: Define FPR*_REGNUM constants.
12585 Fix FPR2_REGNUM constant (18 -> 17).
12586 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12587 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12588 ("trunc<BFP:mode><DFP_ALL:mode>2")
12589 ("trunc<DFP_ALL:mode><BFP:mode>2")
12590 ("*extend<BFP:mode><DFP_ALL:mode>2")
12591 ("*extend<DFP_ALL:mode><BFP:mode>2")
12592 ("extend<BFP:mode><DFP_ALL:mode>2")
12593 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12594 FPR4_REGNUM.
12595
e6ac0270 125962013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 12597
6a2fa4b2 12598 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 12599
b89b22fc
AK		 126002013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12601
12602 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12603 and cfun_fpr_bit_p to cfun_fpr_save_p.
12604 (s390_frame_area, s390_register_info, s390_frame_info)
12605 (s390_emit_prologue, s390_emit_epilogue)
12606 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12607 register numbers.
12608 * config/s390/s390.h: Define *_REGNUM macros for floating point
12609 register numbers.
12610
07f398aa
EB		 126112013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12612
12613 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12614
75fefa91
PCC		 126152013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12616
12617 PR rtl-optimization/57786
12618 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12619 and break out of the loop when it is set to false.
12620
046f1eee
JJ		 126212013-07-08 Jakub Jelinek <jakub@redhat.com>
12622
8140c065
JJ		 12623 PR target/57819
12624 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12625 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12626 (const_int 63)) 0)).
12627 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12628 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12629 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12630
046f1eee
JJ		 12631 PR rtl-optimization/57829
12632 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12633 mask bits outside of mode are just sign-extension from mode to HWI.
12634
930b700b
MZ		 126352013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12636
12637 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12638 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12639 adjust_address instead of change_address to keep info about alignment.
12640 (emit_strmov): Remove.
12641 (emit_memmov): New function.
12642 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12643 (expand_movmem_epilogue): Likewise and return updated rtx for
12644 destination.
12645 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12646 destination and source.
12647 (decide_alignment): Refactor, handle vector_loop.
12648 (ix86_expand_movmem): Likewise.
12649 (ix86_expand_setmem): Likewise.
12650 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12651
7b9d1bd8
UB		 126522013-07-07 Uros Bizjak <ubizjak@gmail.com>
12653
12654 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12655 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12656
eabb5f48
UB		 126572013-07-06 Uros Bizjak <ubizjak@gmail.com>
12658
12659 * config/i386/sse.md (sse_movlhps): Change alternative 3
12660 of operand 2 to "m".
12661
126622013-07-06 Uros Bizjak <ubizjak@gmail.com>
12663
12664 PR target/57807
12665 * config/i386/sse.md (iptr): New mode attribute.
12666 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12667 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12668 (<sse>_vmmul<mode>3): Ditto.
12669 (<sse>_vmdiv<mode>3): Ditto.
12670 (sse_vmrcpv4sf2): Ditto.
12671 (<sse>_vmsqrt<mode>2): Ditto.
12672 (sse_vmrsqrtv4sf2): Ditto.
12673 (<sse>_vm<code><mode>3): Ditto.
12674 (avx_vmcmp<mode>3): Ditto.
12675 (<sse>_vmmaskcmp<mode>3): Ditto.
12676 (<sse>_comi): Ditto.
12677 (<sse>_ucomi): Ditto.
12678 (*xop_vmfrcz_<mode>): Ditto.
12679 (*fmai_fmadd_<mode>): Ditto.
12680 (*fmai_fmsub_<mode>): Ditto.
12681 (*fmai_fnmadd_<mode>): Ditto.
12682 (*fmai_fnmsub_<mode>): Ditto.
12683 (*fma4i_vmfmadd_<mode>): Ditto.
12684 (*fma4i_vmfmsub_<mode>): Ditto.
12685 (*fma4i_vmfnmadd_<mode>): Ditto.
12686 (*fma4i_vmfnmsub_<mode>): Ditto.
12687 (*xop_vmfrcz_<mode>): Ditto.
12688 (sse_cvtps2pi): Ditto.
12689 (sse_cvttps2pi): Ditto.
12690 (sse_cvtss2si): Ditto.
12691 (sse_cvtss2si_2): Ditto.
12692 (sse_cvtss2siq_2): Ditto.
12693 (sse_cvttss2si): Ditto.
12694 (sse_cvttss2siq): Ditto.
12695 (sse_cvtsd2si): Ditto.
12696 (sse_cvtsd2si_2): Ditto.
12697 (sse_cvtsd2siq_2): Ditto.
12698 (sse_cvttsd2si): Ditto.
12699 (sse_cvttsd2siq): Ditto.
12700 (sse_cvtsd2ss): Ditto.
12701 (sse_cvtss2sd): Ditto.
12702 (avx2_pbroadcast<mode>): Ditto.
12703 (avx2_pbroadcast<mode>_1): Ditto.
12704 (*avx_vperm_broadcast_v4sf): Ditto.
12705
12706 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12707 (sse_movlhps): Ditto.
12708 (sse_storehps): Ditto.
12709 (sse_loadhps): Ditto.
12710 (sse_storelps): Ditto.
12711 (sse_loadlps): Ditto.
12712 (*vec_concatv4sf): Ditto.
12713 (*vec_interleave_highv2df): Ditto.
12714 (*vec_interleave_lowv2df): Ditto.
12715 (*vec_extractv2df_1_sse): Ditto.
12716 (*vec_extractv2df_0_sse): Ditto.
12717 (sse2_storelpd): Ditto.
12718 (sse2_loadlpd): Ditto.
12719 (sse2_movsd): Ditto.
12720 (*vec_concatv4si): Ditto.
12721 (vec_concatv2di): Ditto.
12722
12723 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12724 for Intel asm dialect.
12725 (mmx_punpcklwd): Ditto.
12726 (mmx_punpckldq): Ditto.
12727
12728 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12729 for intel assembler dialect.
12730
1f6eac90
JJ		 127312013-07-06 Jakub Jelinek <jakub@redhat.com>
12732
12733 PR target/29776
12734 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12735 for BUILT_IN_C{LZ,LRSB}*.
12736 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12737 * tree-vrp.c (extract_range_basic): Handle
12738 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12739 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12740 fall thru to code calling set_value*.
12741 * builtins.c (expand_builtin): Remove *IMAX cases.
12742 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12743 if width is bigger than 2*HWI.
12744
2b778c9d
VM		 127452013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12746
12747 PR rtl-optimization/55342
12748 * lra-int.h (lra_subreg_reload_pseudos): New.
12749 * lra.c: Add undoing optional reloads to the block diagram.
12750 (lra_subreg_reload_pseudos): New.
12751 (lra_optional_reload_pseudos): Change comments.
12752 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12753 lra_optional_reload_pseudos after undo transformations.
12754 * lra-assigns.c (pseudo_prefix_title): New.
12755 (lra_setup_reg_renumber): Use it.
12756 (spill_for): Ditto. Check subreg reload pseudos too.
12757 (assign_by_spills): Consider subreg reload pseudos too.
12758 * lra-constraints.c (simplify_operand_subreg): Use
12759 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12760 (curr_insn_transform): Recognize and do optional reloads.
12761 (undo_optional_reloads): New.
12762 (lra_undo_inheritance): Call undo_optional_reloads.
12763
207156e4
TQ		 127642013-07-05 Thomas Quinot <quinot@adacore.com>
12765
12766 * tree-complex.c (expand_complex_operations_1): Fix typo.
12767
8990e73a
TB		 127682013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12769
12770 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12771 (tune_params): New member 'const vec_costs'.
12772 * config/aarch64/aarch64.c (generic_vector_cost): New.
12773 (generic_tunings): New member 'generic_vector_cost'.
12774 (aarch64_builtin_vectorization_cost): New.
12775 (aarch64_add_stmt_cost): New.
12776 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12777 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12778
f7bc421d
JJ		 127792013-07-03 Jakub Jelinek <jakub@redhat.com>
12780
12781 PR target/57777
12782 * config/i386/predicates.md (vsib_address_operand): Disallow
12783 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12784
e90d1568
HPN		 127852013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12786
12787 PR middle-end/55030
12788 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12789 expand_builtin_setjmp_receiver.
12790 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12791 with NULL for the label parameter.
12792 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12793 the frame-pointer. Adjust comments.
12794 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12795 only if LABEL is non-NULL.
12796
cbd7413b
YZ		 127972013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12798
12799 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12800 (ARM_ABI_AAPCS64): Ditto.
12801 (arm_abi): Ditto.
12802 (ARM_DEFAULT_ABI): Ditto.
12803
dec11868
JG		 128042013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12805
12806 * config/aarch64/aarch64-builtins.c
12807 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12808 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12809 (st1): Likewise.
12810 * config/aarch64/aarch64-simd.md
12811 (aarch64_ld1<VALL:mode>): New.
12812 (aarch64_st1<VALL:mode>): Likewise.
12813 * config/aarch64/arm_neon.h
12814 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12815
786fb9b6
ST		 128162013-07-02 Sriraman Tallam <tmsriram@google.com>
12817
0ba6aff3
UB		 12818 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12819 target ISA is AVX.
786fb9b6
ST		 12820 (ix86_option_override_internal):Turn on all -mavx target flags by
12821 default as they are dependent on AVX anyway.
12822
4b043553
CC		 128232013-07-02 Cary Coutant <ccoutant@google.com>
12824
12825 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12826 deterministic hash.
12827 (loc_checksum_ordered): Likewise.
12828 (hash_loc_operands): Remove inline keyword.
12829
c4e87a13
JJ		 128302013-07-02 Jakub Jelinek <jakub@redhat.com>
12831
12832 PR tree-optimization/57741
12833 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12834 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12835 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12836 Allow REAL_CST step_exprs if flag_associative_math.
12837 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12838
26366d28
IB		 128392013-07-02 Ian Bolton <ian.bolton@arm.com>
12840
0ba6aff3 12841 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 12842
2879bb2b
IB		 128432013-07-02 Ian Bolton <ian.bolton@arm.com>
12844
12845 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12846
afc5e8a6
KT		 128472013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12848
12849 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12850 encoding.
12851 (iorsi3_insn): Likewise.
12852 (arm_xorsi3): Likewise.
12853
31f8442b
SN		 128542013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12855
12856 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12857 to ...
12858 (attribute "type"): ... here, and prefix with "wmmx_".
12859 (attribute "core_cycles"): Update for attribute changes.
12860 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12861 (tbcstv4hi): Likewise.
12862 (tbcstv2si): Likewise.
12863 (iwmmxt_iordi3): Likewise.
12864 (iwmmxt_xordi3): Likewise.
12865 (iwmmxt_anddi3): Likewise.
12866 (iwmmxt_nanddi3): Likewise.
12867 (iwmmxt_arm_movdi): Likewise.
12868 (iwmmxt_movsi_insn): Likewise.
12869 (mov<mode>_internal): Likewise.
12870 (and<mode>3_iwmmxt): Likewise.
12871 (ior<mode>3_iwmmxt): Likewise.
12872 (xor<mode>3_iwmmxt): Likewise.
12873 (add<mode>3_iwmmxt): Likewise.
12874 (ssaddv8qi3): Likewise.
12875 (ssaddv4hi3): Likewise.
12876 (ssaddv2si3): Likewise.
12877 (usaddv8qi3): Likewise.
12878 (usaddv4hi3): Likewise.
12879 (usaddv2si3): Likewise.
12880 (sub<mode>3_iwmmxt): Likewise.
12881 (sssubv8qi3): Likewise.
12882 (sssubv4hi3): Likewise.
12883 (sssubv2si3): Likewise.
12884 (ussubv8qi3): Likewise.
12885 (ussubv4hi3): Likewise.
12886 (ussubv2si3): Likewise.
12887 (mulv4hi3_iwmmxt): Likewise.
12888 (smulv4hi3_highpart): Likewise.
12889 (umulv4hi3_highpart): Likewise.
12890 (iwmmxt_wmacs): Likewise.
12891 (iwmmxt_wmacsz): Likewise.
12892 (iwmmxt_wmacu): Likewise.
12893 (iwmmxt_wmacuz): Likewise.
12894 (iwmmxt_clrdi): Likewise.
12895 (iwmmxt_clrv8qi): Likewise.
12896 (iwmmxt_clr4hi): Likewise.
12897 (iwmmxt_clr2si): Likewise.
12898 (iwmmxt_uavgrndv8qi3): Likewise.
12899 (iwmmxt_uavgrndv4hi3): Likewise.
12900 (iwmmxt_uavgv8qi3): Likewise.
12901 (iwmmxt_uavgv4hi3): Likewise.
12902 (iwmmxt_tinsrb): Likewise.
12903 (iwmmxt_tinsrh): Likewise.
12904 (iwmmxt_tinsrw): Likewise.
12905 (iwmmxt_textrmub): Likewise.
12906 (iwmmxt_textrmsb): Likewise.
12907 (iwmmxt_textrmuh): Likewise.
12908 (iwmmxt_textrmsh): Likewise.
12909 (iwmmxt_textrmw): Likewise.
12910 (iwmxxt_wshufh): Likewise.
12911 (eqv8qi3): Likewise.
12912 (eqv4hi3): Likewise.
12913 (eqv2si3): Likewise.
12914 (gtuv8qi3): Likewise.
12915 (gtuv4hi3): Likewise.
12916 (gtuv2si3): Likewise.
12917 (gtv8qi3): Likewise.
12918 (gtv4hi3): Likewise.
12919 (gtv2si3): Likewise.
12920 (smax<mode>3_iwmmxt): Likewise.
12921 (umax<mode>3_iwmmxt): Likewise.
12922 (smin<mode>3_iwmmxt): Likewise.
12923 (umin<mode>3_iwmmxt): Likewise.
12924 (iwmmxt_wpackhss): Likewise.
12925 (iwmmxt_wpackwss): Likewise.
12926 (iwmmxt_wpackdss): Likewise.
12927 (iwmmxt_wpackhus): Likewise.
12928 (iwmmxt_wpackwus): Likewise.
12929 (iwmmxt_wpackdus): Likewise.
12930 (iwmmxt_wunpckihb): Likewise.
12931 (iwmmxt_wunpckihh): Likewise.
12932 (iwmmxt_wunpckihw): Likewise.
12933 (iwmmxt_wunpckilb): Likewise.
12934 (iwmmxt_wunpckilh): Likewise.
12935 (iwmmxt_wunpckilw): Likewise.
12936 (iwmmxt_wunpckehub): Likewise.
12937 (iwmmxt_wunpckehuh): Likewise.
12938 (iwmmxt_wunpckehuw): Likewise.
12939 (iwmmxt_wunpckehsb): Likewise.
12940 (iwmmxt_wunpckehsh): Likewise.
12941 (iwmmxt_wunpckehsw): Likewise.
12942 (iwmmxt_wunpckelub): Likewise.
12943 (iwmmxt_wunpckeluh): Likewise.
12944 (iwmmxt_wunpckeluw): Likewise.
12945 (iwmmxt_wunpckelsb): Likewise.
12946 (iwmmxt_wunpckelsh): Likewise.
12947 (iwmmxt_wunpckelsw): Likewise.
12948 (ror<mode>3): Likewise.
12949 (ashr<mode>3_iwmmxt): Likewise.
12950 (lshr<mode>3_iwmmxt): Likewise.
12951 (ashl<mode>3_iwmmxt): Likewise.
12952 (ror<mode>3_di): Likewise.
12953 (ashr<mode>3_di): Likewise.
12954 (lshr<mode>3_di): Likewise.
12955 (ashl<mode>3_di): Likewise.
12956 (iwmmxt_wmadds): Likewise.
12957 (iwmmxt_wmaddu): Likewise.
12958 (iwmmxt_tmia): Likewise.
12959 (iwmmxt_tmiaph): Likewise.
12960 (iwmmxt_tmiabb): Likewise.
12961 (iwmmxt_tmiatb): Likewise.
12962 (iwmmxt_tmiabt): Likewise.
12963 (iwmmxt_tmiatt): Likewise.
12964 (iwmmxt_tmovmskb): Likewise.
12965 (iwmmxt_tmovmskh): Likewise.
12966 (iwmmxt_tmovmskw): Likewise.
12967 (iwmmxt_waccb): Likewise.
12968 (iwmmxt_wacch): Likewise.
12969 (iwmmxt_waccw): Likewise.
12970 (iwmmxt_waligni): Likewise.
12971 (iwmmxt_walignr): Likewise.
12972 (iwmmxt_walignr0): Likewise.
12973 (iwmmxt_walignr1): Likewise.
12974 (iwmmxt_walignr2): Likewise.
12975 (iwmmxt_walignr3): Likewise.
12976 (iwmmxt_wsadb): Likewise.
12977 (iwmmxt_wsadh): Likewise.
12978 (iwmmxt_wsadbz): Likewise.
12979 (iwmmxt_wsadhz): Likewise.
12980 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12981 (iwmmxt_wabsdiffb): Likewise.
12982 (iwmmxt_wabsdiffh): Likewise.
12983 (iwmmxt_wabsdiffw): Likewise.
12984 (iwmmxt_waddsubhx): Likewise
12985 (iwmmxt_wsubaddhx): Likewise.
12986 (addc<mode>3): Likewise.
12987 (iwmmxt_avg4): Likewise.
12988 (iwmmxt_avg4r): Likewise.
12989 (iwmmxt_wmaddsx): Likewise.
12990 (iwmmxt_wmaddux): Likewise.
12991 (iwmmxt_wmaddsn): Likewise.
12992 (iwmmxt_wmaddun): Likewise.
12993 (iwmmxt_wmulwsm): Likewise.
12994 (iwmmxt_wmulwum): Likewise.
12995 (iwmmxt_wmulsmr): Likewise.
12996 (iwmmxt_wmulumr): Likewise.
12997 (iwmmxt_wmulwsmr): Likewise.
12998 (iwmmxt_wmulwumr): Likewise.
12999 (iwmmxt_wmulwl): Likewise.
13000 (iwmmxt_wqmulm): Likewise.
13001 (iwmmxt_wqmulwm): Likewise.
13002 (iwmmxt_wqmulmr): Likewise.
13003 (iwmmxt_wqmulwmr): Likewise.
13004 (iwmmxt_waddbhusm): Likewise.
13005 (iwmmxt_waddbhusl): Likewise.
13006 (iwmmxt_wqmiabb): Likewise.
13007 (iwmmxt_wqmiabt): Likewise.
13008 (iwmmxt_wqmiatb): Likewise.
13009 (iwmmxt_wqmiatt): Likewise.
13010 (iwmmxt_wqmiabbn): Likewise.
13011 (iwmmxt_wqmiabtn): Likewise.
13012 (iwmmxt_wqmiatbn): Likewise.
13013 (iwmmxt_wqmiattn): Likewise.
13014 (iwmmxt_wmiabb): Likewise.
13015 (iwmmxt_wmiabt): Likewise.
13016 (iwmmxt_wmiatb): Likewise.
13017 (iwmmxt_wmiatt): Likewise.
13018 (iwmmxt_wmiabbn): Likewise.
13019 (iwmmxt_wmiabtn): Likewise.
13020 (iwmmxt_wmiatbn): Likewise.
13021 (iwmmxt_wmiattn): Likewise.
13022 (iwmmxt_wmiawbb): Likewise.
13023 (iwmmxt_wmiawbt): Likewise.
13024 (iwmmxt_wmiawtb): Likewise.
13025 (iwmmxt_wmiawtt): Likewise.
13026 (iwmmxt_wmiawbbn): Likewise.
13027 (iwmmxt_wmiawbtn): Likewise.
13028 (iwmmxt_wmiawtbn): Likewise.
13029 (iwmmxt_wmiawttn): Likewise.
13030 (iwmmxt_wmerge): Likewise.
13031 (iwmmxt_tandc<mode>3): Likewise.
13032 (iwmmxt_torc<mode>3): Likewise.
13033 (iwmmxt_torvsc<mode>3): Likewise.
13034 (iwmmxt_textrc<mode>3): Likewise.
13035 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13036 (wmmxt_pack): Likewise.
13037 (wmmxt_mult_c1): Likewise.
13038 (wmmxt_mult_c2): Likewise.
13039 (wmmxt_alu_c1): Likewise.
13040 (wmmxt_alu_c2): Likewise.
13041 (wmmxt_alu_c3): Likewise.
13042 (wmmxt_transfer_c1): Likewise.
13043 (wmmxt_transfer_c2): Likewise.
13044 (wmmxt_transfer_c3): Likewise.
13045 (marvell_f_iwmmxt_wstr): Likewise.
13046 (marvell_f_iwmmxt_wldr): Likewise.
13047
bdb7bf8a
YZ		 130482013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13049
0ba6aff3 13050 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 13051
e6631d38
VM		 130522013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13053
13054 Revert:
13055 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13056 * lra-constraints.c (need_for_split_p): Check call used hard regs
13057 living through calls.
13058
13059 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13060 call used regs for call insn.
13061
53c1275b
JJ		 130622013-06-28 Jakub Jelinek <jakub@redhat.com>
13063
13064 PR target/57736
0ba6aff3
UB		 13065 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13066 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 13067
1141ed3f
BI		 130682013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13069
13070 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 13071
e32e4c4a
VM		 130722013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13073
13074 * lra-constraints.c (need_for_split_p): Check call used hard regs
13075 living through calls.
13076
f161bfd3
MM		 130772013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13078
13079 PR target/57744
13080 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13081 to tie with any other modes. Eliminate Altivec vector mode tests,
13082 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13083 code, to return 0 if testing MODE2 for a condition, if we've
13084 already tested MODE1 for the same condition.
13085
7eda14e1
MS		 130862013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13087
13088 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13089 layout.
13090
c822f852
MS		 130912013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13092
13093 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13094 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13095
da4f13a4
MS		 130962013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13097
0ba6aff3
UB		 13098 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13099 Define.
da4f13a4
MS		 13100 (aarch64_symbolic_constant_p): Remove.
13101 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13102 static. Fix line length and white space.
13103 (aarch64_symbolic_constant_p): Remove.
13104 * config/aarch64/predicates.md (aarch64_valid_symref):
13105 Use aarch64_classify_symbol_expression.
13106
956a95a5
KT		 131072013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13108
13109 * config/arm/constraints.md (Ts): New constraint.
13110 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13111 16-bit encodings.
13112 (compare_scc): Use "Ts" constraint for operand 0.
13113 (ior_scc_scc): Likewise.
13114 (and_scc_scc): Likewise.
13115 (and_scc_scc_nodom): Likewise.
13116 (ior_scc_scc_cmp): Likewise for operand 7.
13117 (and_scc_scc_cmp): Likewise.
13118 * config/arm/thumb2.md (thumb2_movsi_insn):
13119 Add alternatives for 16-bit encodings.
13120 (thumb2_movhi_insn): Likewise.
13121 (thumb2_movsicc_insn): Likewise.
13122 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13123 (thumb2_negscc): Use "Ts" constraint.
13124 Move mvn instruction outside cond_exec block.
13125 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13126 for 16-bit encodings.
13127
31dfce10
KT		 131282013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13129
13130 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13131 encoding.
13132 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13133 (mulsi3subsi): Likewise.
13134 (mulsidi3adddi): Likewise.
13135 (mulsidi3_v6): Likewise.
13136 (umulsidi3_v6): Likewise.
13137 (umulsidi3adddi_v6): Likewise.
13138 (smulsi3_highpart_v6): Likewise.
13139 (umulsi3_highpart_v6): Likewise.
13140 (mulhisi3tb): Likewise.
13141 (mulhisi3bt): Likewise.
13142 (mulhisi3tt): Likewise.
13143 (maddhisi4): Likewise.
13144 (maddhisi4tb): Likewise.
13145 (maddhisi4tt): Likewise.
13146 (maddhidi4): Likewise.
13147 (maddhidi4tb): Likewise.
13148 (maddhidi4tt): Likewise.
13149 (zeroextractsi_compare0_scratch): Likewise.
13150 (insv_zero): Likewise.
13151 (insv_t2): Likewise.
13152 (anddi_notzesidi_di): Likewise.
13153 (anddi_notsesidi_di): Likewise.
13154 (andsi_notsi_si): Likewise.
13155 (iordi_zesidi_di): Likewise.
13156 (xordi_zesidi_di): Likewise.
13157 (andsi_iorsi3_notsi): Likewise.
13158 (smax_0): Likewise.
13159 (smax_m1): Likewise.
13160 (smin_0): Likewise.
13161 (not_shiftsi): Likewise.
13162 (unaligned_loadsi): Likewise.
13163 (unaligned_loadhis): Likewise.
13164 (unaligned_loadhiu): Likewise.
13165 (unaligned_storesi): Likewise.
13166 (unaligned_storehi): Likewise.
13167 (extv_reg): Likewise.
13168 (extzv_t2): Likewise.
13169 (divsi3): Likewise.
13170 (udivsi3): Likewise.
13171 (arm_zero_extendhisi2addsi): Likewise.
13172 (arm_zero_extendqisi2addsi): Likewise.
13173 (compareqi_eq0): Likewise.
13174 (arm_extendhisi2_v6): Likewise.
13175 (arm_extendqisi2addsi): Likewise.
13176 (arm_movt): Likewise.
13177 (thumb2_ldrd): Likewise.
13178 (thumb2_ldrd_base): Likewise.
13179 (thumb2_ldrd_base_neg): Likewise.
13180 (thumb2_strd): Likewise.
13181 (thumb2_strd_base): Likewise.
13182 (thumb2_strd_base_neg): Likewise.
13183 (arm_negsi2): Add alternative for 16-bit encoding.
13184 (arm_one_cmplsi2): Likewise.
13185
1572e697
KT		 131862013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13187
13188 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13189 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13190 (movdfcc): Likewise.
13191 * config/arm/vfp.md (*thumb2_movsf_vfp):
13192 Disable predication for arm_restrict_it.
13193 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13194 (*thumb2_movdfcc_vfp): Likewise.
13195 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13196 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13197 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13198 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13199 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13200 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13201 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13202 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13203 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13204 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13205 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13206 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13207 Disable predication for arm_restrict_it.
13208
ebf8f0ea
KY		 132092013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13210
13211 * config/i386/bmiintrin.h (_bextr_u32): New.
13212 (_bextr_u64): Ditto.
13213
a290fcda
RS		 132142013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13215
13216 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13217 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13218 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13219 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13220 * config/mips/n32-elf.h: ...this new file.
13221
ea169996
MG		 132222013-06-27 Marc Glisse <marc.glisse@inria.fr>
13223
13224 PR target/57224
13225 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13226 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13227
12211b99 132282013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 13229
13230 * config/mips/mips-tables.opt: Regenerate.
13231 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 13232 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 13233 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 13234
2444244e
JJ		 132352013-06-27 Jakub Jelinek <jakub@redhat.com>
13236
cd06d2a2
JJ		 13237 PR target/57623
13238 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13239 constraints of operand 1 and 2.
13240
2444244e
JJ		 13241 PR target/57623
13242 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13243 to match RTL canonicalization. Swap predicates and
13244 constraints of operand 1 and 2.
13245
2f259720
VM		 132462013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13247
0ba6aff3
UB		 13248 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13249 Process OP_INOUT regs for splitting too.
2f259720 13250
fef4d2b3
JJ		 132512013-06-27 Jakub Jelinek <jakub@redhat.com>
13252
13253 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13254 decl before the loop, initialize to NULL.
13255 (vectorizable_load): Initialize ptr_incr to NULL.
13256
79ee9826
MJ		 132572013-06-27 Martin Jambor <mjambor@suse.cz>
13258
13259 PR lto/57208
13260 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13261 * ipa-ref.c (ipa_maybe_record_reference): New function.
13262 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13263 * ipa-cp.c (create_specialized_node): Record potential references from
13264 aggvals.
13265 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13266
e18b4a81
YZ		 132672013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13268
13269 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13270 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13271 'mode' to force_reg.
13272 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13273 (aarch64_expand_mov_immediate): Likewise.
13274
9c023bf0
YZ		 132752013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13276
13277 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13278 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13279
5a3fe9b6
AK		 132802013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13281
13282 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13283 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13284 (struct machine_function): Add tbegin_p.
13285 (s390_canonicalize_comparison): Fold CC mode compares to
13286 conditional jump if possible.
13287 (s390_emit_jump): Return the emitted jump.
13288 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13289 Handle CCRAWmode compares.
13290 (s390_option_override): Default to -mhtm if available.
13291 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13292 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13293 FPRs instead of df_regs_ever_live_p.
13294 (s390_optimize_nonescaping_tx): New function.
13295 (s390_init_frame_layout): Extend clobbered_regs array to cover
13296 FPRs as well.
13297 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13298 (s390_expand_tbegin): New function.
13299 (enum s390_builtin): New enum definition.
13300 (code_for_builtin): New array definition.
13301 (s390_init_builtins): New function.
13302 (s390_expand_builtin): New function.
13303 (TARGET_INIT_BUILTINS): Define.
13304 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 13305 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 13306 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13307 (s390_alc_comparison): Likewise.
13308 * config/s390/s390-modes.def: Add CCRAWmode.
13309 * config/s390/s390.h (processor_flags): Add PF_TX.
13310 (TARGET_CPU_HTM): Define macro.
13311 (TARGET_HTM): Define macro.
13312 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13313 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13314 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13315 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 13316 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 13317 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13318 ("*cc_to_int"): Move up.
13319 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13320 constants other than 0.
13321 ("*ccraw_to_int"): New insn and splitter definition.
13322 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13323 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13324 ("tx_assist"): New expander.
13325 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13326 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13327 * config/s390/s390.opt: Add -mhtm option.
13328 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13329 * config/s390/htmxlintrin.h: New file.
13330 * config/s390/htmintrin.h: New file.
13331 * config/s390/s390intrin.h: New file.
13332 * doc/extend.texi: Document htm builtins.
13333 * config.gcc: Add the new header files to extra_headers.
13334
9e216629
TS		 133352013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13336
13337 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13338 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13339
fd8c65e7 133402013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 13341 Pat Haugen <pthaugen@us.ibm.com>
13342 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 13343
13344 * config/rs6000/power8.md: New.
13345 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13346 setting for power8 entry.
13347 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13348 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13349 test for Power4/Power5 only.
13350 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13351 support.
13352 (force_new_group): Adjust comment.
13353 * config/rs6000/rs6000.md: Include power8.md.
13354
b24a2ce5
GY		 133552013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13356
13357 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13358 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13359 declaration.
13360 (tune_params): Update comment.
13361 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13362 (arm_max_conditional_execute): New function.
13363 (thumb2_final_prescan_insn): Use max_insn_skipped and
13364 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13365
f6fdeb62
JJ		 133662013-06-25 Jakub Jelinek <jakub@redhat.com>
13367
13368 PR tree-optimization/57705
13369 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13370 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13371 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 13372 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13373
c13bc3d9
MJ		 133742013-06-25 Martin Jambor <mjambor@suse.cz>
13375
13376 PR middle-end/57670
13377 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13378 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13379 calls in the dump.
13380 (ipa_note_param_call): Initialize member_ptr flag.
13381 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13382 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13383 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13384 (ipa_read_indirect_edge_info): Likewise.
13385
8a845901
RB		 133862013-06-25 Richard Biener <rguenther@suse.de>
13387
13388 PR middle-end/56977
13389 * passes.c (init_optimization_passes): Move pass_fold_builtins
13390 and pass_dce earlier with -Og.
13391
33be0bec
EB		 133922013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13393
13394 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13395 <BIT_FIELD_REF>: Remove trailing TAB.
13396 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13397 remove blank line.
13398
5fe8e757
MJ		 133992013-06-24 Martin Jambor <mjambor@suse.cz>
13400
13401 PR tree-optimization/57358
13402 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13403 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13404 (ipa_analyze_params_uses): Generate pessimistic info when true.
13405
44a60244
MJ		 134062013-06-24 Martin Jambor <mjambor@suse.cz>
13407
13408 PR tree-optimization/57539
13409 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13410 global.inlined_to of the new node to it. All callers changed.
13411 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13412 inlining_into, pass it to cgraph_clone_node.
13413 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13414 ipa_free_edge_args_substructures.
13415 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13416 rdesc linked list. Do not assert rdesc edges have inlined caller.
13417 Assert we have found an rdesc in the rdesc list.
13418
7c5848b8
RB		 134192013-06-24 Richard Biener <rguenther@suse.de>
13420
0ba6aff3 13421 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 13422 (pointer_set_lookup): Declare.
13423 (class pointer_map): New template class implementing a
13424 generic pointer to T map.
13425 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13426 pointer_map<T>::contains, pointer_map<T>::insert,
13427 pointer_map<T>::traverse): New functions.
13428 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13429 (pointer_set_lookup): New function.
13430 (pointer_set_contains): Use pointer_set_lookup.
13431 (pointer_set_insert): Likewise.
13432 (insert_aux): Remove.
13433 (struct pointer_map_t): Embed a pointer_set_t.
13434 (pointer_map_create): Adjust.
13435 (pointer_map_destroy): Likewise.
13436 (pointer_map_contains): Likewise.
13437 (pointer_map_insert): Likewise.
13438 (pointer_map_traverse): Likewise.
13439 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13440 pointer_map<unsigned> instead of a pointer_map_t.
13441 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13442 (streamer_tree_cache_lookup): Likewise.
13443 (streamer_tree_cache_create): Likewise.
13444 (streamer_tree_cache_delete): Likewise.
13445 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13446 pointer_map<unsigned> instead of a pointer_map_t.
13447 (lto_init_tree_ref_encoder): Adjust.
13448 (lto_destroy_tree_ref_encoder): Likewise.
13449 * lto-section-out.c (lto_output_decl_index): Likewise.
13450 (lto_record_function_out_decl_state): Likewise.
13451 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13452
3dbc97a9
RB		 134532013-06-24 Richard Biener <rguenther@suse.de>
13454
13455 PR tree-optimization/57488
13456 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13457
8345965a
AM		 134582013-06-24 Alan Modra <amodra@gmail.com>
13459
13460 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13461 (gen_easy_altivec_constant): Likewise.
13462 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13463 easy_vector_constant_msb): Likewise.
13464
b7df379f
JJ		 134652013-06-23 Jakub Jelinek <jakub@redhat.com>
13466
13467 PR target/57688
13468 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13469 add missing return true.
13470
12211b99 134712013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 13472
13473 PR target/52483
13474 * config/sh/predicates.md (general_extend_operand): Invoke
13475 general_movsrc_operand for memory operands.
13476 (general_movsrc_operand): Allow reg+reg addressing, do not use
13477 general_operand for memory operands.
13478
97db2bf7
ST		 134792013-06-23 Sriraman Tallam <tmsriram@google.com>
13480
13481 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13482 when current target options does not apply.
13483 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13484 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 13485 * config/i386/bmiintrin.h: Pass appropriate target
13486 attributes to header.
97db2bf7
ST		 13487 * config/i386/mmintrin.h: Ditto.
13488 * config/i386/nmmintrin.h: Ditto.
13489 * config/i386/avx2intrin.h: Ditto.
13490 * config/i386/fxsrintrin.h: Ditto.
13491 * config/i386/tbmintrin.h: Ditto.
13492 * config/i386/xsaveintrin.h: Ditto.
13493 * config/i386/f16cintrin.h: Ditto.
13494 * config/i386/xtestintrin.h: Ditto.
13495 * config/i386/xsaveoptintrin.h: Ditto.
13496 * config/i386/bmi2intrin.h: Ditto.
13497 * config/i386/lzcntintrin.h: Ditto.
13498 * config/i386/smmintrin.h: Ditto.
13499 * config/i386/wmmintrin.h: Ditto.
13500 * config/i386/x86intrin.h: Remove all header include guards.
13501 * config/i386/prfchwintrin.h: Ditto.
13502 * config/i386/pmmintrin.h: Ditto.
13503 * config/i386/tmmintrin.h: Ditto.
13504 * config/i386/xmmintrin.h: Ditto.
13505 * config/i386/popcntintrin.h: Ditto.
13506 * config/i386/rdseedintrin.h: Ditto.
13507 * config/i386/ammintrin.h: Ditto.
13508 * config/i386/emmintrin.h: Ditto.
13509 * config/i386/immintrin.h: Remove all header include guards.
13510 * config/i386/fma4intrin.h: Ditto.
13511 * config/i386/lwpintrin.h: Ditto.
13512 * config/i386/xopintrin.h: Ditto.
13513 * config/i386/ia32intrin.h: Ditto.
13514 * config/i386/avxintrin.h: Ditto.
13515 * config/i386/rtmintrin.h: Ditto.
13516 * config/i386/fmaintrin.h: Ditto.
13517 * config/i386/mm3dnow.h: Ditto.
13518
495e6879
ST		 135192013-06-22 Sriraman Tallam <tmsriram@google.com>
13520
13521 * common/config/i386/i386-common.c: Handle LZCNT.
13522
11e8f687
AK		 135232013-06-22 Andi Kleen <ak@linux.intel.com>
13524
13525 * doc/extend.texi: Use __atomic_store_n instead of
13526 __atomic_store in HLE example.
13527
12211b99 135282013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 13529
13530 * config/sh/sh.c: Remove <cstdlib> workaround.
13531
927f908b
AK		 135322013-06-21 Andi Kleen <ak@linux.intel.com>
13533
0ba6aff3 13534 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13535
cca410b6
AK		 135362013-06-21 Andi Kleen <ak@linux.intel.com>
13537
13538 * doc/extend.texi: Document that __atomic_clear and
13539 __atomic_test_and_set should only be used with bool.
13540
d84db1ed
JH		 135412013-06-20 Jan Hubicka <jh@suse.cz>
13542
13543 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13544 types_same_for_odr.
13545 * tree.c (decls_same_for_odr): New function.
13546 (same_for_edr): New function.
13547 (types_same_for_odr): New function.
13548 (get_binfo_at_offset): Use it.
13549 * tree.h (types_same_for_odr): Declare.
13550
12211b99 135512013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 13552 Jason Merrill <jason@redhat.com>
13553
13554 * system.h: Include <cstdlib> as well as <stdlib.h>.
13555
49b4ff63
UB		 135562013-06-20 Uros Bizjak <ubizjak@gmail.com>
13557
13558 PR target/57655
13559 * config/i386/i386.c (construct_container): Report error if
13560 long double is used with disabled x87 float returns.
13561
7a89ad00
JH		 135622013-06-20 Jan Hubicka <jh@suse.cz>
13563
13564 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13565
12211b99 135662013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 13567
13568 PR rtl-optimization/57425
13569 PR rtl-optimization/57569
13570 * alias.c (write_dependence_p): Remove parameters mem_mode and
13571 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13572 Changed all callers.
13573 (canon_anti_dependence): Get comments and semantics in sync.
13574 Add parameter mem_canonicalized. Changed all callers.
13575 * rtl.h (canon_anti_dependence): Update prototype.
13576
d16e9a99
RB		 135772013-06-20 Richard Biener <rguenther@suse.de>
13578
13579 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13580 byte case, inline streamer_read_uchar and defer section
13581 overrun check.
13582
c3a9b91b
RB		 135832013-06-20 Richard Biener <rguenther@suse.de>
13584
13585 PR tree-optimization/57584
13586 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13587 SSA names into the expanded expression that take part in
13588 abnormal coalescing.
13589
371e77e3 135902013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 13591
13592 * gcov.c (print_usage): Handle new option.
13593 (process_args): Ditto.
13594 (get_gcov_intermediate_filename): New function.
13595 (output_intermediate_file): New function.
13596 (output_gcov_file): New function
13597 (generate_results): Handle new option.
13598 (release_function): Relase demangled name.
13599 (read_graph_file): Handle demangled name.
13600 (output_lines): Ditto.
13601 * doc/gcov.texi: Document gcov intermediate format.
13602
11f5b71a
VM		 136032013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13604
13605 PR bootstrap/57604
13606 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13607 (lra_emit_add): Use the functions. Add comment about Y as an
13608 address segment.
13609
e1f9340b
DE		 136102013-06-19 David Edelsohn <dje.gcc@gmail.com>
13611
13612 PR driver/57652
13613 * collect2.c (collect_atexit): New.
13614 (collect_exit): Delete.
13615 (main): Register collect_atexit with atexit.
13616 (collect_wait): Change collect_exit to exit.
13617 (do_wait): Same.
13618 * collect2.h (collect_exit): Delete.
13619 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13620
e3f9e0ac
WM		 136212013-06-19 Wei Mi <wmi@google.com>
13622
13623 PR rtl-optimization/57518
13624 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13625 if regno is used in paradoxical subreg.
13626 (update_equiv_regs): Check pdx_subregs[regno] before
13627 set a reg to be equivalent with a mem.
13628
a09f1a76
MK		 136292013-06-19 Matthias Klose <doko@ubuntu.com>
13630
13631 PR driver/57651
13632 * file-find.h (find_a_file): Add a mode parameter.
13633 * file-find.c (find_a_file): Likewise.
13634 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13635 with X_OK for the executables.
13636 * collect2.c (main): Call find_a_file with X_OK.
13637
7d18b0ad
SE		 136382013-06-19 Steve Ellcey <sellcey@mips.com>
13639
13640 PR target/56942
13641 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13642 Use NEXT_INSN instead of next_real_insn.
13643
6a6dac52
JH		 136442013-06-19 Jan Hubicka <jh@suse.cz>
13645
13646 * cgraph.h (const_value_known_p): Replace by ...
13647 (ctor_for_folding): .. this one.
13648 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13649 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13650 * expr.c (expand_expr_real_1): Likewise.
13651 (string_constant): Likewise.
13652 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13653 * ipa.c (process_references): Likewise.
13654 (symtab_remove_unreachable_nodes): Likewise.
13655 * ipa-inline-analysis.c (param_change_prob): Likewise.
13656 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13657 (get_base_constructor): Likwise.
13658 * varpool.c (varpool_remove_node): Likewise.
13659 (varpool_remove_initializer): LIkewise.
13660 (dump_varpool_node): LIkwise.
13661 (const_value_known_p): Rewrite to ...
13662 (ctor_for_folding): ... this one.
13663
216c12ab
JJ		 136642013-06-19 Jakub Jelinek <jakub@redhat.com>
13665
13666 PR driver/57651
13667 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13668 PERSONALITY in $PATH derived prefixes.
13669
d53e2f99
JL		 136702013-06-19 Jeff Law <law@redhat.com>
13671
ecdbf306
JL		 13672 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13673 in comment.
13674
d53e2f99
JL		 13675 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13676 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13677 booleans.
13678
89e7fc04
SN		 136792013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13680
13681 * config/arm/vfp.md: Move VFP instruction classification documentation
13682 to ...
13683 * config/arm/arm.md: ... here. Update instruction classification
13684 documentation.
13685
5e7f6aaa
RE		 136862013-06-19 Richard Earnshaw <rearnsha@arm.com>
13687
13688 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13689 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13690 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13691
5fdc1e5d
SB		 136922013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13693
13694 PR target/57609
13695 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13696 with NEXT_INSN. Use tablejump_p to check for jump table data
13697 insns.
13698
e952afba
PC		 136992013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13700
13701 PR c++/56544
13702 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13703 that now in C++ the value is correct per the C++ standards.
13704
3fd0303a
RB		 137052013-06-19 Richard Biener <rguenther@suse.de>
13706
13707 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13708 for global context.
13709
b04e4b08
AK		 137102013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13711
13712 Revert:
13713 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13714
13715 PR target/57609
13716 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13717 with next_active_insn.
13718
bef8491a
ST		 137192013-06-18 Sriraman Tallam <tmsriram@google.com>
13720
13721 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13722 functions are inlined during failures to flag an error.
13723 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13724 in early inline pass.
22fc7d3d 13725
c21b257d
L		 137262013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13727
13728 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13729 in comments.
13730
33255ae3
JB		 137312013-06-18 Julian Brown <julian@codesourcery.com>
13732
13733 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13734 Permit virtual register pre-reload if !strict.
13735 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13736 change.
13737 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13738 prototype.
13739 * config/arm/neon.md (movmisalign<mode>): Use
13740 neon_perm_struct_or_reg_operand instead of
13741 neon_struct_or_register_operand.
13742 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13743 neon_permissive_struct_operand instead of neon_struct_operand.
13744 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13745 neon_vector_mem_operand.
13746 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13747 neon_vector_mem_operand.
13748 (neon_permissive_struct_operand): New.
13749 (neon_struct_or_register_operand): Rename to...
13750 (neon_perm_struct_or_reg_operand): This. Adjust call to
13751 neon_vector_mem_operand.
13752
d579fcda
RB		 137532013-06-18 Richard Biener <rguenther@suse.de>
13754
13755 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13756 * lto-streamer.h: Include pointer-set.h.
13757 (struct lto_decl_slot): Remove.
13758 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13759 Remove next_index entry.
13760 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13761 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13762 (lto_init_tree_ref_encoder): Adjust.
13763 (lto_destroy_tree_ref_encoder): Likewise.
13764 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13765 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13766 (lto_output_decl_index): Adjust.
13767 (lto_new_out_decl_state): Likewise.
13768 (lto_record_function_out_decl_state): Likewise.
13769 * lto-streamer-out.c (copy_function): Likewise.
13770
31ee20ba
RB		 137712013-06-18 Richard Biener <rguenther@suse.de>
13772
13773 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13774 * cgraphunit.c: Include cfgloop.h.
13775 (init_lowered_empty_function): Initialize the loop tree.
13776 (assemble_thunk): Insert new BBs into loops.
13777
ac0511f2
RB		 137782013-06-18 Richard Biener <rguenther@suse.de>
13779
13780 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13781 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13782 the map from cache entry to cache index optional.
13783 (streamer_tree_cache_replace_tree): Adjust accordingly.
13784 (streamer_tree_cache_append): Likewise.
13785 (streamer_tree_cache_delete): Likewise.
13786 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13787 streamer cache map from cache entry to cache index.
13788 * lto-streamer-out.c (create_output_block): Adjust.
13789
09485a08
SN		 137902013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13791
22fc7d3d
UB		 13792 * config/arm/arm.md (attribute "insn"): Move multiplication and
13793 division attributes to...
09485a08
SN		 13794 (attribute "type"): ... here. Remove mult.
13795 (attribute "mul32"): New attribute.
13796 (attribute "mul64"): Add umaal.
13797 (*arm_mulsi3): Update attributes.
13798 (*arm_mulsi3_v6): Likewise.
13799 (*thumb_mulsi3): Likewise.
13800 (*thumb_mulsi3_v6): Likewise.
13801 (*mulsi3_compare0): Likewise.
13802 (*mulsi3_compare0_v6): Likewise.
13803 (*mulsi_compare0_scratch): Likewise.
13804 (*mulsi_compare0_scratch_v6): Likewise.
13805 (*mulsi3addsi): Likewise.
13806 (*mulsi3addsi_v6): Likewise.
13807 (*mulsi3addsi_compare0): Likewise.
13808 (*mulsi3addsi_compare0_v6): Likewise.
13809 (*mulsi3addsi_compare0_scratch): Likewise.
13810 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13811 (*mulsi3subsi): Likewise.
13812 (*mulsidi3adddi): Likewise.
13813 (*mulsi3addsi_v6): Likewise.
13814 (*mulsidi3adddi_v6): Likewise.
13815 (*mulsidi3_nov6): Likewise.
13816 (*mulsidi3_v6): Likewise.
13817 (*umulsidi3_nov6): Likewise.
13818 (*umulsidi3_v6): Likewise.
13819 (*umulsidi3adddi): Likewise.
13820 (*umulsidi3adddi_v6): Likewise.
13821 (*smulsi3_highpart_nov6): Likewise.
13822 (*smulsi3_highpart_v6): Likewise.
13823 (*umulsi3_highpart_nov6): Likewise.
13824 (*umulsi3_highpart_v6): Likewise.
13825 (mulhisi3): Likewise.
13826 (*mulhisi3tb): Likewise.
13827 (*mulhisi3bt): Likewise.
13828 (*mulhisi3tt): Likewise.
13829 (maddhisi4): Likewise.
13830 (*maddhisi4tb): Likewise.
13831 (*maddhisi4tt): Likewise.
13832 (maddhidi4): Likewise.
13833 (*maddhidi4tb): Likewise.
13834 (*maddhidi4tt): Likewise.
13835 (divsi3): Likewise.
13836 (udivsi3): Likewise.
13837 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13838 (thumb2_mulsi_short_compare0): Likewise.
13839 (thumb2_mulsi_short_compare0_scratch): Likewise.
13840 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13841 (1020mult2): Likewise.
13842 (1020mult3): Likewise.
13843 (1020mult4): Likewise.
13844 (1020mult5): Likewise.
13845 (1020mult6): Likewise.
22fc7d3d
UB		 13846 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13847 change.
09485a08
SN		 13848 (cortex_a15_mult64): Likewise.
13849 (cortex_a15_sdiv): Likewise.
13850 (cortex_a15_udiv): Likewise.
13851 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13852 (mult2): Likewise.
13853 (mult3): Likewise.
13854 (mult4): Likewise.
13855 (mult5): Likewise.
13856 (mult6): Likewise.
13857 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13858 (pj4_ir_div): Likewise.
13859 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13860 (11_mult2): Likewise.
13861 (11_mult3): Likewise.
13862 (11_mult4): Likewise.
13863 (11_mult5): Likewise.
13864 (11_mult6): Likewise.
13865 (11_mult7): Likewise.
13866 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13867 (cortex_a8_mla): Likewise.
13868 (cortex_a8_mull): Likewise.
13869 (cortex_a8_smulwy): Likewise.
13870 (cortex_a8_smlald): Likewise.
13871 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13872 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13873 (cortex_r4_mul_3): Likewise.
13874 (cortex_r4_mla_4): Likewise.
13875 (cortex_r4_mla_3): Likewise.
13876 (cortex_r4_smlald): Likewise.
13877 (cortex_r4_mull): Likewise.
13878 (cortex_r4_sdiv): Likewise.
13879 (cortex_r4_udiv): Likewise.
13880 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13881 (cortex_a7_idiv): Likewise.
13882 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13883 (9_mult2): Likewise.
13884 (9_mult3): Likewise.
13885 (9_mult4): Likewise.
13886 (9_mult5): Likewise.
13887 (9_mult6): Likewise.
13888 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13889 (cortex_a53_sdiv): Likewise.
13890 (cortex_a53_udiv): Likewise.
13891 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13892 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13893 (mp626_mult2): Likewise.
13894 (mp626_mult3): Likewise.
13895 (mp626_mult4): Likewise.
13896 * config/arm/fa526.md (526_mult1): Update attribute change.
13897 (526_mult2): Likewise.
13898 * config/arm/arm-generic.md (mult): Update attribute change.
13899 (mult_ldsched_strongarm): Likewise.
13900 (mult_ldsched): Likewise.
13901 (multi_cycle): Likewise.
13902 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13903 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13904 (606te_mult2): Likewise.
13905 (606te_mult3): Likewise.
13906 (606te_mult4): Likewise.
13907 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13908 (cortex_a9_mac16): Likewise.
13909 (cortex_a9_multiply): Likewise.
13910 (cortex_a9_mac): Likewise.
13911 (cortex_a9_multiply_long): Likewise.
13912 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13913 (626te_mult2): Likewise.
13914 (626te_mult3): Likewise.
13915 (626te_mult4): Likewise.
13916
8951f345
RB		 139172013-06-18 Richard Biener <rguenther@suse.de>
13918
13919 PR lto/57334
13920 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13921
d8101d05
AK		 139222013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13923
13924 PR target/57609
13925 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13926 with next_active_insn.
13927
69eff9da
AM		 139282013-06-18 Alan Modra <amodra@gmail.com>
13929
13930 * config/rs6000/rs6000.h (enum data_align): New.
13931 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13932 (DATA_ABI_ALIGNMENT): Define.
13933 (CONSTANT_ALIGNMENT): Correct comment.
13934 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13935 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13936
8ed00d76
DM		 139372013-06-17 David Malcolm <dmalcolm@redhat.com>
13938
13939 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13940 ATTRIBUTE_UNUSED marking.
13941
3d33d06b
SN		 139422013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13943
13944 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13945 alternative and update.
13946 (aarch64_dup_lanedi): Delete.
13947 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13948 * config/aarch64/aarch64-simd-builtins.def: Update.
13949
ee03e71d
RB		 139502013-06-17 Richard Biener <rguenther@suse.de>
13951
13952 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13953 (lto_input_scc): Declare.
13954 (lto_input_tree_1): Likewise.
13955 (struct lto_stats_d): Add num_tree_bodies_output and
13956 num_pickle_refs_output.
13957 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13958 (lto_read_tree_1): Split out from ...
13959 (lto_read_tree): ... this.
13960 (lto_input_scc): New function.
13961 (lto_input_tree_1): Split out from ...
13962 (lto_input_tree): ... this. Handle LTO_tree_scc.
13963 (lto_data_in_create): Create the streamer cache without hashes.
13964 * lto-streamer-out.c (create_output_block): Create the streamer
13965 cache with hashes when not doing WPA.
13966 (lto_write_tree_1): Split out from ...
13967 (lto_write_tree): ... this.
13968 (get_symbol_initial_value): New function.
13969 (lto_output_tree_1): Split out from ...
13970 (lto_output_tree): ... this. Write trees as series of SCCs
13971 using a DFS walk via DFS_write_tree.
13972 (struct sccs, struct scc_entry): New types.
13973 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13974 (DFS_write_tree_body): New function.
13975 (DFS_write_tree): Likewise.
13976 (hash_tree): Likewise.
13977 (scc_entry_compare): Likewise.
13978 (hash_scc): Likewise.
13979 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13980 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13981 TREE_CHAIN as regular reference.
13982 (streamer_read_integer_cst): Remove.
13983 (streamer_get_pickled_tree): Adjust.
13984 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13985 of DECL_EXTERNALs in BLOCK_VARS for now.
13986 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13987 reference.
13988 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13989 Add hash value argument and record that if hashes are recorded
13990 in the cache.
13991 (streamer_tree_cache_insert_1): Adjust.
13992 (streamer_tree_cache_insert): Likewise.
13993 (streamer_tree_cache_insert_at): Rename to ...
13994 (streamer_tree_cache_replace_tree): ... this and adjust.
13995 (streamer_tree_cache_append): Adjust.
13996 (record_common_node): Likewise.
13997 (streamer_tree_cache_create): Add argument whether to
13998 record hash values together with trees.
13999 (streamer_tree_cache_delete): Adjust.
14000 * tree-streamer.h (struct streamer_tree_cache_d): Add
14001 vector of hashes.
14002 (streamer_read_integer_cst): Remove.
14003 (streamer_tree_cache_insert): Adjust.
14004 (streamer_tree_cache_append): Likewise.
14005 (streamer_tree_cache_insert_at): Rename to ...
14006 (streamer_tree_cache_replace_tree): ... this and adjust.
14007 (streamer_tree_cache_create): Add argument whether to record hashes.
14008 (streamer_tree_cache_get): Rename to ...
14009 (streamer_tree_cache_get_tree): ... this.
14010 (streamer_tree_cache_get_hash): New function.
14011 * tree.c (cache_integer_cst): New function.
14012 * tree.h (cache_integer_cst): Declare.
14013 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14014 * lto-symtab.c (lto_varpool_replace_node): Only release
14015 DECL_INITIAL of non-prevailing decls.
14016 * varpool.c (varpool_remove_initializer): Do not release
14017 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14018
0de86a92
JU		 140192013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14020
14021 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14022 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14023 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14024 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14025 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14026 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14027 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14028 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14029 instead of TARGET_64BIT.
14030 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14031 Require ISA_HAS_<D>DIV.
14032
74240413
RS		 140332013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14034
14035 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14036 (mips*-*-linux*): Move default with_llsc setting to where other
14037 defaults are set.
14038 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14039 with_arch block.
14040 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14041 Likewise. Remove default with_tune setting. Move default float
14042 setting to its own block. Handle with_llsc in the same block as above.
14043
12211b99 140442013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 14045
14046 PR rtl-optimization/57425
14047 PR rtl-optimization/57569
299a5f6a 14048 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 14049 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14050 Changed all callers.
14051 (canon_anti_dependence): New function.
14052 * cse.c (check_dependence): Use canon_anti_dependence.
14053 * cselib.c (cselib_invalidate_mem): Likewise.
14054 * rtl.h (canon_anti_dependence): Declare.
14055
b259d352
JU		 140562013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14057
14058 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14059 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14060 ".set mips3" for 64-bit targets.
14061
4de09b85
DC		 140622013-06-15 Dehao Chen <dehao@google.com>
14063
14064 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14065 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14066 (gimple_check_call_args): Likewise.
14067 * value-prof.c (check_ic_target): Likewise.
14068 * ipa-inline.c (early_inliner): Likewise.
14069 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14070 * cgraph.c (cgraph_create_edge_1): Likewise.
14071 (cgraph_make_edge_direct): Likewise.
14072
d1aa4a76
MM		 140732013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14074
14075 PR target/57615
14076 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14077 rs6000_output_move_128bit to handle emitting quad memory
14078 operations. Set attribute length to 8 bytes.
14079
12211b99 140802013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 14081
14082 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14083 New pattern.
14084 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14085 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14086 (aarch64_<su>mlsl<mode>): Likewise.
14087
f5273574
MS		 140882013-06-14 Mike Stump <mikestump@comcast.net>
14089
14090 * Makefile.in (TARGET_H): Add insn-codes.h.
14091
ef7e7ab2
AM		 140922013-06-14 Alan Modra <amodra@gmail.com>
14093
14094 PR middle-end/57134
14095 PR middle-end/57586
14096 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14097 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14098 bitfield expansion when EXPAND_MEMORY.
14099 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14100
135faab6
MM		 141012013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14102
14103 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14104 test for clearing quad memory on 32-bit later.
14105
948a1fd9
MG		 141062013-06-13 Marc Glisse <marc.glisse@inria.fr>
14107
14108 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14109 (fold_negate_expr): Likewise.
14110 (fold_real_zero_addition_p): Handle vectors.
14111 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14112
0597fb9c
AM		 141132013-06-14 Alan Modra <amodra@gmail.com>
14114
14115 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14116
e5b962d0
JH		 141172013-06-13 Jan Hubicka <jh@suse.cz>
14118
14119 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14120 Local comdats are not externally visible.
14121 * symtab.c (dump_symtab_base): Dump externally visible.
14122 (verify_symtab_base): Verify back links in the symtab hash.
14123
a8355e51
BC		 141242013-06-13 Bin Cheng <bin.cheng@arm.com>
14125
14126 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14127 CONVERT_EXPR as equal nodes.
14128
8f7e6e33
BC		 141292013-06-13 Bin Cheng <bin.cheng@arm.com>
14130
22fc7d3d 14131 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 14132
5d418483
MG		 141332013-06-13 Marc Glisse <marc.glisse@inria.fr>
14134
14135 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14136 Generalize to complex and vector.
14137 * tree.c (build_all_ones_cst): New function.
14138 * tree.h (build_all_ones_cst): Declare it.
14139
50751417
AM		 141402013-06-13 Alan Modra <amodra@gmail.com>
14141
14142 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14143 * config/rs6000/rs6000.md (signbittf2): New insn.
14144 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14145 (abstf2_internal, cmptf_internal2): Likewise.
14146 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14147
4b02c962 141482013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14149 Pat Haugen <pthaugen@us.ibm.com>
14150 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 14151
14152 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14153 power8 byte, half-word, and quad-word atomic instructions.
14154 (emit_store_conditional): Likewise.
14155 (rs6000_expand_atomic_compare_and_swap): Likewise.
14156 (rs6000_expand_atomic_op): Likewise.
14157
14158 * config/rs6000/sync.md (larx): Add new modes for power8.
14159 (stcx): Likewise.
14160 (AINT): New mode iterator to include TImode as well as normal
14161 integer modes on power8.
14162 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14163 that VSX registers are not considered. Use AINT mode iterator
14164 instead of INT1 to allow inclusion of quad word atomic operations
14165 on power8.
14166 (load_locked<mode>): Likewise.
14167 (store_conditional<mode>): Likewise.
14168 (atomic_compare_and_swap<mode>): Likewise.
14169 (atomic_exchange<mode>): Likewise.
14170 (atomic_nand<mode>): Likewise.
14171 (atomic_fetch_<fetchop_name><mode>): Likewise.
14172 (atomic_nand_fetch<mode>): Likewise.
14173 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14174 each type.
14175 (ATOMIC): On power8, add QImode, HImode modes.
14176 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14177 modes that promote to SImode.
14178 (load_lockedti): Convert TImode arguments to PTImode, so that we
14179 get a guaranteed even/odd register pair.
14180 (load_lockedpti): Likewise.
14181 (store_conditionalti): Likewise.
14182 (store_conditionalpti): Likewise.
14183
14184 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14185 atomic load/store instructions.
14186 (HSI): Likewise.
14187
710b6cc1
RS		 141882013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14189
14190 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14191 loads.
14192 (insn_count): New attribute, with most cases extracted from...
14193 (length): ...here. Redefine most cases in terms of insn_count.
14194 (single_insn): Delete.
14195 (can_delay): Use insn_count to check for single instructions.
14196 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14197 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14198 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14199 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14200 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14201 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14202 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14203 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14204 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14205 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14206 rather than "length".
14207 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14208 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14209 Use "insn_count" rather than "length".
14210 * config/mips/mips-dsp.md
14211 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14212 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14213 length attributes.
14214
1951f101
MG		 142152013-06-12 Marc Glisse <marc.glisse@inria.fr>
14216
14217 PR tree-optimization/57361
14218 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14219
8b033a8a
SN		 142202013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14221
22fc7d3d
UB		 14222 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14223 to split.
8b033a8a
SN		 14224 (aarch64_simd_combine<mode>): New instruction expansion.
14225 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14226 function prototype.
14227 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14228 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14229
ca0f62a8
JH		 142302013-06-12 Jan Hubicka <jh@suse.cz>
14231
14232 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14233 decl has when in streaming stage.
14234 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14235 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14236
12211b99 142372013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 14238
14239 PR target/57578
14240 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14241
bfdeda2c
JJ		 142422013-06-12 Jakub Jelinek <jakub@redhat.com>
14243
14244 PR tree-optimization/57537
14245 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14246 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14247
a4fa02d1
RB		 142482013-06-12 Richard Biener <rguenther@suse.de>
14249
14250 * data-streamer.h (streamer_write_char_stream): CSE
14251 obs->current_pointer.
14252 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14253 streamer_write_char_stream manually and optimize the resulting loop.
14254 (streamer_write_hwi_stream): Likewise.
14255
bbf9ad07
JH		 142562013-06-12 Jan Hubicka <jh@suse.cz>
14257
14258 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14259 * cgraph.h (varpool_create_empty_node): Declare.
14260 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14261 duplicated nodes.
14262 * symtab.c (symtab_unregister_node): Be lax about missin entries
14263 in node hash.
14264 (symtab_get_node): Update comment.
14265 * varpool.c (varpool_create_empty_node): Break out from ...
14266 (varpool_node_for_decl): ... here.
14267 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14268
a53f90ad
EB		 142692013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14270
14271 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14272 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14273 part. Use straight-line flow at the end.
14274 <COMPONENT_REF>: Remove superfluous else.
14275 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14276
946f9306
JJ		 142772013-06-12 Jakub Jelinek <jakub@redhat.com>
14278
14279 PR target/56564
14280 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14281 target hook even for !TREE_PUBLIC decls. If no resolution info
14282 is available, return false for common and external decls.
14283
a7fca6f0
KP		 142842013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14285
14286 * config/rl78/constraints.md (U): New constraint.
14287 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14288 valloc attribute.
14289
c7d8f446
MM		 142902013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14291
14292 PR target/57589
14293 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14294 to allow returning address to AT_PLATFORM name.
14295
08346abd
JH		 142962013-06-11 Jan Hubicka <jh@suse.cz>
14297
14298 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14299 * cgraph.h (symtab_node_base): Add weakref flag.
14300 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14301 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14302 (output_weakrefs): Use weakref flag.
14303 * fold-const.c (simple_operand_p): Handle WEAK.
14304 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14305 * ipa.c (varpool_externally_visible_p): Drop weakref.
14306 (function_and_variable_visibility): Update comment; fix weakref
14307 sanity checks; do not clear DECL_WEAK on them.
14308 * lto-cgraph.c (lto_output_node): update.
14309 (lto_output_varpool_node): Update.
14310 (input_overwrite_node): Update.
14311 (input_node): Update.
14312 (input_varpool_node): Update.
14313 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14314 (lto_symtab_merge_symbols): Add sanity check.
14315 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14316 * passes.c (rest_of_decl_compilation): Set static flag, too.
14317 * symtab.c (dump_symtab_base): Dump weakref.
14318 (verify_symtab_base): Sanity check weakrefs.
14319 (symtab_make_decl_local): Remove duplicated code.
14320 (symtab_alias_ultimate_target): Simplify.
14321 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 14322
2e6a2f88
TV		 143232013-06-11 Tom de Vries <tom@codesourcery.com>
14324
14325 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14326 sequence_vect == NULL.
14327
d9af3652
DD		 143282013-06-11 DJ Delorie <dj@redhat.com>
14329
14330 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14331 (rl78_unwind_word_mode): New.
14332
ffa66012
DM		 143332013-06-11 David Malcolm <dmalcolm@redhat.com>
14334
14335 * final.c (debug_prefix_maps): Make static.
14336
783f2d35
DM		 143372013-06-11 David Malcolm <dmalcolm@redhat.com>
14338
14339 * function.c (initial_trampoline): Remove stray copy.
14340
7af26def
SN		 143412013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14342
14343 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14344
97756c0e
MJ		 143452013-06-11 Martin Jambor <mjambor@suse.cz>
14346
14347 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14348 within bounds at the beginning of the function.
14349
2c7b8bf6
AM		 143502013-06-11 Alan Modra <amodra@gmail.com>
14351
14352 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14353 reporting.
14354 (get_named_section): Don't NULL !DECL_P decl.
14355
77cea46e
IZ		 143562013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14357
14358 * doc/invoke.texi (core-avx2): Document.
14359 (slm): Likewise.
14360 (atom): Updated with MOVBE.
14361
888f0920
RB		 143622013-06-11 Richard Biener <rguenther@suse.de>
14363
22fc7d3d 14364 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14365
6b438d58
AB		 143662013-06-11 Anton Blanchard <anton@samba.org>
14367
14368 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14369 correct shift value in little-endian mode.
14370
3bc66938
JJ		 143712013-06-11 Jakub Jelinek <jakub@redhat.com>
14372
14373 PR target/56564
14374 * varasm.c (get_variable_align): Move #endif to the right place.
14375
a5dc7f88
CC		 143762013-06-10 Cary Coutant <ccoutant@google.com>
14377
14378 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14379 for hash so that hash table traversal order is deterministic.
14380
19be72ab 143812013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14382 Pat Haugen <pthaugen@us.ibm.com>
14383 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 14384
14385 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14386 of vectors in GPRS if they are direct moves or quad word load or
14387 store moves.
14388
14389 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14390 declaration.
14391 (direct_move_p): Likewise.
14392 (quad_load_store_p): Likewise.
14393
14394 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14395 classes into bins based on the physical register type.
14396 (reg_class_to_reg_type): Likewise.
14397 (IS_STD_REG_TYPE): Likewise.
14398 (IS_FP_VECT_REG_TYPE): Likewise.
14399 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14400 use direct move instructions.
14401 (reload_gpr_vsx): Likewise.
14402 (reload_vsx_gpr): Likewise.
14403 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14404 information that is a simplification of register classes. Also
14405 precalculate direct move reload helpers.
14406 (direct_move_p): New function to return true if the operation can
14407 be done as a direct move instruciton.
14408 (quad_load_store_p): New function to return true if the operation
14409 is a quad memory operation.
14410 (rs6000_legitimize_address): If quad memory, only allow register
14411 indirect for TImode addresses.
14412 (rs6000_legitimate_address_p): Likewise.
14413 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14414 (rs6000_reload_register_type): Likewise.
14415 (register_to_reg_type): Return register type.
14416 (rs6000_secondary_reload_simple_move): New helper function for
14417 secondary reload and secondary memory needed to identify anything
14418 that is a simple move, and does not need reloading.
14419 (rs6000_secondary_reload_direct_move): New helper function for
14420 secondary reload to identify cases that can be done with several
14421 instructions via the direct move instructions.
14422 (rs6000_secondary_reload_move): New helper function for secondary
14423 reload to identify moves between register types that can be done.
14424 (rs6000_secondary_reload): Add support for quad memory operations
14425 and for direct move.
14426 (rs6000_secondary_memory_needed): Likewise.
14427 (rs6000_debug_secondary_memory_needed): Change argument names.
14428 (rs6000_output_move_128bit): New function to return the move to
14429 use for 128-bit moves, including knowing about the various
14430 limitations of quad memory operations.
14431
14432 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14433 memory operations. call rs6000_output_move_128bit for the actual
14434 instruciton(s) to generate.
14435 (vsx_movti_64bit): Likewise.
14436
14437 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14438 (UNSPEC_P8V_MTVSRWZ): Likewise.
14439 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14440 (UNSPEC_P8V_MTVSRD): Likewise.
14441 (UNSPEC_P8V_XXPERMDI): Likewise.
14442 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14443 (UNSPEC_FUSION_GPR): Likewise.
14444 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14445 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 14446 (f32_sv): Likewise.
14447 (f32_dm): Likewise.
14448 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14449 loads and direct move instructions.
14450 (zero_extendsidi2_lfiwzx): Likewise.
14451 (extendsidi2_lfiwax): Likewise.
14452 (extendsidi2_nocell): Likewise.
14453 (floatsi<mode>2_lfiwax): Likewise.
14454 (lfiwax): Likewise.
14455 (floatunssi<mode>2_lfiwzx): Likewise.
14456 (lfiwzx): Likewise.
14457 (fix_trunc<mode>_stfiwx): Likewise.
14458 (fixuns_trunc<mode>_stfiwx): Likewise.
14459 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14460 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14461 (parity<mode>2_cmpb): Set length/type attr.
14462 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14463 for 'mr.' to fast_compare.
14464 (bpermd_<mode>): Change type attr to popcnt.
14465 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14466 (p8_mtvsrwz_1): Likewise.
14467 (p8_mtvsrwz_2): Likewise.
14468 (reload_fpr_from_gpr<mode>): Likewise.
14469 (p8_mtvsrd_1): Likewise.
14470 (p8_mtvsrd_2): Likewise.
14471 (p8_xxpermdi_<mode>): Likewise.
14472 (reload_vsx_from_gpr<mode>): Likewise.
14473 (reload_vsx_from_gprsf): Likewise.
14474 (p8_mfvsrd_3_<mode>): LIkewise.
14475 (reload_gpr_from_vsx<mode>): Likewise.
14476 (reload_gpr_from_vsxsf): Likewise.
14477 (p8_mfvsrd_4_disf): Likewise.
14478 (multi-word GPR splits): Do not split direct moves or quad memory
14479 operations.
14480
16876bdc
DM		 144812013-06-10 David Malcolm <dmalcolm@redhat.com>
14482
14483 * tree-into-ssa.c (interesting_blocks): Make static.
14484
df8a1d28
JJ		 144852013-06-10 Jakub Jelinek <jakub@redhat.com>
14486
14487 PR target/56564
14488 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14489 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14490 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14491 (get_variable_align): New function.
14492 (get_variable_section, emit_bss, emit_common,
14493 assemble_variable_contents, place_block_symbol): Use
14494 get_variable_align instead of DECL_ALIGN.
14495 (assemble_noswitch_variable): Add align argument, use it
14496 instead of DECL_ALIGN.
14497 (assemble_variable): Adjust caller. Use get_variable_align
14498 instead of DECL_ALIGN.
14499 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14500 caller.
14501 (DATA_ABI_ALIGNMENT): Define.
14502 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14503 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14504 opt is false, only return the psABI mandated alignment increase.
14505 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14506 (DATA_ABI_ALIGNMENT): ... this.
14507 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14508 (DATA_ABI_ALIGNMENT): ... this.
14509 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14510 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14511 (DATA_ABI_ALIGNMENT): ... this.
14512 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14513 * doc/tm.texi: Regenerated.
14514
57b29ca6
UB		 145152013-06-10 Uros Bizjak <ubizjak@gmail.com>
14516
14517 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14518 cmp_code to construct REG_EQUAL note.
14519
3e56ed50
JJ		 145202013-06-09 Jakub Jelinek <jakub@redhat.com>
14521
14522 PR target/57568
14523 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14524 that operands[2] doesn't overlap with operands[0].
14525
74fe2a1d
DE		 145262013-06-09 David Edelsohn <dje.gcc@gmail.com>
14527 Jan Hubicka <jh@suse.cz>
14528
14529 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14530 hack to mark symbols as used.
14531
ed52a84e
VM		 145322013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14533
14534 PR rtl-optimization/57559
14535 * lra-constraints.c (process_alt_operands): Don't discourage
14536 memory with known offset for offsetable memory constraint.
14537 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14538
299316ed
EB		 145392013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14540
14541 * varasm.c (struct oc_local_state): Reorder fields.
14542 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14543 and adjust accordingly.
14544 (output_constructor): Reorder initialization code and adjust call to
14545 output_constructor_bitfield.
14546
88ac60d3
JH		 145472013-06-07 Jan Hubicka <jh@suse.cz>
14548
14549 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14550
4cdce1a8
DM		 145512013-06-07 David Malcolm <dmalcolm@redhat.com>
14552
14553 * tree-object-size.c (unknown): Make const.
14554
b5e0425c
AK		 145552013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14556
14557 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14558 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14559 for last alternative in the cpu_facility attribute.
14560
79678d04
KT		 145612013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14562
14563 PR target/56315
14564 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14565 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14566 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14567 * config/arm/constraints.md (Dg): New constraint.
14568 * config/arm/neon.md (xordi3_neon): Remove.
14569 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14570 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14571
b31ddbdb
KT		 145722013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14573
14574 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14575 Clean up alternatives.
14576
8f90b7d4
AM		 145772013-06-07 Alan Modra <amodra@gmail.com>
14578
14579 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14580 va_list_gpr_size.
14581
961ce119
AM		 145822013-06-07 Alan Modra <amodra@gmail.com>
14583
14584 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14585
33e49835
KT		 145862013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14587
14588 * config/arm/constraints.md (Df): New constraint.
14589 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14590 Correct length attribute for last two alternatives.
14591
acd17ae6
AM		 145922013-06-07 Alan Modra <amodra@gmail.com>
14593
14594 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14595 override user -mfp-in-toc.
14596 (offsettable_ok_by_alignment): Consider just the current access
14597 rather than the whole object, unless BLKmode. Handle
14598 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14599 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14600 for -mcmodel=medium.
14601 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14602 override user -mfp-in-toc or -msum-in-toc. Default to
14603 -mno-fp-in-toc for -mcmodel=medium.
14604
73310b0e
DD		 146052013-06-06 DJ Delorie <dj@redhat.com>
14606
14607 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14608 TARGET_VALID_POINTER_MODE.
14609
0bd62dca 146102013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14611 Pat Haugen <pthaugen@us.ibm.com>
14612 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 14613
14614 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14615 Document new power8 builtins.
14616
14617 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14618 condition code register, to allow 128-bit logical operations to be
14619 done in the VSX or GPR registers.
14620 (nor<mode>3): Use the canonical form for nor.
14621 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14622 vclz*, and vpopcnt* vector instructions.
14623 (nand<mode>3): Likewise.
14624 (orc<mode>3): Likewise.
14625 (clz<mode>2): LIkewise.
14626 (popcount<mode>2): Likewise.
14627
14628 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14629 that only the GPRs are recognized.
14630
14631 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14632 support for new power8 builtins.
14633
14634 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14635 builtin functions.
14636 (xscvdpspn): Likewise.
14637 (vclz): Likewise.
14638 (vclzb): Likewise.
14639 (vclzh): Likewise.
14640 (vclzw): Likewise.
14641 (vclzd): Likewise.
14642 (vpopcnt): Likewise.
14643 (vpopcntb): Likewise.
14644 (vpopcnth): Likewise.
14645 (vpopcntw): Likewise.
14646 (vpopcntd): Likewise.
14647 (vgbbd): Likewise.
14648 (vmrgew): Likewise.
14649 (vmrgow): Likewise.
14650 (eqv): Likewise.
14651 (eqv_v16qi3): Likewise.
14652 (eqv_v8hi3): Likewise.
14653 (eqv_v4si3): Likewise.
14654 (eqv_v2di3): Likewise.
14655 (eqv_v4sf3): Likewise.
14656 (eqv_v2df3): Likewise.
14657 (nand): Likewise.
14658 (nand_v16qi3): Likewise.
14659 (nand_v8hi3): Likewise.
14660 (nand_v4si3): Likewise.
14661 (nand_v2di3): Likewise.
14662 (nand_v4sf3): Likewise.
14663 (nand_v2df3): Likewise.
14664 (orc): Likewise.
14665 (orc_v16qi3): Likewise.
14666 (orc_v8hi3): Likewise.
14667 (orc_v4si3): Likewise.
14668 (orc_v2di3): Likewise.
14669 (orc_v4sf3): Likewise.
14670 (orc_v2df3): Likewise.
14671
14672 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14673 allow power8 quad mode in 64-bit.
14674 (rs6000_builtin_vectorized_function): Add support to vectorize
14675 ISA 2.07 count leading zeros, population count builtins.
14676 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14677 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14678 (builtin_function_type): Add vgbbd builtin function which takes an
14679 unsigned argument.
14680 (altivec_expand_vec_perm_const): Add support for new power8 merge
14681 instructions.
14682
14683 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14684 that does not include TImdoe for use with 32-bit.
14685 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14686 instructions.
14687 (UNSPEC_VSX_CVDPSPN): Likewise.
14688 (vsx_xscvdpspn): Likewise.
14689 (vsx_xscvspdpn): Likewise.
14690 (vsx_xscvdpspn_scalar): Likewise.
14691 (vsx_xscvspdpn_directmove): Likewise.
14692 (vsx_and<mode>3): Split logical operations into 32-bit and
14693 64-bit. Add support to do logical operations on TImode as well as
14694 VSX vector types. Allow logical operations to be done in either
14695 VSX registers or in general purpose registers in 64-bit mode. Add
14696 splitters if GPRs were used. For AND, add clobber of CCmode to
14697 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14698 encoding.
14699 (vsx_and<mode>3_32bit): Likewise.
14700 (vsx_and<mode>3_64bit): Likewise.
14701 (vsx_ior<mode>3): Likewise.
14702 (vsx_ior<mode>3_32bit): Likewise.
14703 (vsx_ior<mode>3_64bit): Likewise.
14704 (vsx_xor<mode>3): Likewise.
14705 (vsx_xor<mode>3_32bit): Likewise.
14706 (vsx_xor<mode>3_64bit): Likewise.
14707 (vsx_one_cmpl<mode>2): Likewise.
14708 (vsx_one_cmpl<mode>2_32bit): Likewise.
14709 (vsx_one_cmpl<mode>2_64bit): Likewise.
14710 (vsx_nor<mode>3): Likewise.
14711 (vsx_nor<mode>3_32bit): Likewise.
14712 (vsx_nor<mode>3_64bit): Likewise.
14713 (vsx_andc<mode>3): Likewise.
14714 (vsx_andc<mode>3_32bit): Likewise.
14715 (vsx_andc<mode>3_64bit): Likewise.
14716 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14717 and xxlorc instructions.
14718 (vsx_eqv<mode>3_64bit): Likewise.
14719 (vsx_nand<mode>3_32bit): Likewise.
14720 (vsx_nand<mode>3_64bit): Likewise.
14721 (vsx_orc<mode>3_32bit): Likewise.
14722 (vsx_orc<mode>3_64bit): Likewise.
14723
14724 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14725
14726 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14727 instruction.
14728 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14729 (p8_vmrgow): Likewise.
14730 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14731 GPRs to be split under VSX.
14732 (p8v_clz<mode>2): Add power8 count leading zero support.
14733 (p8v_popcount<mode>2): Add power8 population count support.
14734 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14735 support.
14736
14737 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14738 instruction.
14739
14740 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14741 builtin functions.
14742 (vec_nand): Likewise.
14743 (vec_vclz): Likewise.
14744 (vec_vclzb): Likewise.
14745 (vec_vclzd): Likewise.
14746 (vec_vclzh): Likewise.
14747 (vec_vclzw): Likewise.
14748 (vec_vgbbd): Likewise.
14749 (vec_vmrgew): Likewise.
14750 (vec_vmrgow): Likewise.
14751 (vec_vpopcnt): Likewise.
14752 (vec_vpopcntb): Likewise.
14753 (vec_vpopcntd): Likewise.
14754 (vec_vpopcnth): Likewise.
14755 (vec_vpopcntw): Likewise.
14756
37684c46
VM		 147572013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14758
14759 PR rtl-optimization/57468
14760 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14761 spilled pseudos.
14762
d89ae27c
VM		 147632013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14764
14765 PR rtl-optimization/57459
14766 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14767 type when setting live regs.
14768
3597e113
VM		 147692013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14770
14771 * config/s390/s390.opt (mlra): New option.
14772 * config/s390/s390.c (s390_decompose_address): Check displacement
14773 for all registers for LRA.
14774 (s390_secondary_reload): Don't used secondary reloads for LRA.
14775 (s390_lra_p): New function.
14776 (TARGET_LRA_P): Define.
14777 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14778 of attribute cpu_facility to zarch for the last alternative.
14779 (*cmpmem_short): Ditto.
14780
01037aeb
EB		 147812013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14782
14783 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14784 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14785 (arm_expand_prologue): Likewise.
14786
3371a64f
TJ		 147872013-06-06 Teresa Johnson <tejohnson@google.com>
14788
14789 PR c++/53743
14790 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14791 as this is now done by redirect_edge_and_branch_force.
14792 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14793 barriers, and fix interaction with splitting.
14794 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14795 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14796 reflect changes made in the routine.
14797 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14798 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14799 since this is called in cfglayout mode, and replace partition fixup
14800 with assert as that is now done by force_nonfallthru_and_redirect.
14801 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14802 already be marked with region crossing note.
14803 (insert_section_boundary_note): Make non-static, gate on flag
14804 has_bb_partition, rewrite to also check for multiple partitions.
14805 (rest_of_handle_reorder_blocks): Remove call to
14806 insert_section_boundary_note, now done later during free_cfg.
14807 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14808 * bb-reorder.h (insert_section_boundary_note): Declare.
14809 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14810 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14811 invoke insert_section_boundary_note.
14812 (try_redirect_by_replacing_jump): Remove unnecessary
14813 check for region crossing note.
14814 (fixup_partition_crossing): New function.
14815 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14816 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14817 in non-cfglayout mode.
14818 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14819 remove old code that tried to do this. Emit barrier correctly
14820 when we are in cfglayout mode.
14821 (last_bb_in_partition): New function.
14822 (rtl_split_edge): Correctly fixup partition boundaries.
14823 (commit_one_edge_insertion): Remove old code that tried to
14824 fixup region crossing edge since this is now handled in
14825 split_block, and set up insertion point correctly since
14826 block may now end in a jump.
14827 (verify_hot_cold_block_grouping): Guard against checking when not in
14828 linearized RTL mode.
14829 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14830 notes.
14831 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14832 rtl_verify_flow_info, so not called in cfglayout mode.
14833 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14834 (fixup_reorder_chain): Remove old code that attempted to fixup region
14835 crossing note as this is now handled in force_nonfallthru_and_redirect.
14836 (duplicate_insn_chain): Don't duplicate switch section notes.
14837 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14838 note.
14839 * basic-block.h (emit_barrier_after_bb): Declare.
14840
66071e10
KT		 148412013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14842
14843 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14844 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14845 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14846
0e26bf3d
KT		 148472013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14848
14849 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14850 where appropriate.
14851 * config/arm/ldmstm.md: Regenerate.
14852
12b4e7ef
KT		 148532013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14854
14855 * config/arm/sync.md (atomic_loaddi_1):
14856 Disable predication for arm_restrict_it.
14857 (arm_load_exclusive<mode>): Likewise.
14858 (arm_load_exclusivesi): Likewise.
14859 (arm_load_exclusivedi): Likewise.
14860 (arm_load_acquire_exclusive<mode>): Likewise.
14861 (arm_load_acquire_exclusivesi): Likewise.
14862 (arm_load_acquire_exclusivedi): Likewise.
14863 (arm_store_exclusive<mode>): Likewise.
14864 (arm_store_exclusive<mode>): Likewise.
14865 (arm_store_release_exclusivedi): Likewise.
14866 (arm_store_release_exclusive<mode>): Likewise.
14867
15d16c8a
RB		 148682013-06-06 Richard Biener <rguenther@suse.de>
14869
14870 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14871 after LTO_null.
14872 (lto_tag_is_tree_code_p): Adjust.
14873 (lto_tag_is_gimple_code_p): Likewise.
14874 (lto_gimple_code_to_tag): Likewise.
14875 (lto_tag_to_gimple_code): Likewise.
14876 (lto_tree_code_to_tag): Likewise.
14877 (lto_tag_to_tree_code): Likewise.
14878 * data-streamer.h (streamer_write_hwi_in_range): Use
14879 uhwi streaming to stream the normalized range.
14880 (streamer_read_hwi_in_range): Likewise.
14881
17a7fc37
KT		 148822013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14883
14884 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14885 (predicable_short_it): Likewise.
14886 (predicated): Likewise.
14887 (enabled): Handle above.
14888 (define_cond_exec): Set predicated attribute to yes.
14889
b57ca59b
MS		 148902013-06-05 Mike Stump <mikestump@comcast.net>
14891
14892 * gdbinit.in (__FUNCTION__): Add.
14893
c1e183a9
UB		 148942013-06-05 Uros Bizjak <ubizjak@gmail.com>
14895
14896 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14897 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14898
0378bddb
JH		 148992013-06-05 Jan Hubicka <jh@suse.cz>
14900
14901 * varasm.c (mark_decl_referenced): Revert the removal until targets
14902 are fixed.
14903
8f8a2057
DE		 149042013-06-05 David Edelsohn <dje.gcc@gmail.com>
14905
14906 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14907 instead of mark_decl_referenced.
14908
edb983b2
JH		 149092013-06-05 Jan Hubicka <jh@suse.cz>
14910
14911 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14912 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14913 and symtab_used_from_object_file_p.
14914 (cgraph_make_node_local_1): Clear forced_by_abi.
14915 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14916 * cgraph.h (symtab_node_base): Add forced_by_abi.
14917 (decide_is_variable_needed): Remove.
14918 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14919 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14920 (decide_is_symbol_needed): ... this one; handle symbols in general;
14921 always analyze virtuals; honnor forced_by_abi.
14922 (cgraph_finalize_function): Update.
14923 (varpool_finalize_decl): Update.
14924 (symbol_defined_and_needed): Remove.
14925 (analyze_functions): Update.
14926 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14927 output_refs, input_overwrite_node): Handle forced_by_abi.
14928 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14929 (address_taken_from_non_vtable_p): ... this one.
14930 (comdat_can_be_unshared_p_1): New function.
14931 (cgraph_comdat_can_be_unshared_p): Rename to ...
14932 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14933 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14934 (function_and_variable_visibility): Clear forced_by_abi as needed.
14935 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14936 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 14937 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14938 * varpool.c (decide_is_variable_needed): Remove.
14939
9912dbe5
KT		 149402013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14941
14942 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14943 (arm_option_override): Override arm_restrict_it where appropriate.
14944 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14945 * config/arm/arm.opt (mrestrict-it): New command-line option.
14946 * doc/invoke.texi: Document -mrestrict-it.
14947
34c136b6
DM		 149482013-06-05 David Malcolm <dmalcolm@redhat.com>
14949
14950 * tsan.c (tsan_atomic_table): Make const.
14951
a367df53
RB		 149522013-06-05 Richard Biener <rguenther@suse.de>
14953
14954 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14955 index associated with the tree we are supposed to replace.
14956 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14957 TREE_ASM_WRITTEN as zero for everything but SSA names.
14958
70d51a19
DM		 149592013-06-05 David Malcolm <dmalcolm@redhat.com>
14960
14961 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14962
b9bd2075
JH		 149632013-06-04 Jan Hubicka <jh@suse.cz>
14964
14965 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14966 (input_node, input_varpool_node): Handle correctly external same
14967 body aliases.
14968 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14969 nodes at ltrans stage.
14970
aaae719d
JH		 149712013-06-04 Jan Hubicka <jh@suse.cz>
14972
14973 * ipa-inline.c (update_caller_keys): Fix availability test.
14974 (update_callee_keys): Likewise.
14975 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14976 to follow ELF standard.
14977
107eea2c
JU		 149782013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14979
14980 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14981 (mips64r5900el-*-elf*): New configurations.
14982 * config/mips/mips-cpus.def (r5900): New processor.
14983 * config/mips/mips-tables.opt: Regenerate.
14984 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14985 (mips_issue_rate): Handle PROCESSOR_R5900.
14986 (mips_reorg_process_insns): Force reorder mode for the R5900.
14987 * config/mips/mips.h (TARGET_MIPS5900): Define.
14988 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14989 TARGET_MIPS5900.
14990 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14991 TARGET_MIPS5900.
14992 * config/mips/mips.md (processor): Add r5900.
14993 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14994
b7342d25
IB		 149952013-06-04 Ian Bolton <ian.bolton@arm.com>
14996
14997 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14998 into function to generate MOVI instruction.
c1e183a9 14999 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 15000 (aarch64_preferred_simd_mode): Turn into wrapper.
15001 (aarch64_output_scalar_simd_mov_immediate): New function.
15002 * config/aarch64/aarch64-protos.h: Add prototype for above.
15003
81c2dfb9
IB		 150042013-06-04 Ian Bolton <ian.bolton@arm.com>
15005
15006 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15007 element_char member.
15008 (sizetochar): Return signed char.
15009 (aarch64_simd_valid_immediate): Remove elchar and other
15010 unnecessary variables.
15011 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15012 Calculate element_char as required.
15013 * config/aarch64/aarch64-protos.h: Update and move prototype
15014 for aarch64_output_simd_mov_immediate.
15015 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15016 Update arguments.
15017
48063b9d
IB		 150182013-06-04 Ian Bolton <ian.bolton@arm.com>
15019
15020 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15021 information completed by aarch64_simd_valid_immediate.
15022 (aarch64_legitimate_constant_p): Update arguments.
15023 (aarch64_simd_valid_immediate): Work with struct rather than many
15024 pointers.
15025 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15026 (aarch64_simd_make_constant): Update arguments.
15027 (aarch64_output_simd_mov_immediate): Work with struct rather than
15028 many pointers. Output immediate directly rather than as operand.
15029 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15030 Update prototype.
15031 * config/aarch64/constraints.md (Dn): Update arguments.
15032
3ea63f60
IB		 150332013-06-04 Ian Bolton <ian.bolton@arm.com>
15034
15035 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15036 longer static.
15037 (aarch64_simd_immediate_valid_for_move): Remove.
15038 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15039 (aarch64_simd_make_constant): Update call.
15040 (aarch64_output_simd_mov_immediate): Update call.
15041 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15042 Add prototype.
15043 * config/aarch64/constraints.md (Dn): Update call.
15044
d8edd899
IB		 150452013-06-04 Ian Bolton <ian.bolton@arm.com>
15046
15047 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15048 return type to bool for prototype.
15049 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15050 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15051 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15052
44b20bb8
CM		 150532013-06-04 Catherine Moore <clm@codesourcery.com>
15054
15055 * config/mips/mips.opt (meva): New.
c1e183a9 15056 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 15057 (ASM_SPEC): Handle -meva.
15058 * doc/invoke.texi (meva): Document.
15059
52befbd8
AM		 150602013-06-04 Alan Modra <amodra@gmail.com>
15061
15062 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15063 constant output.
15064
aadaf24e
KT		 150652013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15066
15067 * rtl.def: Add extra fourth optional field to define_cond_exec.
15068 * gensupport.c (process_one_cond_exec): Process attributes from
15069 define_cond_exec.
15070 * doc/md.texi: Document fourth field in define_cond_exec.
15071
0cd9e9ee
EB		 150722013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15073
15074 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15075 out the processing order as in store_bit_field_1.
15076
cb0f665d
JH		 150772013-06-04 Jan Hubicka <jh@suse.cz>
15078
15079 PR middle-end/57500
15080 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15081 non-VAR_DECL node if it does not exist yet.
15082
53984b9b
RS		 150832013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15084
15085 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15086 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15087 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15088 target_cpu_default setting.
15089
5d318fd4
TJ		 150902013-06-03 Teresa Johnson <tejohnson@google.com>
15091
15092 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15093 default to -fopt-info=optimized instead of all.
15094 * doc/invoke.texi: Ditto.
15095 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15096 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15097 (execute_vect_slp): Emit BB vectorization success under
15098 MSG_OPTIMIZED_LOCATIONS.
15099 * tree-vect-slp.c (vect_slp_transform_bb): Change
15100 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15101 * tree-vect-loop.c (vect_transform_loop): Ditto.
15102
ec3af349
JM		 151032013-06-03 Jason Merrill <jason@redhat.com>
15104
15105 PR c++/57415
15106 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15107 Use TARGET_EXPR for C++.
15108
1baf9159
JJ		 151092013-06-03 Jakub Jelinek <jakub@redhat.com>
15110
15111 PR rtl-optimization/57268
0cd9e9ee 15112 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 15113 if DEBUG_INSN_P (insn).
15114
15115 Reapply
15116 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15117
15118 PR rtl-optimization/57268
15119 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15120 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15121
0f1d3965
YR		 151222013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15123
15124 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15125 (ix86_avoid_lea_for_addr): Likewise.
15126 (exact_dependency_1): Likewise.
15127 (ix86_adjust_cost): Likewise.
15128 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15129 removed.
15130 (do_reorder_for_imul): Fix typo, formatting and
15131 !reload_completed check removed.
15132 (ix86_sched_reorder): Fix typo and formatting.
15133 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15134 list.
15135
3bc30db4
SN		 151362013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15137
15138 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15139
19cc6fac
EB		 151402013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15141
15142 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15143 <STRING_CST>: Likewise.
15144 <VECTOR_CST>: Likewise.
15145
c68a6e08
JW		 151462013-06-01 Janus Weil <janus@gcc.gnu.org>
15147 Mikael Morin <mikael@gcc.gnu.org>
15148
15149 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15150 * config.in: Regenerated.
15151 * configure: Regenerated.
15152
38e55e5c
JH		 151532013-06-01 Jan Hubicka <jh@suse.cz>
15154
15155 PR middle-end/57366
15156 * cgraphunit.c (compile): When weakref is not supported,
15157 set up transparent aliases before final output pass.
15158 * varasm.c (assemble_alias): Do not try to do it here.
15159
eb51d2ff
JH		 151602013-06-01 Jan Hubicka <jh@suse.cz>
15161
15162 PR middle-end/57467
15163 * passes.c (for_per_function): Skip unanalyzed functions.
15164
40a7fe1e
JH		 151652013-06-01 Jan Hubicka <jh@suse.cz>
15166
15167 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15168 (lto_symtab_merge_symbols_1): ... this one.
15169 (lto_symtab_merge_cgraph_nodes): Rename to ...
15170 (lto_symtab_merge_symbols): ... this one; simplify.
15171 * cgraph.c (same_body_aliases_done): Rename to ...
15172 (cpp_implicit_aliases_done): ... this one.
15173 (cgraph_create_function_alias): Update.
15174 (cgraph_same_body_alias): Update.
c1e183a9 15175 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 15176 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 15177 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 15178 (cgraph_node): Remove same_body_alias.
15179 (varpool_node): Remove alias_of and extra_name_alias.
15180 (same_body_aliases_done): Rename to ..
15181 (cpp_implicit_aliases_done): ... this one.
15182 (symtab_alias_ultimate_target): Add default parameter.
15183 (symtab_resolve_alias): New function.
15184 (fixup_same_cpp_alias_visibility): Declare.
15185 (cgraph_function_node): Add default parameter.
15186 (cgraph_node_asm_name): Likewise.
15187 (cgraph_function_or_thunk_node): Add default parameter; do
15188 not ICE when it is NULL.
15189 (varpool_variable_node): Likewise.
15190 * tree-emutls.c (create_emultls_var): Update.
15191 (ipa_lower_emutls): Update.
15192 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15193 (cgraph_reset_node): Reset alias info.
15194 (cgraph_finalize_function): Update.
15195 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15196 (analyze_function): Simplify.
15197 (cgraph_process_same_body_aliases): Simplify.
15198 (analyze_functions): Fixup same body aliases.
15199 (handle_alias_pairs): Simplify.
15200 (assemble_thunk): Update.
15201 (assemble_thunks_and_aliases): Update.
15202 (output_weakrefs): Rewrite.
15203 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15204 (lto_output_varpool_node): Likewise.
15205 (compute_ltrans_boundary): Remve assert.
15206 (get_alias_symbol): New functoin.
15207 (input_node): Rewrite alias handling.
15208 (input_varpool_node): Likewise.
15209 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15210 * ipa.c (process_references): Handle weakrefs correctly.
15211 (symtab_remove_unreachable_nodes): Likewise.
15212 * trans-mem.c (get_cg_data): Update.
15213 (ipa_tm_create_version_alias): Update.
15214 (ipa_tm_execute): Update.
15215 * symtab.c (dump_symtab_base): Dump aliases.
15216 (verify_symtab_base): Verify aliases.
15217 (symtab_node_availability): New function.
15218 (symtab_alias_ultimate_target): Simplify.
15219 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15220 handle all the fixup cases.
15221 (symtab_resolve_alias): New function.
15222 * passes.c (ipa_write_summaries): Handle weakrefs.
15223 * varpool.c (varpool_analyze_node): Simplify.
15224 (assemble_aliases): Update.
15225 (varpool_create_variable_alias): Simplify.
15226 (varpool_extra_name_alias): Simplify.
15227 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15228 (lto_symtab_merge_symbols): ... this one.
15229
21759881
DT		 152302013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15231
15232 Revert
15233 PR rtl-optimization/57268
15234 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15235 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15236
daf4e940
TB		 152372013-06-01 Tobias Burnus <burnus@net-b.de>
15238
15239 Partially reverted:
15240 2013-05-31 Tobias Burnus <burnus@net-b.de>
15241
15242 PR middle-end/57073
15243 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15244 further up.
15245
c359d8d0
DT		 152462013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15247
15248 PR rtl-optimization/57268
15249 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15250 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15251
4172215d
EB		 152522013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15253
15254 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15255 unordered comparison operators when -fno-trapping-math is in effect
15256 on the e500.
15257 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15258 and implement unordered comparison operators properly on the e500.
15259
a8c50132
EB		 152602013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15261
15262 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15263 for constant scalar integers.
15264 (simplify_relational_operation_1): Likewise.
15265
44626634
SB		 152662013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15267
15268 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15269 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15270 Fix comment.
15271
55805e54
YR		 152722013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15273 Igor Zamyatin <igor.zamyatin@intel.com>
15274
15275 Silvermont (SLM) architecture performance tuning.
15276 * config/i386/i386.h (enum ix86_tune_indices): Add
15277 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15278 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15279
15280 * config/i386/i386.c (initial_ix86_tune_features)
15281 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15282 (ix86_lea_outperforms): Handle Silvermont tuning.
15283 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15284 call.
15285 (ix86_use_lea_for_mov): Likewise.
15286 (ix86_avoid_lea_for_addr): Likewise.
15287 (ix86_lea_for_add_ok): Likewise.
15288 (exact_dependency_1): New function.
15289 (exact_store_load_dependency): Likewise.
15290 (ix86_adjust_cost): Handle Silvermont tuning.
15291 (do_reoder_for_imul): Likewise.
15292 (swap_top_of_ready_list): New function.
15293 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15294
15295 * config/i386/i386.md (peepholes that split memory operand in fp
15296 converts): New.
15297
e19c9de2
MS		 152982013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15299
15300 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15301 Remove un-necessary braces.
15302
38e6c9a6
MS		 153032013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15304
15305 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15306 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15307
e3530904
TB		 153082013-05-31 Tobias Burnus <burnus@net-b.de>
15309
15310 PR middle-end/57073
c1e183a9 15311 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 15312
b6af05a9
KT		 153132013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15314
15315 PR target/56315
15316 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15317 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15318 * config/arm/neon.md (iordi3_neon): Remove.
15319 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15320 * config/arm/predicates.md (imm_for_neon_logic_operand):
15321 Move to earlier in the file.
15322 (neon_logic_op2): Likewise.
15323 (arm_iordi_operand_neon): New predicate.
15324
f800c166
RB		 153252013-05-31 Richard Biener <rguenther@suse.de>
15326
15327 PR tree-optimization/57478
15328 PR tree-optimization/57453
15329 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15330 are life as well.
15331
e571fa59
KP		 153322013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15333
15334 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15335 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15336
0fa6e0ef
TB		 153372013-05-30 Tobias Burnus <burnus@net-b.de>
15338 Thomas Koenig <tkoenig@gcc.gnu.org>
15339
15340 PR middle-end/57073
15341 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15342 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15343
4688ddf5
SB		 153442013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15345
0fa6e0ef 15346 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15347
3b9ceb4b
VM		 153482013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15349
15350 * target.def (register_usage_leveling_p): New hook.
15351 * targhooks.c (default_register_usage_leveling_p): New.
15352 * targhooks.h (default_register_usage_leveling_p): New prototype.
15353 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15354 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15355 * doc/tm.texi: Update.
15356 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15357
65f9e789
IB		 153582013-05-30 Ian Bolton <ian.bolton@arm.com>
15359
15360 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15361 (*insv_reg<mode>): New define_insn.
15362
12211b99 153632013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 15364
15365 PR rtl-optimization/57439
15366 * postreload.c (move2add_valid_value_p): Check that we have
15367 a zero subreg_regno_offset when accessing the register in
15368 the requested mode.
15369
0b871ccf
YR		 153702013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15371 Igor Zamyatin <igor.zamyatin@intel.com>
15372
15373 Silvermont (SLM) architecture pipeline model, tuning and
15374 insn selection.
15375 * config.gcc: Add slm config options and target.
15376
15377 * config/i386/slm.md: New.
15378
15379 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15380
f43245d1 15381 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR		 15382 PROCESSOR_SLM.
15383 (ix86_target_macros_internal): Likewise.
15384
f43245d1 15385 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR		 15386 (m_SLM): New macro flag.
15387 (initial_ix86_tune_features): Set m_SLM.
15388 (x86_accumulate_outgoing_args): Likewise.
15389 (x86_arch_always_fancy_math_387): Likewise.
15390 (processor_target_table): Add slm cost.
15391 (cpu_names): Add slm cpu name.
15392 (x86_option_override_internal): Set SLM ISA.
15393 (ix86_issue_rate): New case PROCESSOR_SLM.
15394 (ia32_multipass_dfa_lookahead): Likewise.
15395 (fold_builtin_cpu): Add slm.
15396
15397 * config/i386/i386.h (TARGET_SLM): New target macro.
15398 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15399 (processor_type): Add PROCESSOR_SLM.
15400
15401 * config/i386/i386.md (cpu): Add new value "slm".
15402 (slm.md): Include slm.md.
15403
24d5b097
XG		 154042013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15405 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15406
15407 * config/arm/arm-protos.h: Add and update function protos.
15408 * config/arm/arm.c (use_simple_return_p): New added.
15409 (thumb2_expand_return): Check simple_return flag.
15410 * config/arm/arm.md: Add simple_return and conditional simple_return.
15411 * config/arm/iterators.md: Add iterator for return and simple_return.
15412
c1cccc15
ZC		 154132013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15414
15415 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15416 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15417 (arm_emit_vfp_multi_reg_pop): Likewise.
15418 (thumb2_emit_ldrd_pop): Likewise.
15419 (arm_expand_epilogue): Add misc REG_CFA notes.
15420 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15421
3a4f280b
LC		 154222013-05-29 Lawrence Crowl <crowl@google.com>
15423
15424 * config/arm/t-arm: Update for below.
15425
15426 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15427 Change type to hash_table. Update dependent calls and types.
15428
15429 * config/i386/t-cygming: Update for below.
15430
15431 * config/i386/t-interix: Update for below.
15432
15433 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15434 Change type to hash_table. Update dependent calls and types.
15435 (i386_find_on_wrapper_list::wrappers): Likewise.
15436
15437 * config/ia64/t-ia64: Update for below.
15438
15439 * config/ia64/ia64.c (bundle_state_table):
15440 Change type to hash_table. Update dependent calls and types.
15441
15442 * config/mips/mips.c (mips_reorg_process_insns::htab):
15443 Change type to hash_table. Update dependent calls and types.
15444
15445 * config/sol2.c (solaris_comdat_htab):
15446 Change type to hash_table. Update dependent calls and types.
15447
15448 * config/t-sol2: Update for above.
15449
1388a0e3
TJ		 154502013-05-29 Teresa Johnson <tejohnson@google.com>
15451
15452 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15453 functions are not yet marked as defined.
15454
a5965b52 154552013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15456 Pat Haugen <pthaugen@us.ibm.com>
15457 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 15458
15459 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15460 instructions.
15461 (VEC_A): Likewise.
15462 (VEC_C): Likewise.
15463 (vrotl<mode>3): Likewise.
15464 (vashl<mode>3): Likewise.
15465 (vlshr<mode>3): Likewise.
15466 (vashr<mode>3): Likewise.
15467
15468 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15469 support for power8 V2DI builtins.
15470
15471 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15472 power8 V2DI builtins.
15473 (vupkhsw): Likewise.
15474 (vupklsw): Likewise.
15475 (vaddudm): Likewise.
15476 (vminsd): Likewise.
15477 (vmaxsd): Likewise.
15478 (vminud): Likewise.
15479 (vmaxud): Likewise.
15480 (vpkudum): Likewise.
15481 (vpksdss): Likewise.
15482 (vpkudus): Likewise.
15483 (vpksdus): Likewise.
15484 (vrld): Likewise.
15485 (vsld): Likewise.
15486 (vsrd): Likewise.
15487 (vsrad): Likewise.
15488 (vsubudm): Likewise.
15489 (vcmpequd): Likewise.
15490 (vcmpgtsd): Likewise.
15491 (vcmpgtud): Likewise.
15492 (vcmpequd_p): Likewise.
15493 (vcmpgtsd_p): Likewise.
15494 (vcmpgtud_p): Likewise.
15495 (vupkhsw): Likewise.
15496 (vupklsw): Likewise.
15497 (vaddudm): Likewise.
15498 (vmaxsd): Likewise.
15499 (vmaxud): Likewise.
15500 (vminsd): Likewise.
15501 (vminud): Likewise.
15502 (vpksdss): Likewise.
15503 (vpksdus): Likewise.
15504 (vpkudum): Likewise.
15505 (vpkudus): Likewise.
15506 (vrld): Likewise.
15507 (vsld): Likewise.
15508 (vsrad): Likewise.
15509 (vsrd): Likewise.
15510 (vsubudm): Likewise.
15511
15512 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15513 support for power8 V2DI instructions.
15514
15515 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15516 power8 V2DI instructions. Combine pack and unpack insns to use an
15517 iterator for each mode. Check whether a particular mode supports
15518 Altivec instructions instead of just checking TARGET_ALTIVEC.
15519 (UNSPEC_VPKUWUM): Likewise.
15520 (UNSPEC_VPKSHSS): Likewise.
15521 (UNSPEC_VPKSWSS): Likewise.
15522 (UNSPEC_VPKUHUS): Likewise.
15523 (UNSPEC_VPKSHUS): Likewise.
15524 (UNSPEC_VPKUWUS): Likewise.
15525 (UNSPEC_VPKSWUS): Likewise.
15526 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15527 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15528 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15529 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15530 (UNSPEC_VUPKHSB): Likewise.
15531 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15532 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15533 (UNSPEC_VUPKHSH): Likewise.
15534 (UNSPEC_VUPKLSB): Likewise.
15535 (UNSPEC_VUPKLSH): Likewise.
15536 (VI2): Likewise.
15537 (VI_char): Likewise.
15538 (VI_scalar): Likewise.
15539 (VI_unit): Likewise.
15540 (VP): Likewise.
15541 (VP_small): Likewise.
15542 (VP_small_lc): Likewise.
15543 (VU_char): Likewise.
15544 (add<mode>3): Likewise.
15545 (altivec_vaddcuw): Likewise.
15546 (altivec_vaddu<VI_char>s): Likewise.
15547 (altivec_vadds<VI_char>s): Likewise.
15548 (sub<mode>3): Likewise.
15549 (altivec_vsubcuw): Likewise.
15550 (altivec_vsubu<VI_char>s): Likewise.
15551 (altivec_vsubs<VI_char>s): Likewise.
15552 (altivec_vavgs<VI_char>): Likewise.
15553 (altivec_vcmpbfp): Likewise.
15554 (altivec_eq<mode>): Likewise.
15555 (altivec_gt<mode>): Likewise.
15556 (altivec_gtu<mode>): Likewise.
15557 (umax<mode>3): Likewise.
15558 (smax<mode>3): Likewise.
15559 (umin<mode>3): Likewise.
15560 (smin<mode>3): Likewise.
15561 (altivec_vpkuhum): Likewise.
15562 (altivec_vpkuwum): Likewise.
15563 (altivec_vpkshss): Likewise.
15564 (altivec_vpkswss): Likewise.
15565 (altivec_vpkuhus): Likewise.
15566 (altivec_vpkshus): Likewise.
15567 (altivec_vpkuwus): Likewise.
15568 (altivec_vpkswus): Likewise.
15569 (altivec_vpks<VI_char>ss): Likewise.
15570 (altivec_vpks<VI_char>us): Likewise.
15571 (altivec_vpku<VI_char>us): Likewise.
15572 (altivec_vpku<VI_char>um): Likewise.
15573 (altivec_vrl<VI_char>): Likewise.
15574 (altivec_vsl<VI_char>): Likewise.
15575 (altivec_vsr<VI_char>): Likewise.
15576 (altivec_vsra<VI_char>): Likewise.
15577 (altivec_vsldoi_<mode>): Likewise.
15578 (altivec_vupkhsb): Likewise.
15579 (altivec_vupkhs<VU_char>): Likewise.
15580 (altivec_vupkls<VU_char>): Likewise.
15581 (altivec_vupkhsh): Likewise.
15582 (altivec_vupklsb): Likewise.
15583 (altivec_vupklsh): Likewise.
15584 (altivec_vcmpequ<VI_char>_p): Likewise.
15585 (altivec_vcmpgts<VI_char>_p): Likewise.
15586 (altivec_vcmpgtu<VI_char>_p): Likewise.
15587 (abs<mode>2): Likewise.
15588 (vec_unpacks_hi_v16qi): Likewise.
15589 (vec_unpacks_hi_v8hi): Likewise.
15590 (vec_unpacks_lo_v16qi): Likewise.
15591 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15592 (vec_unpacks_lo_v8hi): Likewise.
15593 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15594 (vec_pack_trunc_v8h): Likewise.
15595 (vec_pack_trunc_v4si): Likewise.
15596 (vec_pack_trunc_<mode>): Likewise.
15597
15598 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15599 V2DI builtins.
15600 (vec_vmaxsd): Likewise.
15601 (vec_vmaxud): Likewise.
15602 (vec_vminsd): Likewise.
15603 (vec_vminud): Likewise.
15604 (vec_vpksdss): Likewise.
15605 (vec_vpksdus): Likewise.
15606 (vec_vpkudum): Likewise.
15607 (vec_vpkudus): Likewise.
15608 (vec_vrld): Likewise.
15609 (vec_vsld): Likewise.
15610 (vec_vsrad): Likewise.
15611 (vec_vsrd): Likewise.
15612 (vec_vsubudm): Likewise.
15613 (vec_vupkhsw): Likewise.
15614 (vec_vupklsw): Likewise.
15615
e70670cf
JH		 156162013-05-29 Jan Hubicka <jh@suse.cz>
15617
15618 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15619 flags; reorder rest of fields in more consistent way.
15620 (varpool_node): Remove analyzed, finalized and alias.
15621 (cgraph_ndoe): Likewise.
15622 (symtab_alias_ultimate_target): New function.
15623 (cgraph_function_node): Move offline.
15624 (cgraph_reset_node): Declare.
15625 (cgraph_comdat_can_be_unshared_p): Remove.
15626 (varpool_remove_initializer): Declare.
15627 (varpool_first_defined_variable, varpool_next_defined_variable
15628 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15629 (cgraph_function_with_gimple_body_p): Update.
15630 (varpool_all_refs_explicit_p): Update.
15631 (symtab_alias_target): New function.
15632 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15633 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 15634 (cgraph_function_or_thunk_node): Simplify using
15635 symtab_alias_ultimate_target.
e70670cf
JH		 15636 (varpool_variable_node): Likewise.
15637 * cgraph.c (cgraph_create_function_alias): Update.
15638 (cgraph_add_thunk): Update.
15639 (cgraph_remove_node): Update.
15640 (dump_cgraph_node): Do not dump removed flags.
15641 (cgraph_function_body_availability): Update.
15642 (cgraph_propagate_frequency): Update.
15643 (verify_cgraph_node): Check sanity of local flag.
15644 (cgraph_function_node): Move here from cgraph.h; revamp for
15645 cgraph_function_or_thunk_node.
15646 * lto-symtab.c (lto_varpool_replace_node): Update.
15647 (lto_symtab_resolve_can_prevail_p): Update.
15648 (lto_symtab_merge_cgraph_nodes): Update.
15649 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15650 propagate_constants_accross_call, devirtualization_time_bonus,
15651 ipcp_propagate_stage): Update.
15652 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 15653 * ipa-inline-transform.c (clone_inlined_nodes,
15654 preserve_function_body_p): Update.
e70670cf
JH		 15655 * ipa-reference.c (propagate): Update.
15656 (write_node_summary_p): Update.
15657 * toplev.c (wrapup_global_declaration_2): Update.
15658 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15659 (analyze_function) ... this one.
15660 (cgraph_process_new_functions): Update.
15661 (cgraph_reset_node): Export.
15662 (cgraph_finalize_function): Update.
15663 (cgraph_add_new_function): Update.
15664 (process_function_and_variable_attributes): Update.
15665 (varpool_finalize_decl): Update.
15666 (symbol_finalized): Remove.
15667 (symbol_finalized_and_needed): Rename to ...
15668 (symbol_defined_and_needed): ... update.
15669 (cgraph_analyze_functions): Update.
15670 (handle_alias_pairs): Update.
15671 (mark_functions_to_output): Update.
15672 (assemble_thunk): Update.
15673 (output_in_order): Update.
15674 (output_weakrefs): Update.
15675 (finalize_compilation_unit): Update.
15676 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15677 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15678 input_node, input_varpool_node): Update.
15679 * dbxout.c (dbxout_expand_expr): Update.
15680 * cgraphclones.c (cgraph_clone_node): Update.
15681 (cgraph_copy_node_for_versioning): Update.
15682 (cgraph_materialize_clone): Update.
15683 (cgraph_materialize_all_clones): Update.
15684 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15685 propagate_pure_const, propagate_nothrow): Update.
15686 * lto-streamer-out.c (lto_output, write_symbol): Update.
15687 * ipa-utils.c (ipa_reverse_postorder): Update.
15688 * ipa-inline.c (can_inline_edge_p): Update.
15689 (update_caller_keys, ipa_inline): Update.
15690 * dwarf2out.c (reference_to_unused,
15691 premark_types_used_by_global_vars_helper): Update.
15692 * tree-eh.c (tree_could_trap_p): Update.
15693 * ipa-split.c (consider_split, execute_split_functions): Update.
15694 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 15695 has_addr_references_p): Update; move ahead in file for better
15696 readability.
e70670cf
JH		 15697 (process_references): Simplify.
15698 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15699 bodies are removed.
15700 (cgraph_comdat_can_be_unshared_p): Make static.
15701 (cgraph_externally_visible_p): Update.
15702 (varpool_externally_visible_p): Update.
15703 (function_and_variable_visibility): Update.
15704 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15705 ipa_tm_mark_force_output_node): Update.
15706 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15707 estimate_edge_devirt_benefit, inline_generate_summary,
15708 inline_write_summary): Update.
15709 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15710 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 15711 (ipa_print_node_params, ipa_prop_read_section,
15712 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 15713 * varasm.c (mark_decl_referenced): Update.
15714 (assemble_alias, dump_tm_clone_pairs): Update.
15715 * tree-inline.c (copy_bb): Update.
15716 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15717 Update.
15718 * symtab.c (dump_symtab_base): Print new flags.
15719 (verify_symtab_base): Verify new flags.
15720 (symtab_alias_ultimate_target): New function.
15721 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15722 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15723 Update.
c1e183a9
UB		 15724 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15725 Update.
e70670cf
JH		 15726 * i386.c (ix86_get_function_versions_dispatcher,
15727 ix86_generate_version_dispatcher_body): Update.
15728 (fold_builtin_cpu): Use varpool_add_new_variable.
15729 * varpool.c (varpool_remove_initializer): Break out from ...
15730 (varpool_remove_node): ... this one.
15731 (dump_varpool_node, varpool_node_for_asm,
15732 cgraph_variable_initializer_availability, varpool_analyze_node,
15733 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 15734 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15735 Update.
e70670cf 15736
182802ad
JH		 157372013-05-29 Jan Hubicka <jh@suse.cz>
15738
15739 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15740
8190b609
ER		 157412013-05-29 Easwaran Raman <eraman@google.com>
15742
15743 PR tree-optimization/57442
15744 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15745 when control exits the main loop.
15746
69f5aa9b
SKS		 157472013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15748
15749 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15750 and RX600.
c1e183a9 15751 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 15752 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15753 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15754 option.
15755
4b847da9
BS		 157562013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15757
15758 PR tree-optimization/57441
15759 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15760 Don't limit size of incr_vec to number of candidates.
15761
4fc43c39
SE		 157622013-05-29 Steve Ellcey <sellcey@imgtec.com>
15763
15764 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15765 and mips16 directories.
c1e183a9 15766 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 15767 (MULTILIB_DIRNAMES): Ditto.
15768 (MULTILIB_EXCEPTIONS): Add new exceptions.
15769 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15770 (MULTILIB_DIRNAMES): Ditto.
15771 (MULTILIB_EXCEPTIONS): Add new exceptions.
15772
12211b99 157732012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 15774 Marcus Shawcroft <marcus.shawcroft@arm.com>
15775
15776 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15777 SYMBOL_TINY_ABSOLUTE.
15778 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15779 SYMBOL_TINY_ABSOLUTE.
15780 (aarch64_expand_mov_immediate): Likewise.
15781 (aarch64_classify_symbol): Likewise.
15782 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15783 Permit SYMBOL_TINY_ABSOLUTE.
15784 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15785
12211b99 157862013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 15787 Marcus Shawcroft <marcus.shawcroft@arm.com>
15788
15789 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15790 Refactor if/switch. Replace gcc_assert with if.
15791
c0186656
GG		 157922013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15793
15794 * config/i386/i386.c (initial_ix86_tune_features): Enable
15795 FP Reassociation for AMD bdver1 and bdver2.
15796
d20188f3
MJ		 157972013-05-29 Martin Jambor <mjambor@suse.cz>
15798
15799 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15800 and IMAGPART_EXPR do not occur within other handled_components.
15801
292cba13
RB		 158022013-05-29 Richard Biener <rguenther@suse.de>
15803
15804 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15805 access on whether the use is in the BB we currently try to
15806 vectorize.
15807 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15808 vectorize to vect_bb_slp_scalar_cost.
15809
6eddf228
RB		 158102013-05-29 Richard Biener <rguenther@suse.de>
15811
15812 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15813 computing scalar cost offsetted by stmts that are kept live
15814 by scalar uses.
15815 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15816 for computation of scalar cost.
15817
7df36117
SE		 158182013-05-28 Steve Ellcey <sellcey@mips.com>
15819
15820 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15821
36536d79
BI		 158222013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15823
15824 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15825 array notation built-in reduction functions.
15826 * doc/passes.texi (Passes): Added documentation about changes done
15827 for Cilk Plus.
15828 * doc/invoke.texi (C Dialect Options): Added documentation about
15829 the -fcilkplus flag.
15830 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15831 (BUILTINS_DEF): Depend on cilkplus.def.
15832 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15833 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15834 * cilkplus.def: New file.
15835
12211b99 158362013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 15837
15838 PR rtl-optimization/57439
15839 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15840
9daf6dbb
ER		 158412013-05-28 Easwaran Raman <eraman@google.com>
15842
15843 PR tree-optimization/57337
15844 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15845 (find_insert_point): Correctly identify the insertion point
15846 when two statements with the same UID is compared.
15847
fbd7e877
RB		 158482013-05-28 Richard Biener <rguenther@suse.de>
15849
15850 PR tree-optimization/56787
15851 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15852 from the list of data references.
15853 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15854 clobbers.
15855 (vect_analyze_loop_operations): Likewise.
15856 (vect_transform_loop): Remove clobbers.
15857
bbba1117
MJ		 158582013-05-28 Martin Jambor <mjambor@suse.cz>
15859
15860 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15861 and REALPART_EXPRs have scalar type.
15862
bd388c2a
RB		 158632013-05-28 Richard Biener <rguenther@suse.de>
15864
15865 PR tree-optimization/57411
15866 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15867 virtual operands.
15868 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15869 virtual operand propagation.
15870
2f56a311
EB		 158712013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15872
15873 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15874 destination register for bmasksi_vis.
15875 (vector_init_bshuffle): Likewise.
15876 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15877
5c3eacbb
EB		 158782013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15879
15880 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15881 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15882 mode if the instruction isn't available in the original mode.
15883 * config/sparc/sparc.opt (mfix-ut699): New option.
15884 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15885 (divdf3): Turn into expander.
15886 (divdf3_nofix): New insn.
15887 (divdf3_fix): Likewise.
15888 (divsf3): Disable if -mfix-ut699.
15889 (sqrtdf2): Turn into expander.
15890 (sqrtdf2_nofix): New insn.
15891 (sqrtdf2_fix): Likewise.
15892 (sqrtsf2): Disable if -mfix-ut699.
15893
a1756c0a
RB		 158942013-05-27 Richard Biener <rguenther@suse.de>
15895
15896 PR middle-end/57412
15897 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15898 block for the new loop.
15899
5a892248
RB		 159002013-05-27 Richard Biener <rguenther@suse.de>
15901
15902 PR tree-optimization/57343
15903 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15904 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15905 (number_of_iterations_cond): Do not build the folded tree.
15906
d1de852b
RB		 159072013-05-27 Richard Biener <rguenther@suse.de>
15908
15909 Revert
15910 PR middle-end/57381
15911 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15912 OEP_CONSTANT_ADDRESS_OF retained.
15913
15914 PR tree-optimization/57417
15915 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15916 for unchanged base.
15917 (set_ssa_val_to): Compare addresses using
15918 get_addr_base_and_unit_offset.
15919
12211b99 159202013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 15921
15922 PR rtl-optimization/56833
15923 * postreload.c (move2add_record_mode): New function.
15924 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15925 (move2add_use_add2_insn): Use move2add_record_sym_value.
15926 (move2add_use_add3_insn): Likewise.
15927 (reload_cse_move2add): Use move2add_valid_value_p and
15928 move2add_record_mode. Invalidate call-clobbered and REG_INC
15929 affected regs by setting reg_mode to VOIDmode.
15930 (move2add_note_store): Don't pretend the inside of a SUBREG is
15931 the actual destination. Invalidate single/leading registers by
15932 setting reg_mode to VOIDmode.
15933 Use move2add_record_sym_value, move2add_valid_value_p and
15934 move2add_record_mode.
15935
b03be25f
RB		 159362013-05-27 Richard Biener <rguenther@suse.de>
15937
15938 PR tree-optimization/57396
15939 * tree-affine.c (double_int_constant_multiple_p): Properly
15940 return false for val == 0 and div != 0.
15941
44e88db2
RS		 159422013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15943
15944 * config/mips/mips.h: Use #elif in preprocessor conditions.
15945
3b859704
RS		 159462013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15947
15948 PR target/53916
15949 * config/mips/constraints.md (kl): New constraint.
15950 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15951 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15952 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15953 from LO for MIPS16.
15954 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15955
c979d5f5
RS		 159562013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15957
15958 PR target/55777
15959 * config/mips/mips.c (mips_can_inline_p): New function.
15960 (TARGET_CAN_INLINE_P): Define.
15961
8e90de43
SB		 159622013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15963
15964 * sched-int.h (ds_t, dw_t): Make unsigned int.
15965 Fix documentation that describes how all the ds_t bits are used.
15966 Reserve the last bit for delayed-branch scheduling.
15967 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15968 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15969 (gen_dep_weak_1): Remove prototype.
15970 * sched-deps.c (get_dep_weak_1): Make static.
15971 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15972 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15973 * doc/tm.texi: Regenerate.
15974 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15975
cb5cb194
SB		 159762013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15977
15978 PR debug/56950
15979 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15980
5bd93ff6
NS		 159812013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15982 Sandra Loosemore <sandra@codesourcery.com>
15983
15984 * config.gcc (powerpc-*): Allow native for with-cpu.
15985
2343af65
JL		 159862013-05-24 Jeff Law <law@redhat.com>
15987
15988 PR tree-optimization/57124
15989 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15990 conversion feeding a condition if the range has an overflow
15991 if -fstrict-overflow. Add warnings for when we do make the
15992 transformation.
15993
3ad695b9
DC		 159942013-05-24 Dehao Chen <dehao@google.com>
15995
04960246 15996 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 15997 (locus_discrim_hasher): Likewise.
15998 (locus_discrim_hasher::hash): Likewise.
15999 (locus_discrim_hasher::equal): Likewise.
16000
82338059
MJ		 160012013-05-24 Martin Jambor <mjambor@suse.cz>
16002
16003 PR tree-optimization/57294
16004 * cgraph.h (ipa_record_stmt_references): Declare.
16005 * cgraphbuild.c (ipa_record_stmt_references): New function.
16006 (build_cgraph_edges): Use ipa_record_stmt_references.
16007 (rebuild_cgraph_edges): Likewise.
16008 (cgraph_rebuild_references): Likewise.
16009 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16010 associated with the old statement and build references from the
16011 newly built statements.
16012 * ipa-ref.c (ipa_remove_stmt_references): New function.
16013 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16014
1ccd4874
VM		 160152013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16016
55805e54 16017 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 16018 mem-mem moves.
55805e54 16019 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 16020 too.
55805e54 16021 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 16022 secondary memory moves.
16023 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16024 reg set up in the current insn.
16025
25e25c73
DC		 160262013-05-24 Dehao Chen <dehao@google.com>
16027
04960246 16028 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 16029 hash function.
16030 (locus_descrim_hasher::equal): Likewise.
04960246 16031 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 16032 (make_edges): Likewise.
16033 (next_discriminator_for_locus): Likewise.
16034 (same_line_p): Likewise.
16035 (assign_discriminators): Likewise.
16036 (make_cond_expr_edges): Likewise.
16037 (make_gimple_switch_edges): Likewise.
16038 (make_goto_expr_edges): Likewise.
16039 (make_gimple_asm_edges): Likewise.
16040
50d38551
IB		 160412013-05-24 Ian Bolton <ian.bolton@arm.com>
16042
16043 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16044 X format specifier to only display bottom 16 bits.
16045 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16046 immediate to match for operand 2, since it will be masked.
16047
aea0101d
RB		 160482013-05-24 Richard Biener <rguenther@suse.de>
16049
16050 PR tree-optimization/57287
16051 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16052 all SSA names that occur in abnormal PHIs.
16053
634e03d3
AI		 160542013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16055
16056 PR tree-ssa/57385
16057 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16058 that index is not negative.
16059
b17c024f
EB		 160602013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16061
16062 PR rtl-optimization/55177
16063 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16064 (simplify_byte_swapping_operation): New.
16065 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16066 (simplify_relational_operation_1): Deal with BSWAP.
16067
46aeac1b
RH		 160682013-05-23 Richard Henderson <rth@redhat.com>
16069
16070 PR target/56742
16071 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16072 (ix86_reorg): Call it.
16073
70cc1536
UB		 160742013-05-23 Uros Bizjak <ubizjak@gmail.com>
16075
16076 PR target/57379
16077 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16078 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16079 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16080
b3851501
CB		 160812013-05-23 Christian Bruel <christian.bruel@st.com>
16082
16083 PR debug/57351
16084 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16085
12211b99 160862013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 16087 Marcus Shawcroft <marcus.shawcroft@arm.com>
16088
16089 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16090 * config/aarch64/constraints.md (Usa): Remove.
16091 * doc/md.texi (AArch64 Usa): Remove.
16092
12211b99 160932013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 16094 Marcus Shawcroft <marcus.shawcroft@arm.com>
16095
16096 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16097 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16098 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16099 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16100
12211b99 161012013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 16102
16103 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16104 instruction (AdvSIMD).
16105 * config/aarch64/aarch64-builtins.c
16106 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16107 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16108
76c36cb1 161092013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 16110
16111 PR middle-end/57347
16112 * tree.h (contains_bitfld_component_ref_p): Declare.
16113 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 16114 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16115 caller.
7d2fb524
MJ		 16116 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16117 not access a bit-field. Assert all final offsets are byte-aligned.
16118
ce521ae6
RB		 161192013-05-23 Richard Biener <rguenther@suse.de>
16120
16121 PR tree-optimization/57380
16122 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16123 least one invariant or re-used load.
16124 * passes.c (init_optimization_passes): Move pass_phiprop before
16125 pass_forwprop.
16126
75c7257f
JG		 161272013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16128
16129 * config/aarch64/aarch64-simd.md
16130 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16131
af6d44b5
RB		 161322013-05-23 Richard Biener <rguenther@suse.de>
16133
16134 PR middle-end/57381
16135 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16136 OEP_CONSTANT_ADDRESS_OF retained.
16137
bd3647bf
JJ		 161382013-05-23 Jakub Jelinek <jakub@redhat.com>
16139
16140 PR middle-end/57344
70cc1536
UB		 16141 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16142 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 16143
a55757ea
RB		 161442013-05-23 Richard Biener <rguenther@suse.de>
16145
16146 PR rtl-optimization/57341
16147 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16148 instead of true_dependence.
16149
cfb00b41
DM		 161502013-05-22 David Malcolm <dmalcolm@redhat.com>
16151
dd1c676f
DM		 16152 * bb-reorder.c (branch_threshold): Make const.
16153 (exec_threshold): Ditto.
cfb00b41 16154
f62511da 161552013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16156 Pat Haugen <pthaugen@us.ibm.com>
16157 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16158
16159 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16160 documentation for the power8 crypto builtins.
16161
16162 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16163
16164 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16165 macros for defining power8 builtin functions.
16166 (BU_P8V_AV_2): Likewise.
16167 (BU_P8V_AV_P): Likewise.
16168 (BU_P8V_VSX_1): Likewise.
16169 (BU_P8V_OVERLOAD_1): Likewise.
16170 (BU_P8V_OVERLOAD_2): Likewise.
16171 (BU_CRYPTO_1): Likewise.
16172 (BU_CRYPTO_2): Likewise.
16173 (BU_CRYPTO_3): Likewise.
16174 (BU_CRYPTO_OVERLOAD_1): Likewise.
16175 (BU_CRYPTO_OVERLOAD_2): Likewise.
16176 (XSCVSPDP): Fix typo, point to the correct instruction.
16177 (VCIPHER): Add power8 crypto builtins.
16178 (VCIPHERLAST): Likewise.
16179 (VNCIPHER): Likewise.
16180 (VNCIPHERLAST): Likewise.
16181 (VPMSUMB): Likewise.
16182 (VPMSUMH): Likewise.
16183 (VPMSUMW): Likewise.
16184 (VPERMXOR_V2DI): Likewise.
16185 (VPERMXOR_V4SI: Likewise.
16186 (VPERMXOR_V8HI: Likewise.
16187 (VPERMXOR_V16QI: Likewise.
16188 (VSHASIGMAW): Likewise.
16189 (VSHASIGMAD): Likewise.
16190 (VPMSUM): Likewise.
16191 (VPERMXOR): Likewise.
16192 (VSHASIGMA): Likewise.
16193
16194 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16195 __CRYPTO__ if the crypto instructions are available.
16196 (altivec_overloaded_builtins): Add support for overloaded power8
16197 builtins.
16198
16199 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16200 support for power8 crypto builtins.
16201 (builtin_function_type): Likewise.
16202 (altivec_init_builtins): Add support for builtins that take vector
16203 long long (V2DI) arguments.
16204
16205 * config/rs6000/crypto.md: New file, define power8 crypto
16206 instructions.
16207
162082013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16209 Pat Haugen <pthaugen@us.ibm.com>
16210 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16211
16212 * doc/invoke.texi (Option Summary): Add power8 options.
16213 (RS/6000 and PowerPC Options): Likewise.
16214
16215 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16216 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16217 wm, wn, wr documentation.
16218
f43245d1 16219 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM		 16220 registers if direct move instructions are enabled.
16221 (wn): New constraint for no registers.
16222 (wq): New constraint for quad word even GPR registers.
16223 (wr): New constraint if 64-bit instructions are enabled.
16224 (wv): New constraint if power8 vector instructions are enabled.
16225 (wQ): New constraint for quad word memory locations.
16226
f43245d1 16227 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM		 16228 constraint for 0..15 for crypto instructions.
16229 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16230 as GPR and floating point registers.
16231 (int_reg_operand): New predicate to match only GPR registers.
16232 (base_reg_operand): New predicate to match base registers.
16233 (quad_int_reg_operand): New predicate to match even GPR registers
16234 for quad memory operations.
16235 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16236 operations in both GPR and VSX registers.
16237 (quad_memory_operand): New predicate for quad memory operations.
16238 (reg_or_indexed_operand): New predicate for direct move support.
16239
f43245d1 16240 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM		 16241 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16242 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16243 (POWERPC_MASKS): Add power8 options.
16244 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16245 various options.
16246
f43245d1 16247 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM		 16248 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16249
f43245d1 16250 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM		 16251 (-mpower8-fusion): New power8 options.
16252 (-mpower8-fusion-sign): Likewise.
16253 (-mpower8-vector): Likewise.
16254 (-mcrypto): Likewise.
16255 (-mdirect-move): Likewise.
16256 (-mquad-memory): Likewise.
16257
f43245d1 16258 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM		 16259 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16260 registers.
70cc1536 16261 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 16262 (rs6000_debug_vector_unit): Add p8_vector.
16263 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16264 definitions. Also print fusion state.
16265 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16266 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16267 (rs6000_option_override_internal): Add support for power8.
16268 (rs6000_common_init_builtins): Add debugging for skipped builtins
16269 if -mdebug=builtin.
16270 (rs6000_adjust_cost): Add power8 support.
16271 (rs6000_issue_rate): Likewise.
16272 (insn_must_be_first_in_group): Likewise.
16273 (insn_must_be_last_in_group): Likewise.
16274 (force_new_group): Likewise.
16275 (rs6000_register_move_cost): Likewise.
16276 (rs6000_opt_masks): Likewise.
16277
16278 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16279 power8 capable assembler, default to power7 options.
16280 (TARGET_DIRECT_MOVE): Likewise.
16281 (TARGET_CRYPTO): Likewise.
16282 (TARGET_P8_VECTOR): Likewise.
16283 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16284 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16285 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16286 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16287 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16288 (TARGET_XSCVDPSPN): Likewise.
16289 (TARGET_XSCVSPDPN): Likewsie.
16290 (TARGET_SYNC_HI_QI): Likewise.
16291 (TARGET_SYNC_TI): Likewise.
16292 (MASK_CRYPTO): Likewise.
16293 (MASK_DIRECT_MOVE): Likewise.
16294 (MASK_P8_FUSION): Likewise.
16295 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 16296 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16297 temporary used by some of the direct move instructions to get two FP
16298 temporary registers does not force creation of a stack frame.
f62511da
MM		 16299 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16300 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16301 that any VSX registers are tieable, even if they are also an
16302 Altivec vector mode.
16303 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16304 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16305 (RS6000_BTM_CRYPTO): Likewise.
16306 (RS6000_BTM_COMMON): Likewise.
16307
16308 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16309 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16310 (enum rs6000_vector): Add power8 vector support.
16311
73a1a707
RR		 163122013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16313
16314 PR target/19599
16315 PR target/57340
16316 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16317 (any_sibcall_could_use_r3): this and handle indirect calls.
16318 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16319
d28073d4
BS		 163202013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16321
16322 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16323
bee0b10c
RB		 163242013-05-22 Richard Biener <rguenther@suse.de>
16325
16326 PR middle-end/57349
16327 * profile.c (branch_prob): Do not split blocks that are
16328 abnormally receiving from ECF_RETURNS_TWICE functions.
16329
98409b51
RS		 163302013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16331
16332 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16333 address mode. Move pointer mode initialization to the same place.
16334
c0602ab8
MZ		 163352013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16336
16337 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16338 while it has any effect.
16339
4f7a634e
ER		 163402013-05-21 Easwaran Raman <eraman@google.com>
16341
16342 PR tree-optimization/57322
9daf6dbb
ER		 16343 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16344 UID of the statement added to the BB to be 1.
4f7a634e 16345
a4ce1258
JJ		 163462013-05-21 Jakub Jelinek <jakub@redhat.com>
16347
16348 PR tree-optimization/57331
70cc1536
UB		 16349 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16350 of conversion from pointer type to integral type with integer.
a4ce1258 16351
1b14621a
MJ		 163522013-05-21 Martin Jambor <mjambor@suse.cz>
16353
16354 PR lto/57289
16355 * ipa-prop.c (ipa_read_node_info): Process param_used and
16356 controlled_uses in the same order as when writing.
16357
e60661f0
MG		 163582013-05-21 Magnus Granberg <baldrick@free.fr>
16359
16360 PR plugins/56754
ee49aa34 16361 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16362
4a61be9a
RB		 163632013-05-21 Richard Biener <rguenther@suse.de>
16364
16365 PR tree-optimization/57318
16366 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16367 estimate stmts with side-effects as likely eliminated.
16368
c52da5f7
RB		 163692013-05-21 Richard Biener <rguenther@suse.de>
16370
16371 PR tree-optimization/57330
16372 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16373 preserve the call stmts fntype.
16374
7ec67e2a
RB		 163752013-05-21 Richard Biener <rguenther@suse.de>
16376
16377 PR tree-optimization/57303
16378 * tree-ssa-sink.c (statement_sink_location): Improve killing
16379 stmt detection and properly handle self-assignments.
16380
b112d513
CB		 163812013-05-21 Christian Bruel <christian.bruel@st.com>
16382
55805e54
YR		 16383 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16384 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 16385 registers. Set register size out of the PARALLEL loop.
16386
14c2ec26
OE		 163872013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16388
16389 PR target/56547
16390 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16391 (*fmasf4, *fmasf4_media): New insns.
16392
da734fa1
RS		 163932013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16394
16395 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16396 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16397 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16398 (mips_idiv_insns): Update the comments to say that the returned
16399 instruction counts are in units of BASE_INSN_LENGTH.
16400 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16401 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16402 using 2 rather than 4 as the length of indirect MIPS16 and
16403 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16404 length of a NOP. Don't divide MIPS16 lengths by 2.
16405 (mips16_split_long_branches): Assume a branch is long if the
16406 length is greater than 4 rather than 8.
16407 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16408 rather than multiplying them by 2. Multiply instruction counts
16409 by BASE_INSN_LENGTH rather than 4.
16410 (*jump_mips16, tls_get_tp_mips16_<mode>)
16411 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16412
13719e8b
RS		 164132013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16414
16415 * config/mips/mips.md (extended_mips16): Remove branch case.
16416 (length): Remove duplicated extended_mips16 test.
16417
c3850d14
RS		 164182013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16419
16420 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16421
8da2e059
RS		 164222013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16423
16424 * recog.h (Recog_data): Rename to...
16425 (recog_data_d): ...this.
16426 (recog_data): Update accordingly.
16427 * recog.c (recog_data): Likewise.
16428 * reload.c (save_recog_data): Likewise.
16429 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16430 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16431
2df013f3
JB		 164322013-05-17 Julian Brown <julian@codesourcery.com>
16433
16434 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16435 found in a REG_EQUAL note, invalidate it.
16436
371e77e3 164372013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 16438
16439 * tree-ssa-reassoc.c (find_insert_point): New function.
16440 (insert_stmt_after): Likewise.
16441 (get_def_stmt): Likewise.
16442 (ensure_ops_are_available): Likewise.
16443 (not_dominated_by): Likewise.
16444 (rewrite_expr_tree): Do not move statements beyond what is
16445 necessary. Remove call to swap_ops_for_binary_stmt...
16446 (reassociate_bb): ... and move it here.
16447 (build_and_add_sum): Assign UIDs for new statements.
16448 (linearize_expr): Likewise.
16449 (do_reassoc): Renumber gimple statement UIDs.
16450
e01c7cca
JH		 164512013-05-17 Jan Hubicka <jh@suse.cz>
16452
16453 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16454 weakrefs.
16455 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 16456 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16457 weakrefs.
e01c7cca
JH		 16458 (output_weakrefs): Update.
16459
c3272a92
PCC		 164602013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16461 Martin Jambor <mjambor@suse.cz>
16462
16463 PR middle-end/57276
16464 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16465 value that corresponds to the given aggval is found in values vector.
16466
11c2aa39
UB		 164672013-05-17 Uros Bizjak <ubizjak@gmail.com>
16468
16469 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16470 sse, sse2, sse3, ssse3 and sse4a flags to options.
16471
abecc8c6
DM		 164722013-05-17 David Malcolm <dmalcolm@redhat.com>
16473
16474 * gengtype-state.c: (s_expr_writer): New class, to handle
16475 prettifying of output layout of s-expressions.
16476 (state_writer): New class, to write out gtype.state.
16477 (state_written_type_count): Move this variable into member data of
16478 state_writer.
16479 (s_expr_writer::s_expr_writer): New code: constructor for new class
16480 (state_writer::state_writer(): ditto
16481 (s_expr_writer::write_new_line): New function
16482 (s_expr_writer::write_any_indent): ditto
16483 (s_expr_writer::begin_s_expr): ditto
16484 (s_expr_writer::end_s_expr): ditto
16485 (write_state_fileloc): convert to method of state_writer...
16486 (state_writer:: write_state_fileloc): ...and use methods of
16487 s_expr_writer to write indentation into the gtype.state output file
16488 to visually represent the hierarchical structure of the list
16489 structures
16490 (write_state_fields): ditto, renaming to...
16491 (state_writer::write_state_fields)
16492 (write_state_a_string): ditto, renaming to...
16493 (state_writer::write_state_a_string)
16494 (write_state_string_option): ditto, renaming to...
16495 (state_writer::write_state_string_option)
16496 (write_state_type_option): ditto, renaming to...
16497 (state_writer::write_state_type_option)
16498 (write_state_nested_option): ditto, renaming to...
16499 (state_writer::write_state_nested_option)
16500 (write_state_option): ditto, renaming to...
16501 (state_writer::write_state_option)
16502 (write_state_options): ditto, renaming to...
16503 (state_writer::write_state_options)
16504 (write_state_lang_bitmap): ditto, renaming to...
16505 (state_writer::write_state_lang_bitmap)
16506 (write_state_version): ditto, renaming to...
16507 (state_writer::write_state_version)
16508 (write_state_scalar_type): ditto, renaming to...
16509 (state_writer::write_state_scalar_type)
16510 (write_state_string_type): ditto, renaming to...
16511 (state_writer::write_state_string_type)
16512 (write_state_undefined_type): ditto, renaming to...
16513 (state_writer::write_state_undefined_type)
16514 (write_state_struct_union_type): ditto, renaming to...
16515 (state_writer::write_state_struct_union_type)
16516 (write_state_struct_type): ditto, renaming to...
16517 (state_writer::write_state_struct_type)
16518 (write_state_user_struct_type): ditto, renaming to...
16519 (state_writer::write_state_user_struct_type)
16520 (write_state_lang_struct_type): ditto, renaming to...
16521 (state_writer::write_state_lang_struct_type)
16522 (write_state_param_struct_type): ditto, renaming to...
16523 (state_writer::write_state_param_struct_type)
16524 (write_state_pointer_type): ditto, renaming to...
16525 (state_writer::write_state_pointer_type)
16526 (write_state_array_type): ditto, renaming to...
16527 (state_writer::write_state_array_type)
16528 (write_state_gc_used): ditto, renaming to...
16529 (state_writer::write_state_gc_used)
16530 (write_state_common_type_content): ditto, renaming to...
16531 (state_writer::write_state_common_type_content)
16532 (write_state_type): ditto, renaming to...
16533 (state_writer::write_state_type)
16534 (write_state_pair_list): ditto, renaming to...
16535 (state_writer::write_state_pair_list)
16536 (write_state_pair): ditto, renaming to...
16537 (state_writer::write_state_pair)
16538 (write_state_typedefs): ditto, renaming to...
16539 (state_writer::write_state_typedefs)
16540 (write_state_structures): ditto, renaming to...
16541 (state_writer::write_state_structures)
16542 (write_state_param_structs): ditto, renaming to...
16543 (state_writer::write_state_param_structs)
16544 (write_state_variables): ditto, renaming to...
16545 (state_writer::write_state_variables)
16546 (write_state_srcdir): ditto, renaming to...
16547 (state_writer::write_state_srcdir)
16548 (write_state_files_list): ditto, renaming to...
16549 (state_writer::write_state_files_list)
16550 (write_state_languages): ditto, renaming to...
16551 (state_writer::write_state_languages)
16552 (write_state): create a state_writer instance and use it when
16553 writing out the state file
16554
d6545f29
MS		 165552013-05-17 Mike Stump <mikestump@comcast.net>
16556
816a3f73 16557 PR rtl-optimization/57304
d6545f29
MS		 16558 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16559 accessing DF_REF_REAL_LOC.
16560
38c821cf
JJ		 165612013-05-17 Jakub Jelinek <jakub@redhat.com>
16562
16563 PR rtl-optimization/57281
16564 PR rtl-optimization/57300
16565 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16566 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16567 what the other splitter did if the registers are dead.
16568
2a293391
RB		 165692013-05-17 Richard Biener <rguenther@suse.de>
16570
16571 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16572 MEM_REF offsets.
16573
17042d2b
JJ		 165742013-05-17 Jakub Jelinek <jakub@redhat.com>
16575
16576 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16577 linking.
16578
5b115c1f
MP		 165792013-05-17 Marek Polacek <polacek@redhat.com>
16580
11c2aa39
UB		 16581 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16582 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 16583
52d84413
JJ		 165842013-05-17 Jakub Jelinek <jakub@redhat.com>
16585
68119618
JJ		 16586 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16587 vect_external_def oprnd1 with loop_vinfo, try to emit
16588 optional cast, negation and and stmts on the loop preheader
16589 edge instead of into the pattern def seq.
16590
52d84413
JJ		 16591 PR tree-optimization/57051
16592 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16593 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16594
17b962bd
NC		 165952013-05-16 Nick Clifton <nickc@redhat.com>
16596
16597 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16598 (rl78_is_naked_func): New function.
56aefbf7
UB		 16599 (rl78_expand_prologue): Skip prologue generation for naked functions.
16600 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 16601 * doc/extend.texi (naked): Add RL78 to the list of processors
16602 that supports this attribute.
16603
b1a0f84e
JL		 166042013-05-16 Jeff Law <law@redhat.com>
16605
16606 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16607
af0e415b
UB		 166082013-05-16 Uros Bizjak <ubizjak@gmail.com>
16609
16610 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16611 cache parameters using detect_caches_amd also for CYRIX,
16612 NSC and TM2 signatures.
16613
19db293a
UB		 166142013-05-16 Uros Bizjak <ubizjak@gmail.com>
16615 Dzianis Kahanovich <mahatma@eu.by>
16616
16617 PR target/45359
16618 PR target/46396
16619 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16620 VIA/Centaur processors and determine their cache parameters
16621 using detect_caches_amd.
16622
251a41b9
TJ		 166232013-05-16 Teresa Johnson <tejohnson@google.com>
16624
16625 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16626 (rtl_verify_edges): New function.
16627 (rtl_verify_bb_insns): Ditto.
16628 (rtl_verify_bb_pointers): Ditto.
16629 (rtl_verify_bb_insn_chain): Ditto.
16630 (rtl_verify_fallthru): Ditto.
16631 (rtl_verify_bb_layout): Ditto.
16632 (rtl_verify_flow_info_1): Outline checks into new functions.
16633 (rtl_verify_flow_info): Ditto.
16634
f14540b6
SE		 166352013-05-16 Steve Ellcey <sellcey@imgtec.com>
16636
16637 * cfghooks.c (copy_bbs): Add update_dominance argument.
16638 * cfghooks.h (copy_bbs): Update prototype.
16639 * tree-cfg.c (gimple_duplicate_sese_region):
16640 Add update_dominance argument.
16641 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16642 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16643 gimple_duplicate_sese_region call.
16644 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16645 Update copy_bbs call.
16646 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16647 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16648
7e9a3abb
JJ		 166492013-05-16 Jakub Jelinek <jakub@redhat.com>
16650
16651 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16652 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16653 vect_recog_rotate_pattern.
16654 (vect_recog_rotate_pattern): New function.
16655
427b248d
JM		 166562013-05-16 Jason Merrill <jason@redhat.com>
16657
16658 * Makefile.in (LLINKER): New variable.
16659 (mostlyclean): Remove link mutex.
16660 * configure.ac: Handle --enable-link-mutex.
16661 * lock-and-run.sh: New script.
16662
b871e3d2
RR		 166632013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16664
16665 PR target/19599
16666 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16667 for NULL decl.
16668
ce858126
RO		 166692013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16670
16671 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16672
798d3d04
GY		 166732013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16674
16675 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16676 * config/arm/arm.c (next_consecutive_mem): New function.
16677 (gen_movmem_ldrd_strd): Likewise.
16678 * config/arm/arm.md (movmemqi): Update condition and code.
16679 (unaligned_loaddi, unaligned_storedi): New patterns.
16680
0baddc45
RO		 166812013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16682
16683 * config.gcc: Obsolete *-*-solaris2.9*.
16684 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16685
aa9d5bcf
RB		 166862013-05-16 Richard Biener <rguenther@suse.de>
16687
16688 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16689 earlier, after GRAPHITE transforms and IV canonicalization.
16690
5a0f4dd3
JJ		 166912013-05-16 Jakub Jelinek <jakub@redhat.com>
16692
16693 * omp-low.c (extract_omp_for_data): For collapsed loops,
16694 if at least one of the loops is known at compile time to
16695 iterate zero times, set count to 0.
16696 (expand_omp_regimplify_p): New function.
16697 (expand_omp_for_generic): For collapsed loops, if at least
16698 one of the loops isn't known to iterate at least once,
16699 add runtime check with setting count to 0.
16700 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16701 For unsigned types if it isn't known at compile time that
16702 the loop will iterate at least once, add runtime check to bypass
16703 the whole loop if initial condition isn't true.
16704
e3753785
NS		 167052013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16706
16707 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16708
43bb4dd1
MG		 167092013-05-16 Marc Glisse <marc.glisse@inria.fr>
16710
16711 PR middle-end/57286
16712 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16713 transformations to avoid an infinite loop.
16714
3571dde6
MP		 167152013-05-16 Marek Polacek <polacek@redhat.com>
16716
16717 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16718
b8b3f0ca
LE		 167192013-05-15 Leif Ekblad <leif@rdos.net>
16720
16721 * config/i386/i386.c (ix86_decompose_address): Use
16722 DEFAULT_TLS_SEG_REG to access TLS segment register.
16723 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16724 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16725 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16726
e299a383
RS		 167272013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16728
16729 PR target/57260
16730 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16731 sibling calls to functions that would normally be lazily bound,
16732 unless $gp is call-clobbered.
16733
e7208ea3
UB		 167342013-05-15 Uros Bizjak <ubizjak@gmail.com>
16735
19db293a 16736 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 16737 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 16738 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16739 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 16740 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16741 of TARGET_3DNOW.
16742 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16743
e7413f3d
AS		 167442013-05-15 Andreas Schwab <schwab@suse.de>
16745
16746 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16747 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16748 third operand.
16749
ad4db775
TJ		 167502013-05-15 Teresa Johnson <tejohnson@google.com>
16751
16752 * loop-unroll.c (report_unroll_peel): Check decision before
16753 emitting unroll/peel message.
16754
af205f67
TJ		 167552013-05-15 Teresa Johnson <tejohnson@google.com>
16756
16757 * function.h (has_bb_partition): New rtl_data flag.
16758 (bb_reorder_complete): Ditto.
16759 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16760 instead of flag_reorder_blocks_and_partition.
16761 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16762 with some enhancements.
16763 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16764 * bb-reorder.c (connect_traces): Check for has_bb_partition
16765 instead of flag_reorder_blocks_and_partition.
16766 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16767 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16768 verify_hot_cold_block_grouping.
16769 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16770
9adcfa3c
RR		 167712013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16772
16773 PR target/19599
16774 * config/arm/predicates.md (call_insn_operand): New predicate.
16775 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16776 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16777 if insn is not a tail call.
16778 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16779 registers.
16780 * config/arm/arm.h (enum reg_class): New caller save register class.
16781 (REG_CLASS_NAMES): Likewise.
16782 (REG_CLASS_CONTENTS): Likewise.
16783 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16784 without decls.
16785
ccb3ad87
RB		 167862013-05-15 Richard Biener <rguenther@suse.de>
16787
16788 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16789 of MSG_OPTIMIZED_LOCATIONS.
16790 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16791 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16792 message.
16793 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16794 of MSG_OPTIMIZED_LOCATIONS.
16795 (execute_vect_slp): Likewise.
16796 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16797 (vect_create_cond_for_alias_checks): Likewise.
16798 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16799 (vect_recog_widen_mult_pattern): Likewise.
16800 (vect_recog_widen_sum_pattern): Likewise.
16801 (vect_recog_over_widening_pattern): Likewise.
16802 (vect_recog_widen_shift_pattern): Likewise.
16803 (vect_recog_vector_vector_shift_pattern): Likewise.
16804 (vect_recog_divmod_pattern): Likewise.
16805 (vect_recog_mixed_size_cond_pattern): Likewise.
16806 (vect_recog_bool_pattern): Likewise.
16807 (vect_pattern_recog_1): Likewise.
16808
48b1474e
MJ		 168092013-05-15 Martin Jambor <mjambor@suse.cz>
16810
16811 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16812 non-functions to builtin_unreachable.
16813 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16814 correct when new direct edges were discovered.
16815
9de04252
MJ		 168162013-05-15 Martin Jambor <mjambor@suse.cz>
16817
16818 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16819 header, print symbol order instead of node uid, print more information
16820 about indirect edge targets.
16821 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16822 uids.
16823 (ipa_make_edge_direct_to_target): Likewise.
16824 (remove_described_reference): Likewise.
16825 (propagate_controlled_uses): Likewise.
16826 (ipa_print_node_params): Also print symbol order.
16827 (ipcp_transform_function): Print symbol order instead of node uids.
16828 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16829 (cgraph_get_create_real_symbol_node): Likewise.
16830 * ipa-cp.c (print_lattice): Likewise.
16831 (print_all_lattices): Likewise.
16832 (determine_versionability): Likewise.
16833 (initialize_node_lattices): Likewise.
16834 (estimate_local_effects): Likewise.
16835 (update_profiling_info): Likewise.
16836 (create_specialized_node): Likewise.
16837 (perhaps_add_new_callers): Likewise.
16838 (decide_about_value): Likewise.
16839 (decide_whether_version_node): Likewise.
16840 (identify_dead_nodes): Likewise.
16841 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16842 (dump_inline_summary): Likewise.
16843 (estimate_node_size_and_time): Likewise.
16844 (inline_analyze_function): Likewise.
16845 * ipa-inline.c (report_inline_failed_reason): Likewise.
16846 (want_early_inline_function_p): Likewise.
16847 (edge_badness): Likewise.
16848 (update_edge_key): Likewise.
16849 (inline_small_functions): Likewise. Add dumping of order to two other
16850 dumps.
16851 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16852 instead of node uids.
16853 (propagate_pure_const): Likewise.
16854 (propagate_pure_const): Likewise.
16855 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16856 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16857 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16858 of node uids.
16859 * tree-pretty-print.c (dump_function_header): Likewise.
16860 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16861 Print symbol order instead of node uids.
16862
1dd03b91
AK		 168632013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16864
16865 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16866 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16867
fdf6a7b9
RB		 168682013-05-15 Richard Biener <rguenther@suse.de>
16869
16870 PR tree-optimization/57275
16871 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16872 return value for fail to do runtime alias checks for gather loads.
16873
2d6e4603
JH		 168742013-05-15 Jan Hubicka <jh@suse.cz>
16875
16876 PR lto/57038
16877 PR lto/47375
e7208ea3
UB		 16878 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16879 weakrefs are not external.
16880 (lto_symtab_merge_decls): Fix thinko when dealing with
16881 non-lto_symtab decls.
2d6e4603
JH		 16882 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16883 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16884 * varpool.c (dump_varpool_node): Dump more flags.
16885
83f44b39
GG		 168862013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16887
16888 * config/i386/i386.c (processor_alias_table): Add instruction
16889 FSGSBASE for AMD bdver3 architecture.
16890
2e55d062
JJ		 168912013-05-14 Jakub Jelinek <jakub@redhat.com>
16892
16893 * tree.c (warn_deprecated_use): Print file:line using locus color.
16894 * diagnostic.c (diagnostic_report_current_module): Print file:line
16895 and file:line:column using locus color.
16896
0bfdb81e
MS		 168972013-05-14 Mike Stump <mikestump@comcast.net>
16898
16899 * gdbinit.in: Add __null.
16900
a508ef22
MS		 169012013-05-14 Mike Stump <mikestump@comcast.net>
16902
16903 * recog.h: Rename struct recog_data to Recog_data.
16904 * recog.c: Likewise.
16905 * reload.c (can_reload_into): Likewise.
16906 * config/picochip/picochip.c: Likewise.
16907
e7180acb
MS		 169082013-05-14 Mike Stump <mikestump@comcast.net>
16909
16910 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16911
61216c26
SB		 169122013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16913
fd6beed4
SB		 16914 * resource.h (struct resources): Remove unch_memory member.
16915 (CLEAR_RESOURCE): Don't clear unch_memory.
16916 * resource.c (mark_referenced_resources): Don't set it.
16917 (mark_set_resources): Likewise.
16918 (mark_target_live_regs): Don't clear it.
16919 (init_resource_info): Likewise.
16920 * reorg.c (resource_conflicts_p): Don't compare it.
16921 (redundant_insn): Don't set it.
16922
61216c26
SB		 16923 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16924 Remove prototypes.
16925 * emit-rtl.c (next_label): Remove unused function.
16926 (skip_consecutive_labels, link_cc0_insns): Move to ...
16927 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16928 only place where these functions are used, and make them static.
16929
418d1b87
MG		 169302013-05-14 Marc Glisse <marc.glisse@inria.fr>
16931
16932 * fold-const.c (fold_negate_expr): Handle vectors.
16933 (fold_truth_not_expr): Make it static.
16934 (fold_invert_truthvalue): New static function.
16935 (invert_truthvalue_loc): Handle vectors. Do not call
16936 fold_truth_not_expr directly.
16937 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16938 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16939 (fold_comparison): Handle vector constants.
16940 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16941 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16942 * tree.h (fold_truth_not_expr): Remove declaration.
16943
fc21784d
JG		 169442013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16945
16946 * config/aarch64/aarch64-simd.md
16947 (aarch64_vcond_internal<mode>): Rename to...
16948 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16949 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16950 float modes. Clarify all iterator modes.
16951 (vcond<mode><mode>): Use new name for vcond expanders.
16952 (vcond<v_cmp_result><mode>): Likewise.
16953 (vcondu<mode><mode>: Likewise.
16954 * config/aarch64/iterators.md (VDQF_COND): New.
16955
d4c52634
MG		 169562013-05-14 Marc Glisse <marc.glisse@inria.fr>
16957
16958 PR bootstrap/57266
16959 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16960 variable for the shift amount. Check that we shift by non-negative
16961 amounts.
16962
2b261262
CLT		 169632013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16964
16965 PR target/42017
16966 * config/arm/arm.h (EPILOGUE_USES): Only return true
16967 for LR_REGNUM after epilogue_completed.
16968
12211b99 169692013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 16970
16971 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 16972 is error_mark_node.
6e022d7b 16973
56cf7859
RO		 169742013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16975
16976 PR target/57261
16977 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16978 and Solaris 11+/x86 with gld.
16979 * configure: Regenerate.
16980
75776c6d
JJ		 169812013-05-14 Jakub Jelinek <jakub@redhat.com>
16982
16983 * expmed.c (expand_shift_1): Canonicalize rotates by
16984 constant bitsize / 2 to bitsize - 1.
7f998021 16985 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 16986 case ROTATERT>: Likewise.
16987
16988 Revert:
16989 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16990
16991 * config/i386/i386.md (rotateinv): New code attr.
16992 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16993 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16994 roll $31, %eax, etc.
16995
df35498a
RB		 169962013-05-14 Richard Biener <rguenther@suse.de>
16997
16998 PR middle-end/57235
16999 * tree-eh.c (sink_clobbers): Give up for successors with
17000 multiple predecessors and no virtual uses.
17001
cc6e7ece
EB		 170022013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17003
17004 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17005 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17006
f778dd4d
JJ		 170072013-05-14 Jakub Jelinek <jakub@redhat.com>
17008
17009 PR middle-end/57251
17010 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17011 the case when both op0 and op1 have VOIDmode.
17012
bad4df9b
KP		 170132013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17014
17015 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17016 in multiply-accumulate mode.
17017
56f3e9ac
GW		 170182013-05-13 Guozhi Wei <carrot@google.com>
17019
17020 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17021
a3d7ab92
KT		 170222013-05-13 Kai Tietz <ktietz@redhat.com>
17023
17024 PR target/56975
17025 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 17026 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 17027 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 17028 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 17029 (ix86_expand_prologue): Likewise.
17030 (ix86_expand_split_stack_prologue): Likewise.
17031 (legitimate_pic_address_disp_p): Likewise.
17032 (legitimize_pic_address): Likewise.
17033 (legitimize_tls_address): Likewise.
17034 (legitimize_pe_coff_symbol): Likewise.
17035 (output_pic_addr_const): Likewise.
17036 (construct_plt_address): Likewise.
17037 (ix86_expand_call): Likewise.
17038 (x86_output_mi_thunk): Likewise.
17039 (x86_function_profiler): Likewise.
17040
c59b7e28
SN		 170412013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17042
17043 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17044 similar switch cases.
17045 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17046 (aarch64_simd_mov_to_<mode>low): Delete.
17047 (aarch64_simd_mov_to_<mode>high): Delete.
17048 (move_lo_quad_<mode>): Add w<-r alternative.
17049 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17050 (aarch64_simd_mov_from_*): Update type attribute.
17051 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17052 statement.
17053
5f28524a
JH		 170542013-05-13 Jan Hubicka <jh@suse.cz>
17055
17056 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17057 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 17058 ix86_expand_epilogue, emit_i387_cw_initialization,
17059 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17060 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 17061
ae6fa899
JJ		 170622013-05-13 Jakub Jelinek <jakub@redhat.com>
17063
17064 PR tree-optimization/45216
17065 PR tree-optimization/57157
17066 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17067 the (-Y) & (B - 1) variant if OP is |.
17068 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17069 return shifted. Use (-op1) & (prec - 1) as other_amount
17070 instead of prec - op1.
17071
4502fe8d
MJ		 170722013-05-13 Martin Jambor <mjambor@suse.cz>
17073
17074 PR middle-end/42371
17075 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17076 (ipa_constant_data): New type.
17077 (ipa_jump_func): Use ipa_constant_data to hold information about
17078 constant jump functions.
17079 (ipa_get_jf_constant): Adjust to jump function type changes.
17080 (ipa_get_jf_constant_rdesc): New function.
17081 (ipa_param_descriptor): New field controlled_uses.
17082 (ipa_get_controlled_uses): New function.
17083 (ipa_set_controlled_uses): Likewise.
17084 * ipa-ref.h (ipa_find_reference): Declare.
17085 * ipa-prop.c (ipa_cst_ref_desc): New type.
17086 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17087 changes.
17088 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17089 New parameter cs. Adjust all callers.
17090 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17091 (remove_described_reference): New function.
17092 (jfunc_rdesc_usable): Likewise.
17093 (try_make_edge_direct_simple_call): Decrement controlled use count,
17094 attempt to remove reference if it hits zero.
17095 (combine_controlled_uses_counters): New function.
17096 (propagate_controlled_uses): Likewise.
17097 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17098 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17099 (ipa_print_node_params): Print described use counter.
17100 (ipa_write_jump_function): Adjust to jump function type changes.
17101 (ipa_read_jump_function): New parameter CS, pass it to
17102 ipa_set_jf_constant. Adjust caller.
17103 (ipa_write_node_info): Stream controlled use count
17104 (ipa_read_node_info): Likewise.
17105 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17106 asserting.
17107 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17108 count. Remove cloning-added reference if it reaches zero.
17109 * ipa-ref.c (ipa_find_reference): New function.
17110
0864bfc2
GG		 171112013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17112
e7208ea3 17113 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 17114 alignment values for AMD BD and BT architectures.
17115
640bfeb2
MG		 171162013-05-13 Marc Glisse <marc.glisse@inria.fr>
17117
17118 * tree-vect-generic.c (uniform_vector_p): Move ...
17119 * tree.c (uniform_vector_p): ... here.
17120 * tree.h (uniform_vector_p): Declare it.
17121 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17122 into a scalar.
17123
3a60f32b
JJ		 171242013-05-13 Jakub Jelinek <jakub@redhat.com>
17125
198fe1bf
JJ		 17126 PR tree-optimization/57230
17127 * tree-ssa-strlen.c (handle_char_store): Record length for
17128 array store from STRING_CST.
17129
3a60f32b
JJ		 17130 PR tree-optimization/57230
17131 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17132 check.
17133
566be57c
JR		 171342013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17135
17136 * config/epiphany/epiphany.c (epiphany_init): Check size of
17137 NUM_MODES_FOR_MODE_SWITCHING.
17138 (epiphany_expand_prologue):
17139 Remove CONFIG_REGNUM initial value handling code.
17140 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17141 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 17142 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 17143 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17144 Don't return 1 for FP_MODE_NONE.
17145 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17146 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17147 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17148 * config/epiphany/epiphany.md (save_config): New pattern.
17149
0f2c2331
UB		 171502013-05-12 Uros Bizjak <ubizjak@gmail.com>
17151
17152 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17153
5b3f93c7
UB		 171542013-05-10 Uros Bizjak <ubizjak@gmail.com>
17155
17156 * config/i386/i386.md (memory): Handle sseishft1.
17157 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17158 (*vec_extractv2di_1): Ditto.
17159
1f873f0e
VM		 171602013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17161
17162 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17163 saved registers.
17164
341427fa 171652013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 17166
17167 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17168 Add mthumb/march=armv7-a multilib.
17169 Add mthumb/march=armv7-r multilib.
17170 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17171
9e69bdde
RC		 171722013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17173
17174 * config/v850/t-rtems: Add more multilibs.
17175
9ff09a22
RB		 171762013-05-10 Richard Biener <rguenther@suse.de>
17177
17178 PR tree-optimization/57214
17179 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17180 not propagate from SSA names that occur in abnormal PHI nodes.
17181
a5e0cd1d
MG		 171822013-05-10 Marc Glisse <marc.glisse@inria.fr>
17183
17184 * stor-layout.c (element_precision): New function.
17185 * machmode.h (element_precision): Declare it.
17186 * tree.c (build_minus_one_cst): New function.
17187 (element_precision): Likewise.
17188 * tree.h (build_minus_one_cst): Declare new function.
17189 (element_precision): Likewise.
17190 * fold-const.c (operand_equal_p): Use element_precision.
17191 (fold_binary_loc): Handle vector types.
17192 * convert.c (convert_to_integer): Use element_precision.
17193 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17194 separately.
17195
cb2558bc
RS		 171962013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17197
17198 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17199 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17200 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17201 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17202 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17203 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17204 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17205 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17206 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17207 (Uuw8): New constraints.
17208 (Usb4): Move into alphabetical order.
17209 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17210 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17211 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17212 previously unnamed patterns.
17213 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17214 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17215 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17216 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17217 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17218 of specific lengths.
17219
cb3b8d33
JJ		 172202013-05-10 Jakub Jelinek <jakub@redhat.com>
17221
6f93c008
JJ		 17222 * config/i386/i386.md (rotateinv): New code attr.
17223 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17224 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17225 roll $31, %eax, etc.
17226
cb3b8d33
JJ		 17227 PR tree-optimization/45216
17228 PR tree-optimization/57157
17229 * tree-ssa-forwprop.c (simplify_rotate): New function.
17230 (ssa_forward_propagate_and_combine): Call it.
17231
afb119be
RB		 172322013-05-10 Richard Biener <rguenther@suse.de>
17233
17234 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17235 disable peeling when we version for aliasing.
17236 (vector_alignment_reachable_p): Honor explicit user alignment.
17237 (vect_supportable_dr_alignment): Likewise.
17238 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17239 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17240 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17241 then peeling to arrange for the cost-model check to come first.
17242
01ae4861
AM		 172432013-05-10 Alan Modra <amodra@gmail.com>
17244
17245 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17246 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17247 * configure: Regenerate.
17248
ebc9a431
AM		 172492013-05-10 Alan Modra <amodra@gmail.com>
17250
17251 PR target/55033
17252 * varasm.c (default_elf_select_section): Move !DECL_P check..
17253 (get_named_section): ..to here before calling get_section_name.
17254 Adjust assertion.
17255 (default_section_type_flags): Add DECL_P check.
17256 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17257 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17258
d4bca93c
JR		 172592013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17260
17261 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17262 When using gen_stack_adjust_str with a register offset, add a
17263 REG_FRAME_RELATED_EXPR note.
17264
60ca9a65
UB		 172652013-05-09 Uros Bizjak <ubizjak@gmail.com>
17266
17267 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17268 (*vec_extractv4si_zext_mem): Ditto.
17269 (*vec_extractv2di): Add 0->x and x->x alternatives.
17270 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17271 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17272
a3409c02
JM		 172732013-05-09 Jason Merrill <jason@redhat.com>
17274
0138d6b2
JM		 17275 N3639 C++1y VLA support
17276 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17277 DECL_VALUE_EXPR.
17278
a3409c02
JM		 17279 * tree.c (build_constructor_va): New.
17280 * tree.h: Declare it.
17281
66e6b990
MJ		 172822013-05-09 Martin Jambor <mjambor@suse.cz>
17283
17284 PR lto/57084
17285 * gimple-fold.c (canonicalize_constructor_val): Call
17286 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17287
64cfa6c0
JH		 172882013-05-09 Jan Hubicka <jh@suse.cz>
17289 Richard Biener <rguenther@suse.de>
17290
17291 PR lto/54095
17292 * symtab.c (symtab_make_decl_local): Do not add private names.
17293
c3167b00
JH		 172942013-05-09 Jan Hubicka <jh@suse.cz>
17295
17296 PR lto/54095
17297 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17298 (unlink_from_assembler_name_hash): Likewise.
17299 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17300 symtab_unregister_node, symtab_initialize_asm_name_hash,
17301 change_decl_assembler_name): Update.
17302
12dc6974
SN		 173032013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17304
17305 * config/aarch64/aarch64.md: New movtf split.
17306 (*movtf_aarch64): Update.
17307 (aarch64_movdi_tilow): Handle TF modes and rename to
17308 aarch64_movdi_<mode>low.
17309 (aarch64_movdi_tihigh): Handle TF modes and rename to
17310 aarch64_movdi_<mode>high
17311 (aarch64_movtihigh_di): Handle TF modes and rename to
17312 aarch64_mov<mode>high_di
17313 (aarch64_movtilow_di): Handle TF modes and rename to
17314 aarch64_mov<mode>low_di
17315 (aarch64_movtilow_tilow): Remove spurious whitespace.
17316 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17317 splits.
17318 (aarch64_print_operand): Update.
17319
227eb343
AM		 173202013-05-09 Alan Modra <amodra@gmail.com>
17321
17322 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17323 powerpc64le.
17324 * configure: Regenerate.
17325
0b013847
UB		 173262013-05-08 Uros Bizjak <ubizjak@gmail.com>
17327
17328 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17329 splitter preparation statements.
17330 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17331 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17332 adjust_address_nv.
17333
1dc3d6e9
BS		 173342013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17335
17336 * gimple-ssa-strength-reduction.c (count_candidates): Change
17337 return value to int.
17338 (analyze_candidates_and_replace): Change type of length to int.
17339
e61e7d28
UB		 173402013-05-08 Uros Bizjak <ubizjak@gmail.com>
17341
17342 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17343 (*vec_extract<mode>): Use VI12_128 mode iterator.
17344 (*vec_extract<mode>_mem): Ditto.
17345 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17346 attribute.
17347
4fbfcf44
DN		 173482013-05-08 Diego Novillo <dnovillo@google.com>
17349
17350 PR bootstrap/54659
17351
17352 Revert:
4fbfcf44
DN		 17353 2012-08-17 Diego Novillo <dnovillo@google.com>
17354
e61e7d28
UB		 17355 PR bootstrap/54281
17356 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17357 * config.in: Regenerate.
17358 * configure: Regenerate.
17359 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17360
702d8703
JH		 173612013-05-08 Jan Hubicka <jh@suse.cz>
17362
17363 PR lto/54095
17364 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17365 * cgraph.h (symtab_node_base): Add unique_name.
17366 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17367 input_overwrite_node, input_varpool_node): Stream unique_name.
17368 * cgraphclones.c (cgraph_create_virtual_clone,
17369 cgraph_function_versioning): Set unique_name.
17370 * ipa.c (function_and_variable_visibility): Set unique_name.
17371
8b28cf47
BS		 173722013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17373
17374 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17375 (alloc_cand_and_find_basis): Restrict conditional candidate
17376 processing to CAND_MULTs.
17377
e86074fd
JH		 173782013-05-08 Jan Hubicka <jh@suse.cz>
17379
17380 PR lto/54095
17381 lto-symtab.c (lto_symtab_symbol_p): New function.
17382 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17383 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17384 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17385 Skip static symbols.
17386
44398cbe
PC		 173872013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17388
17389 PR tree-optimization/57200
17390 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17391 Only call inform if the preceding warning_at returns true.
17392
f6bc1c4a
HS		 173932013-05-07 Han Shen <shenhan@google.com>
17394
17395 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17396 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17397 * common.opt (fstack-protector-strong): New option.
17398 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17399 * doc/invoke.texi (Optimization Options): Document
17400 "-fstack-protector-strong".
17401 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17402
4ffecb1f
SB		 174032013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17404
17405 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17406
5a107a0f
VM		 174072013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17408
17409 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17410 it is not lhs of the insn.
17411 (setup_reg_equiv): Remove insn having equiv memory which it is not
17412 lhs of the insn.
17413 * lra-constraints.c (process_address): Try to improve generation
17414 code for address base + disp.
17415 (lra_constraints): Make correct the code for checking insn setting
17416 up backward equivalence. Remove insn only if it is in the init
17417 insn list.
17418 * lra-eliminations.c (update_reg_eliminate): Change return value.
17419 (lra_eliminate): Use the result.
17420
3f5783ea
UB		 174212013-05-07 Uros Bizjak <ubizjak@gmail.com>
17422
17423 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17424 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17425 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17426 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17427 register target operands.
17428 (*vec_extractv8hi_sse2): New pattern.
17429 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17430 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17431 (*vec_extract<mode>_mem): New insn and split pattern.
17432
8a5800b8
CL		 174332013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17434
17435 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17436 (TARGET_ASAN_SHADOW_OFFSET): Define.
17437 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17438 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17439
7bf55a70
BS		 174402013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17441
17442 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17443 (incr_vec_index): Return -1 if increment not found.
17444 (create_add_on_incoming_edge): Assert if increment not found.
17445 (record_increment): Limit number of increments recorded.
17446 (all_phi_incrs_profitable): Return false if an increment not found.
17447 (replace_profitable_candidates): Don't process increments that were
17448 not recorded.
17449 (analyze_candidates_and_replace): Limit size of incr_vec.
17450
3f8825c0
RB		 174512013-05-07 Richard Biener <rguenther@suse.de>
17452
17453 * calls.c (special_function_p): setjmp-like functions are leaf.
17454 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17455 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17456
fd4842cd
SN		 174572013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17458
17459 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17460 (aarch64_simd_mov<mode>): New expander.
17461 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17462 (aarch64_simd_mov_to_<mode>high): Likewise.
17463 (aarch64_simd_mov_from_<mode>low): Likewise.
17464 (aarch64_simd_mov_from_<mode>high): Likewise.
17465 (aarch64_dup_lane<mode>): Update.
17466 (aarch64_dup_lanedi): New instruction pattern.
17467 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17468 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17469
a7a7d10e
BS		 174702013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17471
17472 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17473 (replace_mult_candidate): Remove unnecessary argument; remove
17474 unnecessary parameter from call to introduce_cast_before_cand.
17475 (replace_unconditional_candidate): Remove unnecessary parameter
17476 from call to replace_mult_candidate.
17477 (replace_conditional_candidate): Likewise.
17478 (insert_initializers): Use make_temp_ssa_name.
17479 (introduce_cast_before_cand): Remove unnecessary argument; use
17480 make_temp_ssa_name.
17481 (replace_one_candidate): Remove unnecessary argument; remove
17482 unnecessary parameter from calls to introduce_cast_before_cand.
17483 (replace_profitable_candidates): Remove unnecessary parameters
17484 from calls to replace_one_candidate.
17485
29105868
BS		 174862013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17487
17488 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17489 phi def as possibly hiding a basis for a CAND_ADD whose operands
17490 have been commuted in the analysis.
17491 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17492
4095f9fa
N		 174932013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17494
17495 * config/aarch64/aarch64.md
17496 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17497 shift value between 0-4.
17498
39e843e8
RB		 174992013-05-07 Richard Biener <rguenther@suse.de>
17500
17501 * double-int.h (rshift): New overload.
17502 * double-int.c (rshift): New function.
17503 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17504 (create_reference_ops_from_ref): Remove.
17505 (vn_reference_insert): Use shared ops for constructing the
17506 reference and copy it.
17507
0a1a83cb
RB		 175082013-05-07 Richard Biener <rguenther@suse.de>
17509
17510 PR middle-end/57190
17511 * tree-eh.c (sink_clobbers): Properly propagate
17512 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17513
43303d6f
JJ		 175142013-05-07 Jakub Jelinek <jakub@redhat.com>
17515
ba7e83f8
JJ		 17516 PR tree-optimization/57149
17517 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17518 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17519 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17520 uninit_undefined_value_p instead of ssa_undefined_value_p.
17521
43303d6f
JJ		 17522 PR debug/57184
17523 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17524 for modifier == EXPAND_INITIALIZER.
17525
14523c25
AB		 175262013-05-07 Anton Blanchard <anton@samba.org>
17527
17528 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17529 for powerpc64 little endian.
17530 * configure: Regenerate.
17531
cb7c8be9
GS		 175322013-05-06 Graham Stott <grahams@btinternet.com>
17533
17534 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17535 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17536 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17537 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17538
175392013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 17540
17541 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17542 codes which allow non-lvalues.
17543
9a0ee7b0
MG		 175442013-05-06 Marc Glisse <marc.glisse@inria.fr>
17545
17546 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17547 components are all 1s.
17548 (integer_minus_onep): New function.
17549 * tree.h (integer_minus_onep): Declare it.
17550 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17551 integer_minus_onep instead of integer_all_onesp.
17552
f2c17ea9
OE		 175532013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17554
17555 PR target/52933
17556 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17557 variations of these patterns.
17558
f75e6a51
UB		 175592013-05-06 Uros Bizjak <ubizjak@gmail.com>
17560
17561 * config/i386/i386.md (isa): Add x64_sse4 member.
17562 (enabled): Handle x64_sse4.
17563 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17564 instruction for 64bit SSE4_1 targets. Update insn attributes.
17565 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17566 instruction for SSE4_1 targets. Update insn attributes.
17567 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17568 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17569 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17570 const_1 selector.
17571 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17572 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17573 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17574
a986d468
OE		 175752013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17576
17577 PR target/57108
17578 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17579
382522cb
MK		 175802013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17581
17582 * final.c (do_assembler_dialects): Don't handle curly braces and
17583 vertical bar escaped by % as dialect delimiters.
17584 (output_asm_insn): Print curly braces and vertical bar if escaped
17585 by % and ASSEMBLER_DIALECT defined.
17586 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17587 * doc/tm.texi: Regenerated.
17588
fb0d5c60
SB		 175892013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17590
fb0d5c60
SB		 17591 * config/mips/mips.c: Include tree-pass.h.
17592 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17593 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17594 (pass_mips_machine_reorg2): New machine specific pass.
17595 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17596 (mips_option_override): Register the new pass.
17597 * rtl.h (cleanup_barriers): Remove prototype.
17598 (dbr_schedule): Likewise.
17599 * jump.c (cleanup_barriers): Make static.
17600 * reorg.c (dbr_schedule): Likewise.
17601
aa06a978
RB		 176022013-05-06 Richard Biener <rguenther@suse.de>
17603
17604 PR tree-optimization/57185
17605 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17606 SSA names properly.
17607
0b953bec
UB		 176082013-05-06 Uros Bizjak <ubizjak@gmail.com>
17609
17610 PR target/57106
17611 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17612 between "first_arg" and "insn", not "last" and "insn".
17613
28708525
WS		 176142013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17615
17616 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17617 (find_candidates_in_block): Re-enable slsr_process_phi.
17618 (create_phi_basis): Fix double counting of candidate adjustment.
17619
0107dca2
RB		 176202013-05-06 Richard Biener <rguenther@suse.de>
17621
17622 PR middle-end/57147
17623 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17624 the edge is also fallthru, preserve it and just clear the
17625 abnormal flag.
17626 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17627 also complex, preserve that and just clear the fallthru flag.
17628 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17629 update virtual operands.
17630
470d4d13
AM		 176312013-05-06 Alan Modra <amodra@gmail.com>
17632
17633 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17634 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17635 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17636 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17637 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17638 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17639
8f1dbf8d
AM		 176402013-05-06 Alan Modra <amodra@gmail.com>
17641
17642 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17643 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17644 (DEFAULT_ASM_ENDIAN): Define.
17645 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17646 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17647 Update -K PIC clause from sysv4.h.
17648 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17649 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17650
54890767
AM		 176512013-05-06 Alan Modra <amodra@gmail.com>
17652
17653 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17654 twice for little-endian.
17655 (ashrdi3_no_power, ashrdi3): Support little-endian.
17656
2353515d
OE		 176572013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17658
17659 PR target/55303
17660 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17661 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17662 related expanders.
17663 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17664 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17665 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17666 New predicates.
17667
33e67557
SB		 176682013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17669 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17670
17671 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17672 * config/pa/pa.opt: Make mbig-switch a no-op.
17673 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17674 (CASE_VECTOR_MODE): Always return SImode.
17675 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17676 for the !TARGET_BIG_SWITCH case.
17677 * config/pa/pa-linux.h: Likewise.
17678 * config/pa/pa-openbsd.h: Likewise.
17679 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17680 * config/pa/pa.md (short_jump): Remove define_insn.
17681 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17682 (casesi0): Remove define_insn.
17683 (type): Remove btable_branch.
17684 (pa_combine_type): Likewise.
17685 (in_nullified_branch_delay): Likewise.
17686 (in_call_delay): Likewise.
17687 (define_delay): Likewise.
17688 (define_insn_reservation "Z3"): Likewise.
17689 (define_insn_reservation "Z4"): Likewise.
17690 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17691 (pa_adjust_insn_length): Remove adjustment for btable branches.
17692 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17693 and mno-big-switch
17694
3095685e
UB		 176952013-05-05 Uros Bizjak <ubizjak@gmail.com>
17696
17697 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17698 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17699 Add m->r,x alternatives.
17700 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17701 splitters using SWI48x mode iterator.
17702 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17703 TARGET_64BIT. Add m->x alternative.
17704 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17705 Add o->x alternative. Enable for TARGET_SSE.
17706 (sse_storeq): Remove expander.
17707 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17708 with memory input operand.
17709 (*vec_extractv2di_1 splitter): New.
17710 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17711 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17712
4b36ae28
SB		 177132013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17714
17715 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17716 (INT_LOWPART): Delete.
17717 (extract_MB): Adjust.
17718 (extract_ME): Adjust.
17719 (print_operand): Adjust.
17720
da226db2
SB		 177212013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17722
17723 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17724 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17725 (reg_or_logical_cint_operand, easy_fp_constant,
17726 logical_const_operand): Delete "CONST_DOUBLE" case.
17727 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17728 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 17729 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 17730 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17731 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17732 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17733 test.
17734 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17735 CONST_DOUBLE DImode/VOIDmode case.
17736 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17737 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17738 CONST_DOUBLE VOIDmode case.
17739 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17740 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17741 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17742 Delete CONST_DOUBLE case.
17743 (splitters for mov FMOVE64 const_double): Delete
17744 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17745 "HOST_BITS_PER_WIDE_INT >= 64" test.
17746 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17747 case.
17748 (mov DI const_double): Delete.
17749
40de22d6
JJ		 177502013-05-04 Jakub Jelinek <jakub@redhat.com>
17751
17752 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17753 on op shows all bits zero in mode of a lowpart subreg, return zero.
17754
5ec6aff2
MM		 177552013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17756
17757 PR target/57150
17758 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17759 to save TFmode registers and DImode to save TImode registers for
17760 caller save operations.
17761 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17762 mark being partially clobbered since they only use the first
17763 double word.
17764
17765 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17766 and TDmode only use the upper 64-bits of each VSX register.
17767
2cefad90
BS		 177682013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17769
17770 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17771 (find_candidates_in_block): Disable slsr_process_phi.
17772
d6d7eee1
GW		 177732013-05-03 Guozhi Wei <carrot@google.com>
17774
17775 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 17776 constructor to ...
17777 (build_init_ctor): ... here.
d6d7eee1 17778
9b92d12b
BS		 177792013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17780
17781 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17782 (slsr_cand_d): Redefine def_phi.
17783 (stride_status, phi_adjust_status, count_phis_status): New enums.
17784 (find_phi_def): New.
17785 (find_basis_for_base_expr): New.
17786 (find_basis_for_candidate): Handle hidden bases.
17787 (alloc_cand_and_find_basis): Handle phi candidates.
17788 (slsr_process_phi): New.
17789 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17790 (create_mul_imm_cand): Likewise.
17791 (create_add_ssa_cand): Exclude phi base candidates.
17792 (create_add_imm_cand): Likewise.
17793 (slsr_process_cast): Likewise.
17794 (slsr_process_copy): Likewise.
17795 (find_candidates_in_block): Handle phi candidates.
17796 (dump_candidate): Likewise.
17797 (unconditional_cands): Delete.
17798 (unconditional_cands_with_known_stride_p): Delete.
17799 (phi_dependent_cand_p): New.
17800 (cand_increment): Handle phi-dependent candidates.
17801 (replace_dependent): Delete.
17802 (replace_mult_candidate): New.
17803 (replace_unconditional_candidate): New.
17804 (incr_vec_index): Move to avoid forward reference.
17805 (create_add_on_incoming_edge): New.
17806 (create_phi_basis): New.
17807 (replace_dependents): Delete.
17808 (replace_conditional_candidate): New.
17809 (phi_add_costs): New.
17810 (replace_uncond_cands_and_profitable_phis): New.
17811 (record_increment): Handle phi adjustments.
17812 (record_phi_increments): New.
17813 (record_increments): Handle phi adjustments.
17814 (phi_incr_cost): New.
17815 (lowest_cost_path): Handle phis.
17816 (total_savings): Likewise.
17817 (analyze_increments): Likewise.
17818 (ncd_with_phi): New.
17819 (ncd_of_cand_and_phis): New.
17820 (nearest_common_dominator_for_cands): Handle phi increments.
17821 (all_phi_incrs_profitable): New.
17822 (replace_profitable_candidates): Handle phi-dependent candidates.
17823 (analyze_candidates_and_replace): Likewise.
17824
68f073d4
TJ		 178252013-05-03 Teresa Johnson <tejohnson@google.com>
17826
17827 PR bootstrap/57154
17828 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17829 do not exceed REG_BR_PROB_BASE.
17830
a4ee7cb9
JL		 178312013-05-03 Jeff Law <law@redhat.com>
17832
ade67f70 17833 PR tree-optimization/57144
a4ee7cb9
JL		 17834 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17835 operand of the condition will bit into the new type when eliminating
17836 a cast feeding a condition.
17837
47954c4e
JJ		 178382013-05-03 Jakub Jelinek <jakub@redhat.com>
17839
17840 PR rtl-optimization/57130
3095685e
UB		 17841 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17842 of COMPARE as in_code to the recursive call if needed.
47954c4e 17843
3c21604f
UB		 178442013-05-03 Uros Bizjak <ubizjak@gmail.com>
17845
17846 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17847 (enabled): Handle new members.
17848 * config/i386/sse.md (*vec_concatv2si): Merge from
17849 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17850 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17851
12211b99 178522013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 17853
17854 PR tree-optimization/57027
17855 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17856 for fnms opportunity, check we got the prerequisite kind
17857 of tree / gimple before using accessor functions.
17858
07bfc9ec
RB		 178592013-05-03 Richard Biener <rguenther@suse.de>
17860
17861 * double-int.h (lshift): New overload without precision
17862 and arith argument.
17863 (operator *=, operator +=, operator -=): Move ...
17864 * double-int.c (operator *=, operator +=, operator -=): ... here
17865 and implement more efficiently.
17866 (mul_double_with_sign): Remove.
17867 (lshift_double): Adjust to take unsinged shift argument, push
17868 dispatching code to callers.
17869 (mul_double_wide_with_sign): Add early out for callers that
17870 are not interested in high parts or overflow.
17871 (lshift): New function.
17872 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17873 dispatch code here.
17874 (lrotate, rrotate): Use logical shifts.
17875 * expr.c (get_inner_reference): Use lshift.
17876 * fixed-value.c (do_fixed_divide): Likewise.
17877 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17878 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17879 (indirect_refs_may_alias_p): Likewise.
17880 (stmt_kills_ref_p_1): Likewise.
17881
7769bb64
VP		 178822013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17883
17884 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17885
1d0c8e5c
VP		 178862013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17887
17888 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17889 scalar form of FABD instruction.
17890
f15feaf9
VM		 178912013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17892
17893 * lra-constraints.c (process_alt_operands): Add checking alt
17894 number to choose the best alternative.
17895
d90e76d4
RB		 178962013-05-02 Richard Biener <rguenther@suse.de>
17897
17898 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17899 bitmap and its handling.
17900 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17901
9f8e7a96
RB		 179022013-05-02 Richard Biener <rguenther@suse.de>
17903
17904 PR middle-end/57140
17905 * tree-inline.c (copy_loops): Properly handle removed loops.
17906 (copy_cfg_body): Mark destination loops for fixup if source
17907 loops needed fixup.
17908
f3a81b39
GY		 179092013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17910
17911 PR target/56732
17912 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17913 generating simple_return for naked functions.
17914
7b920a9a
MJ		 179152013-05-02 Martin Jambor <mjambor@suse.cz>
17916
17917 PR middle-end/56988
17918 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17919 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17920 flags match.
17921 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17922 ipa_agg_replacement_value structures.
17923 (known_aggs_to_agg_replacement_list): Likewise.
17924 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17925 (read_agg_replacement_chain): Likewise.
17926 (ipcp_transform_function): Also check that by_ref flags match.
17927
2c41c19d
RB		 179282013-05-02 Richard Biener <rguenther@suse.de>
17929
17930 * graphds.h (struct graph): Add obstack member.
17931 * graphds.c (new_graph): Initialize obstack and allocate
17932 vertices from it.
17933 (add_edge): Allocate edge from the obstack.
3c21604f 17934 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 17935
8b47039c
TJ		 179362013-05-02 Teresa Johnson <tejohnson@google.com>
17937
17938 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17939 divides.
17940 * cfg.c (update_bb_profile_for_threading): Ditto.
17941 * tree-inline.c (copy_bb): Ditto.
17942 (copy_edges_for_bb): Ditto.
17943 (initialize_cfun): Ditto.
17944 (copy_cfg_body): Ditto.
17945 (expand_call_inline): Ditto.
17946 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17947 (estimate_node_size_and_time): Ditto.
17948 (inline_merge_summary): Ditto.
17949 * cgraphclones.c (cgraph_clone_edge): Ditto.
17950 (cgraph_clone_node): Ditto.
17951 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17952 (compute_trg_info): Ditto.
17953
da65928c
IB		 179542013-05-02 Ian Bolton <ian.bolton@arm.com>
17955
17956 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17957 S reg when fp attribute set.
17958 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17959
add3c965
IB		 179602013-05-02 Ian Bolton <ian.bolton@arm.com>
17961
17962 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17963 New pattern.
17964 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17965 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17966 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17967
fdd43ac4
RB		 179682013-05-02 Richard Biener <rguenther@suse.de>
17969
17970 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17971 (struct instantiate_cache_entry): New type.
17972 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17973 (struct instantiate_cache_type): New type.
17974 (set_instantiated_value, get_instantiated_value): Remove.
17975 (get_instantiated_value_entry): New function.
17976 (instantiate_scev_name): Use the new cache and adjust.
17977 (instantiate_scev_poly): Adjust.
17978 (instantiate_scev_binary): Likewise.
17979 (instantiate_array_ref): Likewise.
17980 (instantiate_scev_convert): Likewise.
17981 (instantiate_scev_not): Likewise.
17982 (instantiate_scev_3): Likewise.
17983 (instantiate_scev_2): Likewise.
17984 (instantiate_scev_r): Likewise.
17985 (instantiate_scev): Likewise.
17986 (resolve_mixers): Likewise.
17987
36ff9dfb
VM		 179882013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17989
17990 PR target/57091
17991 * lra-constraints.c (best_small_class_operands_num): Remove.
17992 (process_alt_operands): Remove small_class_operands_num. Take
17993 small classes operands into losers and only if the operand is not
17994 matched. Modify debugging output.
17995 (curr_insn_transform): Remove best_small_class_operands_num.
17996 Print insn name.
17997
36054fab
JG		 179982013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17999
18000 * config/aarch64/aarch64-builtins.c
18001 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18002 * config/aarch64/aarch64-simd-builtins.def
18003 (reduc_splus_): Add new modes.
18004 (reduc_uplus_): New.
18005 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18006 (reduc_uplus_v4sf): Likewise.
18007 (reduc_splus_v4sf): Likewise.
18008 (aarch64_addv<mode>): Likewise.
18009 (reduc_uplus_<mode>): Likewise.
18010 (reduc_splus_<mode>): Likewise.
18011 (aarch64_addvv2di): Likewise.
18012 (reduc_uplus_v2di): Likewise.
18013 (reduc_splus_v2di): Likewise.
18014 (aarch64_addvv2si): Likewise.
18015 (reduc_uplus_v2si): Likewise.
18016 (reduc_splus_v2si): Likewise.
18017 (reduc_<sur>plus_<mode>): New.
18018 (reduc_<sur>plus_v2di): Likewise.
18019 (reduc_<sur>plus_v2si): Likewise.
18020 (reduc_<sur>plus_v4sf): Likewise.
18021 (aarch64_addpv4sf): Likewise.
18022 * config/aarch64/arm_neon.h
18023 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18024 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18025 add UNSPEC_SADDV, UNSPEC_UADDV.
18026 (SUADDV): New.
18027 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18028
6dce23a8
JG		 180292013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18030
18031 * config/aarch64/arm_neon.h
18032 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18033
1598945b
JG		 180342013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18035
18036 * config/aarch64/aarch64-builtins
18037 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18038
998eaf97
JG		 180392013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18040
18041 * config/aarch64/aarch64-simd-builtins.def
18042 (reduc_smax_): New.
18043 (reduc_smin_): Likewise.
18044 (reduc_umax_): Likewise.
18045 (reduc_umin_): Likewise.
18046 (reduc_smax_nan_): Likewise.
18047 (reduc_smin_nan_): Likewise.
18048 (fmax): Remove.
18049 (fmin): Likewise.
18050 (smax): Update for V2SF, V4SF and V2DF modes.
18051 (smin): Likewise.
18052 (smax_nan): New.
18053 (smin_nan): Likewise.
18054 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18055 (<su><maxmin><mode>3): ...This, refactor.
18056 (s<maxmin><mode>3): New.
18057 (<maxmin_uns><mode>3): Likewise.
18058 (reduc_<maxmin_uns>_<mode>): Refactor.
18059 (reduc_<maxmin_uns>_v4sf): Likewise.
18060 (reduc_<maxmin_uns>_v2si): Likewise.
18061 (aarch64_<fmaxmin><mode>: Remove.
18062 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18063 new builtin names.
18064 (vmin<q>_f<32,64>): Likewise.
18065 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18066 (FMAXMIN): New.
18067 (su): Add mappings for smax, smin, umax, umin.
18068 (maxmin): New.
18069 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18070 (FMAXMIN): Rename as...
18071 (FMAXMIN_UNS): ...This.
18072 (maxminv): Remove.
18073 (fmaxminv): Likewise.
18074 (fmaxmin): Likewise.
18075 (maxmin_uns): New.
18076 (maxmin_uns_op): Likewise.
18077
bd11644e
JG		 180782013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18079
18080 * config/aarch64/arm_neon.h
18081 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18082 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18083 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18084
75dd5ace
JG		 180852013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18086
18087 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18088 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18089
7c19979f
JG		 180902013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18091
18092 * config/aarch64/aarch64-simd.md
18093 (vcond<mode>_internal): Handle special cases for constant masks.
18094 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18095 (vcondu<mode><mode>): Likewise.
18096 (vcond<v_cmp_result><mode>): New.
18097
bb60efd9
JG		 180982013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18099
18100 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18101 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18102 * config/aarch64/aarch64-simd-builtins.def
18103 (cmeq): Update to BUILTIN_VALLDI.
18104 (cmgt): Likewise.
18105 (cmge): Likewise.
18106 (cmle): Likewise.
18107 (cmlt): Likewise.
18108 * config/aarch64/arm_neon.h
18109 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18110 to builtins or C as appropriate.
18111
889b9412
JG		 181122013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18113
18114 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18115 (cmgeu): ...This.
18116 (cmhi): Rename to...
18117 (cmgtu): ...This.
18118 * config/aarch64/aarch64-simd.md
18119 (simd_mode): Add SF.
18120 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18121 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18122 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18123 (cstore<mode>_neg): ...This.
18124 * config/aarch64/iterators.md
18125 (VALLF): new.
18126 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18127 (COMPARISONS): New.
18128 (UCOMPARISONS): Likewise.
18129 (optab): Add missing comparisons.
18130 (n_optab): New.
18131 (cmp_1): Likewise.
18132 (cmp_2): Likewise.
18133 (CMP): Likewise.
18134 (cmp): Remove.
18135 (VCMP_S): Likewise.
18136 (VCMP_U): Likewise.
18137 (V_cmp_result): Add DF, SF modes.
18138 (v_cmp_result): Likewise.
18139 (v): Likewise.
18140 (vmtype): Likewise.
18141 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18142
0a7dbb76
GY		 181432013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18144
18145 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18146 define_insn to define_insn_and_split.
18147 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18148 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18149 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18150 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18151 (thumb2_negscc): Likewise.
18152
fb614ca6
GY		 181532013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18154
18155 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18156
9e64a0bf
GY		 181572013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18158
18159 * config/arm/thumb2.md: Remove trailing whitespaces.
18160
d6b28156
RS		 181612013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18162
18163 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18164 Use gen_int_mode rather than GEN_INT.
18165
f91674c3
L		 181662013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18167
b0dec607 18168 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 18169 debug_gimple_stmt.
18170
3551257c
RB		 181712013-04-30 Richard Biener <rguenther@suse.de>
18172
18173 PR middle-end/57122
3c21604f 18174 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 18175
4c1aff1c
RB		 181762013-04-30 Richard Biener <rguenther@suse.de>
18177
18178 PR middle-end/57107
18179 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18180
54b8379a
AB		 181812013-04-30 Andrey Belevantsev <abel@ispras.ru>
18182
18183 PR rtl-optimization/56957
18184 PR rtl-optimization/57105
54b8379a
AB		 18185 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18186 variable. Use just INSN_UID for determining whether an insn
18187 should be only disconnected from the insn stream.
18188 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18189
abc27962
JJ		 181902013-04-30 Jakub Jelinek <jakub@redhat.com>
18191
18192 PR tree-optimization/57104
18193 * tsan.c (instrument_expr): Don't instrument accesses to
18194 DECL_HARD_REGISTER VAR_DECLs.
18195
0fc822d0
RB		 181962013-04-30 Richard Biener <rguenther@suse.de>
18197
18198 * function.h (loops_for_fn): New inline function.
18199 (set_loops_for_fn): Likewise.
18200 * cfgloop.h (place_new_loop): Add struct function parameter.
18201 (get_loop): Likewise.
18202 (get_loops): Likewise.
18203 (number_of_loops): Likewise.
18204 (fel_next): Adjust.
18205 (fel_init): Likewise.
18206 * cfg.c (get_loop_copy): Adjust.
18207 * cfgloop.c (flow_loops_dump): Likewise.
18208 (record_loop_exits): Likewise.
18209 (verify_loop_structure): Likewise.
18210 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18211 (estimate_reg_pressure_cost): Likewise.
18212 (mark_loop_exit_edges): Likewise.
18213 * cfgloopmanip.c (place_new_loop): Likewise.
18214 (add_loop): Likewise.
18215 (duplicate_loop): Likewise.
18216 * graph.c (draw_cfg_nodes): Likewise.
18217 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18218 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18219 (extract_affine_chrec): Likewise.
18220 (build_scop_iteration_domain): Likewise.
18221 * graphite.c (graphite_initialize): Likewise.
18222 * ira-build.c (create_loop_tree_nodes): Likewise.
18223 (more_one_region_p): Likewise.
18224 (rebuild_regno_allocno_maps): Likewise.
18225 (mark_loops_for_removal): Likewise.
18226 (mark_all_loops_for_removal): Likewise.
18227 (remove_unnecessary_regions): Likewise.
18228 (ira_build): Likewise.
18229 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18230 * loop-init.c (fix_loop_structure): Likewise.
18231 (gate_rtl_move_loop_invariants): Likewise.
18232 (gate_rtl_unswitch): Likewise.
18233 (gate_rtl_unroll_and_peel_loops): Likewise.
18234 (rtl_doloop): Likewise.
18235 * lto-streamer-in.c (input_cfg): Likewise.
18236 * lto-streamer-out.c (output_cfg): Likewise.
18237 * modulo-sched.c (sms_schedule): Likewise.
18238 * predict.c (tree_estimate_probability): Likewise.
18239 (tree_estimate_probability_driver): Likewise.
18240 (estimate_loops): Likewise.
18241 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18242 (move_sese_region_to_fn): Likewise.
18243 (debug_loop_num): Likewise.
18244 * tree-chrec.c (chrec_evaluate): Likewise.
18245 (hide_evolution_in_other_loops_than_loop): Likewise.
18246 (chrec_component_in_loop_num): Likewise.
18247 (reset_evolution_in_loop): Likewise.
18248 (evolution_function_is_invariant_rec_p): Likewise.
18249 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18250 * tree-inline.c (copy_loops): Likewise.
18251 (copy_cfg_body): Likewise.
18252 (tree_function_versioning): Likewise.
18253 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18254 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18255 Likewise.
18256 (add_to_evolution_1): Likewise.
18257 (scev_const_prop): Likewise.
18258 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18259 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18260 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18261 (tree_ssa_lim_initialize): Likewise.
18262 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18263 (verify_loop_closed_ssa): Likewise.
18264 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18265 (tree_ssa_loop_im): Likewise.
18266 (tree_ssa_loop_unswitch): Likewise.
18267 (tree_vectorize): Likewise.
18268 (check_data_deps): Likewise.
18269 (tree_ssa_loop_ivcanon): Likewise.
18270 (tree_ssa_loop_bounds): Likewise.
18271 (tree_complete_unroll): Likewise.
18272 (tree_complete_unroll_inner): Likewise.
18273 (tree_parallelize_loops): Likewise.
18274 (tree_ssa_loop_prefetch): Likewise.
18275 (tree_ssa_loop_ivopts): Likewise.
18276 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18277 * tree-vectorizer.c (vectorize_loops): Likewise.
18278
37953bd3
MF		 182792013-04-29 Mike Frysinger <vapier@gentoo.org>
18280
18281 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18282 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18283 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18284 with EABI_LINK_SPEC.
18285
f9ed28db
UB		 182862013-04-29 Uros Bizjak <ubizjak@gmail.com>
18287
18288 PR target/44578
18289 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18290 alternative.
18291
deca73f5
VM		 182922013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18293
18294 PR target/57097
37953bd3 18295 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 18296 using memory for pseudos. Print cost dump for alternatives.
18297 Modify cost values for conflicts with early clobbers.
18298 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18299
fdca7d03
UB		 183002013-04-29 Uros Bizjak <ubizjak@gmail.com>
18301
18302 PR target/57098
18303 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18304
0b064172
IB		 183052013-04-29 Ian Bolton <ian.bolton@arm.com>
18306
18307 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18308 from/to S register.
18309 (movdi_aarch64): Support LDR/STR from/to D register.
18310
473cec55
IB		 183112013-04-29 Ian Bolton <ian.bolton@arm.com>
18312
18313 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18314 or higher by default.
18315
a6f30e66
RB		 183162013-04-29 Richard Biener <rguenther@suse.de>
18317
18318 PR middle-end/57075
18319 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18320 even if not adding abnormal edges for calls that can make
18321 abnormal gotos.
18322
0c2b2040
RB		 183232013-04-29 Richard Biener <rguenther@suse.de>
18324
18325 PR middle-end/57103
18326 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18327 TREE_BLOCK.
18328 (move_stmt_r): Remove redundant checking.
18329
f41f80f9
TJ		 183302013-04-29 Teresa Johnson <tejohnson@google.com>
18331
18332 PR bootstrap/57077
18333 * basic-block.h (apply_scale): New function.
18334 (apply_probability): Use apply_scale.
18335 * gimple-streamer-in.c (input_bb): Ditto.
18336 * lto-streamer-in.c (input_cfg): Ditto.
18337 * lto-cgraph.c (merge_profile_summaries): Ditto.
18338 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18339 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 18340 (copy_edges_for_bb): Ditto.
18341 (copy_cfg_body): Ditto.
18342
315bbd2e
TV		 183432013-04-29 Tom de Vries <tom@codesourcery.com>
18344
18345 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18346 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18347 (tail_merge_optimize): Handle current_loops == NULL.
18348
ebbd90d8
JL		 183492013-04-26 Jeff Law <law@redhat.com>
18350
18351 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18352 (simplify_cond_using_ranges): Generalize code to simplify
18353 COND_EXPRs where one argument is a constant and the other
18354 is an SSA_NAME created by an integral type conversion.
18355
8b9b57eb
KT		 183562013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18357
18358 * config/arm/arm.md (store_minmaxsi): Use only when
18359 optimize_insn_for_size_p.
18360
9498e5dc
CB		 183612013-04-29 Christian Bruel <christian.bruel@st.com>
18362
18363 PR target/57108
18364 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18365
6093bc06
RB		 183662013-04-29 Richard Biener <rguenther@suse.de>
18367
18368 PR middle-end/57089
fdca7d03
UB		 18369 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18370 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 18371 (expand_omp_for_generic): Properly add loops.
18372 (expand_omp_for_static_nochunk): Likewise.
18373 (expand_omp_for_static_chunk): Likewise.
18374 (expand_omp_for): For the degenerate case fixup loops.
18375 (expand_omp_sections): Fix default bb placement in loops.
18376 (expand_omp_atomic_pipeline): Properly add loops.
18377
84aacbfd
KT		 183782013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18379
18380 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18381
8dee4479
TV		 183822013-04-29 Tom de Vries <tom@codesourcery.com>
18383
18384 * tree-ssa-tail-merge.c: Update header comment.
18385
47934dc4
JG		 183862013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18387
18388 * config/aarch64/arm_neon.h
18389 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18390 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18391 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18392 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18393 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18394 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18395 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18396 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18397
384be29f
JG		 183982013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18399
18400 * config/aarch64/aarch64-simd.md
18401 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18402 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18403 fix_trunc, fixuns_trunc.
18404 (ftrunc<VDQF:mode>2): New.
18405 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18406 (fix_trunc_optab): New.
18407
0386b123
JG		 184082013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18409
18410 * config/aarch64/aarch64-builtins.c
18411 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18412 iceilf, lround, iroundf.
18413
00fcb892
UB		 184142013-04-29 Uros Bizjak <ubizjak@gmail.com>
18415
18416 PR target/54349
18417 * config/i386/i386.h (enum ix86_tune_indices)
18418 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18419 New, split from X86_TUNE_INTER_UNIT_MOVES.
18420 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18421 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18422 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18423 (TARGET_INTER_UNIT_MOVES): Remove.
18424 * config/i386/i386.c (initial_ix86_tune_features): Update.
18425 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18426 (ix86_expand_convert_uns_didf_sse): Use
18427 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18428 (ix86_expand_vector_init_one_nonzero): Ditto.
18429 (ix86_expand_vector_init_interleave): Ditto.
18430 (inline_secondary_memory_needed): Return true for moves from SSE class
18431 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18432 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18433 * config/i386/constraints.md (Yi, Ym): Depend on
18434 TARGET_INTER_UNIT_MOVES_TO_VEC.
18435 (Yj, Yn): New constraints.
18436 * config/i386/i386.md (*movdi_internal): Change constraints of
18437 operand 1 from Yi to Yj and from Ym to Yn.
18438 (*movsi_internal): Ditto.
18439 (*movdf_internal): Ditto.
18440 (*movsf_internal): Ditto.
18441 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18442 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18443 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18444 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18445 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18446 * config/i386/sse.md (movdi_to_sse): Ditto.
18447 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18448 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18449 TARGET_INTER_UNIT_MOVES.
18450 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18451 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18452 instead of TARGET_INTER_UNIT_MOVES.
18453 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18454 operand 1 from Yi to Yj and from Ym to Yn.
18455
4c871069
JG		 184562013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18457
18458 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18459 (float_truncate_hi_): Likewise.
18460 (float_extend_lo_): Likewise.
18461 (float_truncate_lo_): Likewise.
18462 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18463 (aarch64_float_extend_lo_v2df): Likewise.
18464 (vec_unpacks_hi_v4sf): Likewise.
18465 (aarch64_float_truncate_lo_v2sf): Likewise.
18466 (aarch64_float_truncate_hi_v4sf): Likewise.
18467 (vec_pack_trunc_v2df): Likewise.
18468 (vec_pack_trunc_df): Likewise.
18469
1709ff9b
JG		 184702013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18471
18472 * config/aarch64/aarch64-builtins.c
18473 (aarch64_fold_builtin): Fold float conversions.
18474 * config/aarch64/aarch64-simd-builtins.def
18475 (floatv2si, floatv4si, floatv2di): New.
18476 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18477 * config/aarch64/aarch64-simd.md
18478 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18479 * config/aarch64/iterators.md (FLOATUORS): New.
18480 (optab): Add float, floatuns.
18481 (su_optab): Likewise.
18482
ce966824
JG		 184832013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18484
18485 * config/aarch64/aarch64-builtins.c
18486 (aarch64_builtin_vectorized_function): Use new names for
18487 fcvt builtins.
18488 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18489 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18490 (fcvtzu): Split as...
18491 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18492 (fcvtas): Split as...
18493 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18494 (fcvtau): Split as...
18495 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18496 (fcvtps): Split as...
18497 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18498 (fcvtpu): Split as...
18499 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18500 (fcvtms): Split as...
18501 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18502 (fcvtmu): Split as...
18503 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18504 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18505 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18506 (lfrintnusf, lfrintnudf): Likewise.
18507 * config/aarch64/aarch64-simd.md
18508 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18509 define_insn.
18510 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18511 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18512 (fcvt_pattern): Likewise.
18513
b9de24fe
JG		 185142013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18515
18516 * config/aarch64/aarch64-simd.md
18517 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18518 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18519
77a205be
JG		 185202013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18521
18522 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18523 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18524 (vrnd<a,m,n,p>_f32): Implement using builtins.
18525 (vrnd<i,x><q>_f<32, 64>): New.
18526
0659ce6f
JG		 185272013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18528
18529 * config/aarch64/aarch64-builtins.c
18530 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18531 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18532 (frintz): Rename to...
18533 (btrunc): ...this.
18534 (frintp): Rename to...
18535 (ceil): ...this.
18536 (frintm): Rename to...
18537 (floor): ...this.
18538 (frinti): Rename to...
18539 (nearbyint): ...this.
18540 (frintx): Rename to...
18541 (rint): ...this.
18542 (frinta): Rename to...
18543 (round): ...this.
18544 * config/aarch64/aarch64-simd.md
18545 (aarch64_frint<frint_suffix><mode>): Delete.
18546 (<frint_pattern><mode>2): Convert to insn.
18547 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18548 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18549 (frint_pattern): Likewise.
18550 (frint_suffix): Likewise.
18551
ea78906a
RB		 185522013-04-29 Richard Biener <rguenther@suse.de>
18553
18554 PR tree-optimization/57081
18555 * loop-init.c: Include tree-flow.h.
18556 (loop_optimizer_finalize): Free number of iteration estimates.
18557 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18558
baee1763
JJ		 185592013-04-29 Jakub Jelinek <jakub@redhat.com>
18560
94dc5332
JJ		 18561 PR tree-optimization/57083
18562 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18563 non-singleton shift count range, zero extend low_bound for uns case.
18564
baee1763
JJ		 18565 * config/i386/predicates.md (general_vector_operand): New predicate.
18566 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18567 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18568 if they aren't nonimmediate operands. If their original values
18569 satisfy const_vector_equal_evenodd_p, don't shift them.
18570 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18571 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18572 if not nonimmediate_operand.
18573 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18574 instead of register_operand.
18575 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18576
a9073727 185772013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 18578
18579 * stor-layout.c (finalize_size_functions): Allocate a structure and
18580 reset cfun before dumping the functions.
18581
ba8011e6
JJ		 185822013-04-27 Jakub Jelinek <jakub@redhat.com>
18583
d6fde69e
JJ		 18584 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18585
ba8011e6
JJ		 18586 PR target/56866
18587 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18588 use xop_pmacsdqh if uns_p.
18589 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18590 the immediate rotate count.
18591
2c62cbaa
VM		 185922013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18593
18594 * rtl.h (struct rtx_def): Add comment for field jump.
18595 (LRA_SUBREG_P): New macro.
18596 * recog.c (register_operand): Check LRA_SUBREG_P.
18597 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18598 stack frame.
18599 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18600 (lra_final_code_change): Skip subreg change for operators.
18601 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18602 if there are no operand changes.
18603 * lra-constraints.c (curr_insn_set): New.
18604 (match_reload): Set LRA_SUBREG_P.
18605 (emit_spill_move): Ditto.
18606 (check_and_process_move): Use curr_insn_set. Process only single
18607 set insns. Don't initialize sec_mem_p and change_p.
18608 (simplify_operand_subreg): Use LRA_SUBREG_P.
18609 (reg_in_class_p): New function.
18610 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18611 of #ifdef. Add code to remove cycling.
18612 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18613 non-null disp. Reload inner instead of disp when base and index
18614 are null. Try to put lo_sum into register.
18615 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18616 (check_and_process_move): Move code for move cost check to
18617 simple_move_p. Remove equiv_substitution.
18618 (simple_move_p): New function.
18619 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18620 curr_insn_set. Call check_and_process_move only for single set
18621 insns. Use the new function. Move call of check_and_process_move
18622 after operand equiv substitution and address process.
18623
e7d764f3
JJ		 186242013-04-26 Jakub Jelinek <jakub@redhat.com>
18625
18626 PR go/57045
18627 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18628 with nonlocal goto receivers or returns twice calls, ignore
18629 unininitialized values from abnormal edges to nl goto receiver
18630 or returns twice call.
18631
41e10689
JJ		 186322013-04-26 Jakub Jelinek <jakub@redhat.com>
18633
18634 PR tree-optimization/57051
18635 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18636 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18637 bitsize.
18638
d7ed20db
RB		 186392013-04-26 Richard Biener <rguenther@suse.de>
18640
18641 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18642 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18643 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18644 (fixup_loop_arrays_after_move): New function.
18645 (move_sese_region_to_fn): Properly outline the loop tree parts
18646 of the SESE region.
18647
df93505e
UB		 186482013-04-26 Uros Bizjak <ubizjak@gmail.com>
18649
18650 * config/i386/i386.md (type, unit): Fix long lines.
18651
dd366ec3
RB		 186522013-04-26 Richard Biener <rguenther@suse.de>
18653
18654 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18655 (lto-streamer-out.o): Likewise.
18656 * cfgloop.c (init_loops_structure): Export, add struct function
18657 argument and adjust.
18658 (flow_loops_find): Adjust.
18659 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18660 (init_loops_structure): Declare.
18661 * lto-streamer-in.c: Include cfgloop.h.
18662 (input_cfg): Input the loop tree.
18663 * lto-streamer-out.c: Include cfgloop.h.
18664 (output_cfg): Output the loop tree.
18665 (output_struct_function_base): Do not drop PROP_loops.
18666
a9e0d843
RB		 186672013-03-26 Richard Biener <rguenther@suse.de>
18668
18669 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18670 (pass_build_cfg): Provide PROP_loops.
18671 (move_sese_region_to_fn): Remove loops that are outlined into fn
18672 for now.
18673 * tree-inline.c: Include cfgloop.h.
18674 (initialize_cfun): Do not drop PROP_loops.
18675 (copy_loops): New function.
18676 (copy_cfg_body): Copy loop structure.
18677 (tree_function_versioning): Initialize destination loop tree.
18678 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18679 (pass_parallelize_loops): Do IL verification.
18680 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18681 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18682 the CFG make sure we fixup loops as well.
18683 * tree-ssa-tail-merge.c: Include cfgloop.h.
18684 (replace_block_by): When merging loop latches mark loops for fixup.
18685 * lto-streamer-out.c (output_struct_function_base): Drop
18686 PROP_loops for now.
18687 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18688 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18689 * ipa-split.c: Include cfgloop.h.
18690 (split_function): Add the new return block to the loop tree root.
18691 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18692 whether we have removed the forwarder block.
18693 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18694 * cfgloop.h (place_new_loop): Declare.
18695 * cfgloopmanip.c (place_new_loop): Export.
18696 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18697 (tree-switch-conversion.o): Likewise.
18698 (tree-complex.o): Likewise.
18699 (tree-inline.o): Likewise.
18700 (tree-ssa-tailmerge.o): Likewise.
18701 (ipa-split.o): Likewise.
18702 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18703 (tree-ssa-copy.o): Likewise.
18704 * tree-switch-conversion.c: Include cfgloop.h
18705 (process_switch): If we emit a bit-test cascade, schedule loops
18706 for fixup.
18707 * tree-complex.c: Include cfgloop.h.
18708 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18709 * asan.c: Include cfgloop.h.
18710 (create_cond_insert_point): Properly add new basic-blocks to
18711 loops, schedule loop fixup.
18712 * cfgloop.c (verify_loop_structure): Check that looks are not
18713 marked for fixup.
18714 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18715 to loops.
18716 (expand_omp_for_generic): Likewise.
18717 (expand_omp_sections): Likewise.
18718 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18719 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18720 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18721 is initialized, not when loops are present.
18722 * tree-parloops.c (parallelize_loops): Remove checking here.
18723 * passes.c (init_optimization_passes): Schedule a copy-propagation
18724 pass before complete unrolling of inner loops.
18725
e78e8a0b
JJ		 187262013-04-26 Jakub Jelinek <jakub@redhat.com>
18727
a2e836b2
JJ		 18728 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18729 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18730 (colorize_init): Add argument to _WIN32 version.
18731 * toplev.c: Include diagnostic-color.h.
18732 (process_options): Default to -fdiagnostics-color=auto if
18733 GCC_COLORS env var is in the environment.
18734 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 18735 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18736 env var is in the environment, the default is auto rather than never.
a2e836b2 18737
e78e8a0b
JJ		 18738 * diagnostic.h (file_name_as_prefix): Add context argument.
18739 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18740 the string as locus.
18741 * langhooks.c (lhd_print_error_function): Adjust caller.
18742
013e5ef9
LC		 187432013-04-25 Lawrence Crowl <crowl@google.com>
18744
18745 * var-tracking.c (shared_hash_def::htab):
18746 Change type to hash_table. Update dependent calls and types.
18747
4a8fb1a1
LC		 187482013-04-25 Lawrence Crowl <crowl@google.com>
18749
18750 * Makefile.in: Update as needed below.
18751
18752 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18753 Move declaration to after the type's method definitons.
18754
18755 * attribs.c (htab_t scoped_attributes::attribute_hash):
18756 Change type to hash_table. Update dependent calls and types.
18757
18758 * bitmap.c (htab_t bitmap_desc_hash):
18759 Change type to hash_table. Update dependent calls and types.
18760
18761 * cselib.c (htab_t cselib_hash_table):
18762 Change type to hash_table. Update dependent calls and types.
18763
18764 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18765 (hash_string_slot_node): Move implementation into lto-streamer.h
18766 struct string_slot_hasher.
18767 (eq_string_slot_node): Likewise.
18768
18769 * data-streamer-out.c: Update output_block::string_hash_table
18770 dependent calls and types.
18771
18772 * dwarf2cfi.c (htab_t trace_index):
18773 Change type to hash_table. Update dependent calls and types.
18774
18775 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18776 Change type to hash_table. Update dependent calls and types.
18777 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18778 (htab_t optimize_external_refs::map): Likewise.
18779 (htab_t output_comp_unit::extern_map): Likewise.
18780 (htab_t output_comdat_type_unit::extern_map): Likewise.
18781 (htab_t output_macinfo::macinfo_htab): Likewise.
18782 (htab_t optimize_location_lists::htab): Likewise.
18783 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18784
18785 * except.c (htab_t ehspec_hash_type):
18786 Change type to hash_table. Update dependent calls and types.
18787 (assign_filter_values::ttypes): Likewise.
18788 (assign_filter_values::ehspec): Likewise.
18789 (sjlj_assign_call_site_values::ar_hash): Likewise.
18790 (convert_to_eh_region_ranges::ar_hash): Likewise.
18791
18792 * gcse.c (htab_t pre_ldst_table):
18793 Change type to hash_table. Update dependent calls and types.
18794
18795 * ggc-common.c (htab_t saving_htab):
18796 Change type to hash_table. Update dependent calls and types.
18797 (htab_t loc_hash): Likewise.
18798 (htab_t ptr_hash): Likewise.
18799 (call_count): Rename ggc_call_count.
18800 (call_alloc): Rename ggc_call_alloc.
18801 (loc_descriptor): Rename make_loc_descriptor.
18802 (add_statistics): Rename ggc_add_statistics.
18803
18804 * ggc-common.c (saving_htab):
18805 Change type to hash_table. Update dependent calls and types.
18806
18807 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18808 (push_gimplify_context): Likewise.
18809 (pop_gimplify_context): Likewise.
18810 (struct gimple_temp_hash_elt): Added.
18811 (struct gimplify_hasher): Likewise.
18812 (struct gimplify_ctx.temp_htab):
18813 Change type to hash_table. Update dependent calls and types.
18814
18815 * gimple-fold.c: Include gimplify-ctx.h.
18816
18817 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18818 Change type to hash_table. Update dependent calls and types.
18819 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18820 avoid potential global name collision.
18821
18822 * gimplify.c: Include gimplify-ctx.h.
18823 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18824 (htab_t gimplify_ctx::temp_htab):
18825 Update dependent calls and types for new type hash_table.
18826 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18827 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18828
18829 * gimplify-ctx.h: New.
18830 (struct gimple_temp_hash_elt): Move from gimplify.c.
18831 (class gimplify_hasher): New.
18832 (struct gimplify_ctx): Move from gimple.h.
18833 (htab_t gimplify_ctx::temp_htab):
18834 Change type to hash_table. Update dependent calls and types.
18835
18836 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18837 (htab_t ivs_params::newivs_index):
18838 Change type to hash_table. Update dependent calls and types.
18839 (htab_t ivs_params::params_index): Likewise.
18840 (htab_t print_generated_program::params_index): Likewise.
18841 (htab_t gloog::newivs_index): Likewise.
18842 (htab_t gloog::params_index): Likewise.
18843
18844 * graphite.c: Include graphite-htab.h.
18845 4htab_t graphite_transform_loops::bb_pbb_mapping):
18846 Change type to hash_table. Update dependent calls and types.
18847
18848 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18849 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18850 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18851
18852 * graphite-dependences.c: Include graphite-htab.h.
18853 (loop_is_parallel_p): Change hash table type of parameter.
18854
18855 * graphite-htab.h: New.
18856 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18857 (extern find_pbb_via_hash): Move from graphite-poly.h.
18858 (extern loop_is_parallel_p): Move from graphite-poly.h.
18859 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18860
18861 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18862 (extern loop_is_parallel_p): Move to graphite-htab.h.
18863 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18864
18865 * haifa-sched.c (htab_t delay_htab):
18866 Change type to hash_table. Update dependent calls and types.
18867 (htab_t delay_htab_i2): Likewise.
18868
18869 * ira-color.c (htab_t allocno_hard_regs_htab):
18870 Change type to hash_table. Update dependent calls and types.
18871
18872 * ira-costs.c (htab_t cost_classes_htab):
18873 Change type to hash_table. Update dependent calls and types.
18874
18875 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18876 Change type to hash_table. Update dependent calls and types.
18877
18878 * loop-iv.c (htab_t bivs):
18879 Change type to hash_table. Update dependent calls and types.
18880
18881 * loop-unroll.c (htab_t opt_info::insns_to_split):
18882 Change type to hash_table. Update dependent calls and types.
18883 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18884
18885 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18886 (struct string_slot_hasher): New.
18887 (htab_t output_block::string_hash_table):
18888 Change type to hash_table. Update dependent calls and types.
18889
18890 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18891 (htab_t file_name_hash_table):
18892 Change type to hash_table. Update dependent calls and types.
18893
18894 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18895 calls and types.
18896
18897 * lto-streamer.c (htab_t tree_htab):
18898 Change type to hash_table. Update dependent calls and types.
18899
18900 * omp-low.c: Include gimplify-ctx.h.
18901
18902 * passes.c (htab_t name_to_pass_map):
18903 Change type to hash_table. Update dependent calls and types.
18904 (pass_traverse): Rename to passes_pass_traverse.
18905
18906 * plugin.c (htab_t event_tab):
18907 Change type to hash_table. Update dependent calls and types.
18908
18909 * postreload-gcse.c (htab_t expr_table):
18910 Change type to hash_table. Update dependent calls and types.
18911 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18912
18913 * sese.c (debug_rename_map_1): Make extern.
18914 (htab_t copy_bb_and_scalar_dependences::rename_map):
18915 Change type to hash_table. Update dependent calls and types.
18916
18917 * sese.h (extern debug_rename_map): Move to .c file.
18918
18919 * store-motion.c (htab_t store_motion_mems_table):
18920 Change type to hash_table. Update dependent calls and types.
18921
18922 * trans-mem.c (htab_t tm_new_mem_hash):
18923 Change type to hash_table. Update dependent calls and types.
18924
18925 * tree-browser.c (htab_t TB_up_ht):
18926 Change type to hash_table. Update dependent calls and types.
18927
18928 * tree-cfg.c (htab_t discriminator_per_locus):
18929 Change type to hash_table. Update dependent calls and types.
18930
18931 * tree-complex.c: Include tree-hasher.h
18932 (htab_t complex_variable_components):
18933 Change type to hash_table. Update dependent calls and types.
18934
18935 * tree-eh.c (htab_t finally_tree):
18936 Change type to hash_table. Update dependent calls and types.
18937
18938 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18939 struct int_tree_hasher.
18940 (extern int_tree_map_eq): Likewise.
18941 (uid_decl_map_hash): Removed.
18942 (extern decl_tree_map_eq): Likewise.
18943
18944 * tree-hasher.h: New.
18945 (struct int_tree_hasher): New.
18946 (typedef int_tree_htab_type): New.
18947
18948 * tree-inline.c: Include gimplify-ctx.h.
18949
18950 * tree-mudflap.c: Include gimplify-ctx.h.
18951
18952 * tree-parloops.c: Include tree-hasher.h.
18953 (htab_t eliminate_local_variables_stmt::decl_address):
18954 Change type to hash_table. Update dependent calls and types.
18955 (htab_t separate_decls_in_region::decl_copies): Likewise.
18956
18957 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18958 Change type to hash_table. Update dependent calls and types.
18959
18960 * tree-sra.c (candidates):
18961 Change type to hash_table. Update dependent calls and types.
18962
18963 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18964 in tree-flow.h.
18965 (int_tree_map_hash): Likewise.
18966
18967 * tree-ssa-dom.c (htab_t avail_exprs):
18968 Change type to hash_table. Update dependent calls and types.
18969
18970 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18971 Change type to hash_table. Update dependent calls and types.
18972
18973 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18974 Change type to hash_table. Update dependent calls and types.
18975
18976 * tree-ssa-phiopt.c (seen_ssa_names):
18977 Change type to hash_table. Update dependent calls and types.
18978
18979 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18980 Change type to hash_table. Update dependent calls and types.
18981
18982 * tree-ssa-uncprop.c (equiv):
18983 Change type to hash_table. Update dependent calls and types.
18984
c5a44004
JJ		 189852013-04-25 Jakub Jelinek <jakub@redhat.com>
18986
18987 PR rtl-optimization/57003
18988 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18989 call note_stores with kill_clobbered_value callback again after
18990 killing regs_invalidated_by_call.
18991
09962a4a
JG		 189922013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18993
18994 * config/aarch64/aarch64-simd.md
18995 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18996 (aarch64_simd_bsl<mode>): Likewise.
18997 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18998
ea28bb0b
MP		 189992013-04-25 Marek Polacek <polacek@redhat.com>
19000
19001 PR tree-optimization/57066
3c21604f 19002 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 19003
96659611
JG		 190042013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19005
19006 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19007
9697e620
JG		 190082013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19009
19010 * config/aarch64/aarch64-builtins.c
19011 (aarch64_fold_builtin): New.
19012 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19013 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19014 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19015 * config/aarch64/arm_neon.h
19016 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19017
0ac198d3
JG		 190182013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19019 Tejas Belagod <tejas.belagod@arm.com>
19020
19021 * config/aarch64/aarch64-builtins.c
19022 (aarch64_gimple_fold_builtin): New.
19023 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19024 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19025 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19026 (addvv4sf): Update.
19027 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19028
58cff58c
N		 190292013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19030
df93505e 19031 * config/aarch64/aarch64.md
58cff58c
N		 19032 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19033
7e0228bf
N		 190342013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19035
19036 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19037 (*ngcsi_uxtw): New pattern.
19038
5819f96f 190392013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 19040 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 19041
19042 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19043 (TB_DREG): Add T_V4HF.
19044 (v4hf_UP): New macro.
19045 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 19046 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19047 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19048 built-ins.
df93505e 19049 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19050 (arm_vector_mode_supported_p): Handle V4HF.
19051 (arm_mangle_map): Handle V4HFmode.
19052 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19053 * config/arm/arm_neon_builtins.def: Add entries for
19054 vcvtv4hfv4sf, vcvtv4sfv4hf.
19055 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19056 (neon_vcvtv4hfv4sf): Likewise.
19057 * config/arm/neon-gen.ml: Handle half-precision floating point
19058 features.
19059 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19060 * config/arm/arm_neon.h: Regenerate.
19061 * config/arm/neon.ml (type elts): Add F16.
19062 (type vectype): Add T_float16x4, T_floatHF.
19063 (type vecmode): Add V4HF.
19064 (type features): Add Requires_FP_bit feature.
19065 (elt_width): Handle F16.
19066 (elt_class): Likewise.
19067 (elt_of_class_width): Likewise.
19068 (mode_of_elt): Refactor.
19069 (type_for_elt): Handle F16, fix error messages.
19070 (vectype_size): Handle T_float16x4.
19071 (vcvt_sh): New function.
19072 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19073 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19074 (string_of_mode): Handle V4HF.
19075 * doc/arm-neon-intrinsics.texi: Regenerate.
19076
1ef395e4
JG		 190772013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19078
19079 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19080 format specifier in 'X' case.
19081
41c34e94
AM		 190822013-04-25 Alan Modra <amodra@gmail.com>
19083
19084 PR target/57052
19085 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19086 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19087 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19088 Repeat for many other rotate/shift and mask patterns using subregs.
19089 Name lshiftrt insns.
19090 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19091 on WORDS_BIG_ENDIAN.
19092
b9a7eb5d
AM		 190932013-04-25 Alan Modra <amodra@gmail.com>
19094
19095 * config.gcc: Support little-endian powerpc-linux targets.
19096 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19097 (LINK_OS_LINUX_SPEC): Define.
19098 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19099 Preserve MASK_LITTLE_ENDIAN.
19100 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19101 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19102 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19103 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19104 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19105 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19106 Correct fp word order for little-endian. Don't shift toc entries
19107 smaller than a word for little-endian.
19108 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19109 (bswapdi2 splits): Correct low-part subreg for little-endian.
19110 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19111 low/high where such is correct only for be.
19112 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19113 little-endian for -mcall-aixdesc.
19114
87f73374
AM		 191152013-04-25 Alan Modra <amodra@gmail.com>
19116
19117 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19118 replace_equiv_address_nv.
19119
cabf91cd
AM		 191202013-04-25 Alan Modra <amodra@gmail.com>
19121
19122 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19123
0ae24cc8
VM		 191242013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19125
19126 Revert:
19127 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19128 * rtl.h (struct rtx_def): ...
cabf91cd 19129
77bce07c
VM		 191302013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19131
19132 PR rtl-optimizations/57046
19133 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19134 for multi-reg splits.
19135
0db63e7f
L		 191362013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19137
19138 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19139
3362b6b6
SA		 191402013-04-24 Sterling Augustine <saugustine@google.com>
19141
19142 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19143 (comp_dir_string, debug_str_dwo_section): New.
19144 (DEBUG_STR_DWO_SECTION): Rename to ...
19145 (DEBUG_DWO_STR_SECTION): ... this.
19146 (DEBUG_NORM_STR_SECTION): Delete.
19147 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19148 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19149 (find_AT_string): Move most logic to ...
19150 (find_AT_string_in_table): ... here. New.
19151 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19152 add_skeleton_AT_string. Delete logic.
19153 (output_skeleton_debug_sections): Remove call to
19154 add_top_level_skeleton_die_attrs.
19155 (add_comp_dir_attribute): Move logic to comp_dir_string.
19156 (dwarf2out_init): Initialize debug_str_dwo_section.
19157 (output_indirect_string): Call find_string_form.
19158 (output_indirect_strings): Rewrite.
19159 (prune_unused_types): Empty skeleton_debug_str_hash.
19160 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19161 (dwarf2out_finish): Call output_indirect_strings.
19162
e93e18e9
PC		 191632013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19164
19165 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19166
f6ce35ac
VM		 191672013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19168
cabf91cd 19169 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 19170 (LRA_SUBREG_P): New macro.
19171 * recog.c (register_operand): Check LRA_SUBREG_P.
19172 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19173 stack frame.
19174 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19175 (lra_final_code_change): Skip subreg change for operators.
19176 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19177 if there are no operand changes.
19178 * lra-constraints.c (curr_insn_set): New.
19179 (match_reload): Set LRA_SUBREG_P.
19180 (emit_spill_move): Ditto.
19181 (check_and_process_move): Use curr_insn_set. Process only single
19182 set insns. Don't initialize sec_mem_p and change_p.
19183 (simplify_operand_subreg): Use LRA_SUBREG_P.
19184 (reg_in_class_p): New function.
19185 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19186 of #ifdef. Add code to remove cycling.
19187 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19188 non-null disp. Reload inner instead of disp when base and index
19189 are null. Try to put lo_sum into register.
19190 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 19191 (check_and_process_move): Move code for move cost check to
f6ce35ac 19192 simple_move_p. Remove equiv_substitution.
cabf91cd 19193 (simple_move_p): New function.
f6ce35ac
VM		 19194 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19195 curr_insn_set. Call check_and_process_move only for single set
19196 insns. Use the new function. Move call of check_and_process_move
19197 after operand equiv substitution and address process.
19198
38047d90
JG		 191992013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19200
19201 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19202 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19203 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19204
13f39b2e
PC		 192052013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19206
19207 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19208
97c116dc
MP		 192092013-04-24 Marek Polacek <polacek@redhat.com>
19210
19211 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19212 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19213 (select_loops_exit_conditions): Likewise.
19214 (number_of_iterations_for_all_loops): Likewise.
19215 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19216 (scev_analysis): Likewise.
19217
83082391 192182013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 19219 Chao-ying Fu <fu@mips.com>
83082391 19220
cabf91cd
AM		 19221 * config/mips/micromips.md (jraddiusp): New pattern.
19222 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19223 instruction if possible.
83082391 19224
19e34aa2
AM		 192252013-04-24 Alan Modra <amodra@gmail.com>
19226
19227 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19228
fdb6603c
JB		 192292013-04-24 Julian Brown <julian@codesourcery.com>
19230 Chung-Lin Tang <cltang@codesourcery.com>
19231
19232 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 19233 dependency behavior in enumeration type DIE generation. Add TODO note
19234 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 19235
bf190e8d
LC		 192362013-04-23 Lawrence Crowl <crowl@google.com>
19237
19238 * Makefile.in: Update as needed below.
19239
19240 * hash-table.h (class hash_table):
19241 Correct many methods with parameter types compare_type to the correct
19242 value_type. (Correct code was unlikely to notice the change.)
19243 (hash_table::elements_with_deleted) New.
19244 (class hashtable::iterator): New.
19245 (hashtable::begin()): New.
19246 (hashtable::end()): New.
19247 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19248
19249 * statistics.c (statistics_hashes):
19250 Change type to hash_table. Update dependent calls and types.
19251
19252 * tree-into-ssa.c (var_infos):
19253 Change type to hash_table. Update dependent calls and types.
19254
19255 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19256 Change type to hash_table. Update dependent calls and types.
19257
19258 * tree-ssa-loop-im.c (struct mem_ref.refs):
19259 Change type to hash_table. Update dependent calls and types.
19260
19261 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19262 Change type to hash_table. Update dependent calls and types.
19263
19264 * tree-ssa-sccvn.c (vn_tables_s::nary):
19265 Change type to hash_table. Update dependent calls and types.
19266 (vn_tables_s::phis): Likewise.
19267 (vn_tables_s::references): Likewise.
19268
19269 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19270 (vn_reference_eq): Update parameter and return types.
19271
19272 * tree-ssa-structalias.c (pointer_equiv_class_table):
19273 Change type to hash_table. Update dependent calls and types.
19274 (location_equiv_class_table): Likewise.
19275
19276 * tree-vect-data-refs.c: Consequential changes for making
19277 peeling a hash_table.
19278
19279 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19280 (destroy_loop_vec_info): Dependent hash_table update.
19281
19282 * tree-vectorizer.h (peeling_htab):
19283 Change type to hash_table. Update dependent calls and types.
19284
d70a81dd
SC		 192852013-04-23 Shiva Chen <shiva0217@gmail.com>
19286
cabf91cd
AM		 19287 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19288 to check the register content is equal or not.
19289 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19290 to assign register content record.
19291 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 19292 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 19293 * lra-int.h (struct lra_reg): Add offset member.
19294 (lra_reg_val_equal_p): New static inline function.
19295 (lra_update_reg_val_offset): New static inline function.
19296 (lra_assign_reg_val): New static inline function.
19297 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19298 to assign register content record.
19299 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 19300
b894a1f3
CM		 193012013-04-23 Catherine Moore <clm@codesourcery.com>
19302
19303 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19304 operands. Record compression.
19305
ecd14de9
XDL		 193062013-04-23 Xinliang David Li <davidxl@google.com>
19307
19308 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19309
92e776e9
RB		 193102013-04-23 Richard Biener <rguenther@suse.de>
19311
19312 PR middle-end/57036
19313 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19314 parameter, only add abnormal goto edges from the copied body
19315 if the call could perform abnormal gotos.
19316 (copy_cfg_body): Adjust.
19317
a15ee567
SN		 193182013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19319
19320 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19321
08c52234
AS		 193222013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19323
19324 * coretypes.h (gimple_stmt_iterator): Add struct to make
19325 compatible with C.
19326
999c1171
RB		 193272013-04-23 Richard Biener <rguenther@suse.de>
19328
19329 PR tree-optimization/57026
19330 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19331 from SSA names occuring in abnormal PHI nodes.
19332
53669259
AK		 193332013-04-22 Andi Kleen <ak@linux.intel.com>
19334
19335 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19336
1a0ad150
AK		 193372013-04-22 Andi Kleen <ak@linux.intel.com>
19338
19339 * lto/lto.c (print_lto_report_1): Declare early.
19340 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19341
057f8f20
AK		 193422013-04-22 Andi Kleen <ak@linux.intel.com>
19343
19344 * common.opt (-flto-report-wpa): Add.
19345 * doc/invoke.texi (-flto-report-wpa): Add.
19346 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19347 (lto_main): dito.
19348
473b1e05
XDL		 193492013-04-22 Xinliang David Li <davidxl@google.com>
19350
19351 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19352 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19353 * Makefile.in: New dependency
19354
cabf91cd 19355 David Daney <ddaney.cavm@gmail.com>
b1485a33 19356
cabf91cd
AM		 19357 * configure.ac (gcc_cv_as_micromips_support): Use the
19358 --fatal-warnings option.
19359 * configure: Regenerate.
b1485a33 19360
829d0168
MP		 193612013-04-22 Marek Polacek <polacek@redhat.com>
19362
19363 PR sanitizer/56990
19364 * tsan.c (instrument_expr): Don't instrument expression
19365 in case its size is zero.
19366
6d9b7208
UB		 193672013-04-22 Uros Bizjak <ubizjak@gmail.com>
19368
19369 PR target/57032
19370 Revert:
19371 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19372
19373 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19374
ea679d55
JG		 193752013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19376
19377 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19378 (gimple_stmt_iterator): New typedef.
19379 * gimple.h (gimple_stmt_iterator): Rename to...
19380 (gimple_stmt_iterator_d): ... This.
19381 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19382 trees be valid for GIMPLE and GENERIC.
19383 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19384 * gimple-fold.c (gimple_fold_call): Call target hook
19385 gimple_fold_builtin.
19386 * hooks.c (hook_bool_gsiptr_false): New.
19387 * hooks.h (hook_bool_gsiptr_false): New.
19388 * target.def (fold_stmt): New.
19389 * doc/tm.texi: Regenerate.
19390
88a581da
VM		 193912013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19392
19393 PR target/57018
19394 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19395 a set sp if no stack realignment.
19396
92be22dc
NC		 193972013-04-22 Nick Clifton <nickc@redhat.com>
19398
19399 * config.gcc (tilegx-linux): Extend extra_objs rather than
19400 overwriting it.
19401 (tilepro-linux): Likewise.
19402
0ddec79f
JG		 194032013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19404
19405 * config/aarch64/aarch64-builtins.c
19406 (CF): Remove.
19407 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19408 (VAR<1-12>): Add MAP parameter.
19409 (BUILTIN_*): Likewise.
19410 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19411 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19412 (aarch64_ushl_n<mode>): Likewise.
19413 (aarch64_sshr_n<mode>): Likewise.
19414 (aarch64_ushr_n<mode>): Likewise.
19415 (aarch64_<maxmin><mode>): Likewise.
19416 (aarch64_sqrt<mode>): Likewise.
19417 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19418 (vshr<q>_n_*): Likewise.
19419
0050faf8
JG		 194202013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19421
19422 * config/aarch64/aarch64-builtins.c
19423 (aarch64_simd_builtin_type_mode): Handle SF types.
19424 (sf_UP): Define.
19425 (BUILTIN_GPF): Define.
19426 (aarch64_init_simd_builtins): Handle SF types.
19427 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19428 (frecps): Likewise.
19429 (frecpx): Likewise.
19430 * config/aarch64/aarch64-simd.md
19431 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19432 (aarch64_frecpe<mode>): New.
19433 (aarch64_frecps<mode>): Likewise.
19434 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19435 (v8type): Add frecp<esx>.
19436 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19437 (aarch64_frecps<mode>): Likewise.
19438 * config/aarch64/iterators.md (FRECP): New.
19439 (frecp_suffix): Likewise.
19440 * config/aarch64/arm_neon.h
19441 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19442
0fad3dbc 194432013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 19444
19445 PR target/56995
19446 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19447 (REG_CLASS_NAMES): Idem.
19448 (REG_CLASS_CONTENTS): Idem.
19449 (REGCLASS_HAS_FP_REG): Idem.
19450 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19451 (sh_conditional_register_usage): Idem.
19452
3e8a33f9
JL		 194532013-04-21 Jeff Law <law@redhat.com>
19454
19455 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19456 (ssa_forward_propagate_and_combine): Use it.
19457
f38e1b0a
VM		 194582013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19459
19460 * lra.c: Update the flow chart diagram.
19461
682303da
VM		 194622013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19463
19464 PR rtl-optimization/56847
19465 * lra-constraints.c (process_alt_operands): Discourage alternative
19466 with non-matche doffsettable memory constraint fro memory with
19467 known offset.
19468
f6b64c35
RB		 194692013-04-19 Richard Biener <rguenther@suse.de>
19470
19471 PR tree-optimization/56982
19472 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19473 function.
19474 * gimplify.c (gimplify_call_expr): Notice special calls.
19475 (gimplify_modify_expr): Likewise.
19476 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19477 abnormal control flow receivers.
19478 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19479 in the same way as cfun->has_nonlocal_labels.
19480 (gimple_purge_dead_abnormal_call_edges): Likewise.
19481 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19482 receivers start a basic-block.
19483
01d8bf07
RB		 194842013-04-19 Richard Biener <rguenther@suse.de>
19485
19486 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19487 member ...
19488 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19489 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19490 (SLP_TREE_LOAD_PERMUTATION): Add.
19491 (vect_transform_slp_perm_load): Adjust prototype.
19492 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19493 (vect_free_slp_instance): Likewise.
19494 (vect_create_new_slp_node): Likewise.
19495 (vect_supported_slp_permutation_p): Remove.
19496 (vect_slp_rearrange_stmts): Adjust.
19497 (vect_supported_load_permutation_p): Likewise. Inline
19498 vect_supported_slp_permutation_p here.
19499 (vect_analyze_slp_instance): Compute load permutations per
19500 slp node instead of per instance.
19501 (vect_get_slp_defs): Adjust.
19502 (vect_transform_slp_perm_load): Likewise.
19503 (vect_schedule_slp_instance): Remove redundant code.
19504 (vect_schedule_slp): Remove hack for PR56270, add it ...
19505 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19506 CSE loads for SLP. Adjust.
19507
ede22fc3
GY		 195082013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19509
19510 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19511 spelling in two comments.
19512
67bc84fb
GY		 195132013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19514
19515 PR target/56797
19516 * config/arm/arm.c (load_multiple_sequence): Require SP
19517 as base register for loads if SP is in the register list.
19518
e248d83f
MJ		 195192013-04-19 Martin Jambor <mjambor@suse.cz>
19520
19521 PR tree-optimization/56718
19522 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19523 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19524 and made public. Adjusted all callers.
19525 (ipa_intraprocedural_devirtualization): New function.
19526 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19527 (ipa_intraprocedural_devirtualization): Likewise.
19528 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19529
4891e8f8
RB		 195302013-04-19 Richard Biener <rguenther@suse.de>
19531
19532 PR tree-optimization/57000
19533 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19534
dad89f7c
TG		 195352013-04-19 Terry Guo <terry.guo@arm.com>
19536
19537 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19538 Replace with ...
19539 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19540 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19541 (cortex_m4_fmacs): Use new reservations.
19542 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19543
72ea0d47
VM		 195442013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19545
f1e6512c 19546 PR rtl-optimization/56999
72ea0d47
VM		 19547 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19548 related code.
19549 (lra_coalesce): Remove split_origin_bitmap and related code.
19550 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19551 ranges if necessary.
19552
780a5b71
UB		 195532013-04-18 Uros Bizjak <ubizjak@gmail.com>
19554
19555 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19556 New array.
19557 (ix86_expand_call): Remove clobbered_registers array and use
19558 x86_64_ms_sysv_extra_clobbered_registers instead.
19559 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19560 Declare here.
19561 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19562 predicate.
19563 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19564 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19565 (*call_value_rex64_ms_sysv): Ditto.
19566
6f5a366a
CC		 195672013-04-18 Cary Coutant <ccoutant@google.com>
19568
19569 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19570 parent instead of die_mark.
19571
475b8f37
DN		 195722013-04-18 Diego Novillo <dnovillo@google.com>
19573
19574 * gimple.c (create_gimple_tmp): New.
19575 (get_expr_type): New.
19576 (build_assign): New.
19577 (build_type_cast): New.
19578 * gimple.h (enum ssa_mode): Define.
19579 (gimple_seq_set_location): New.
19580 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 19581 to use build_assign and build_type_cast.
475b8f37 19582
08940f33
RB		 195832013-04-18 Richard Biener <rguenther@suse.de>
19584
19585 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19586 handle negative step. Remove redundant checks.
19587 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19588 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19589 for negative step and grouped loads fail to vectorize.
19590
0e0f87d4
SB		 195912013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19592
19593 * emit-rtl.c (reset_insn_used_flags): New function.
19594 (reset_all_used_flags): Use it.
19595 (verify_insn_sharing): New function.
19596 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19597
4c445590
JJ		 195982013-04-18 Jakub Jelinek <jakub@redhat.com>
19599
19600 PR tree-optimization/56984
19601 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19602 and (x >> M) >= N don't register any assertion if N << M is the
19603 minimum value.
19604
6873ecab
SB		 196052013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19606
19607 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19608 do not delete_insn insns that have not yet been emitted, only
19609 unlink them with remove_insn.
19610 * df-scan.c (df_insn_delete): Revert r197492.
19611
3ccb989e
SB		 196122013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19613
19614 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19615 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19616
196172013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 19618
19619 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19620 define_insn_and_split.
19621 (and_scc,ior_scc,negscc): Likewise.
19622 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19623
3ccb989e 196242013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 19625
19626 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19627 can benefit from using a sequence of LDRD instructions in epilogue
19628 instead of a single LDM instruction.
19629
6d10a203
MLI		 196302013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19631
19632 PR 45688
19633 * doc/extend.texi: Fix typo.
19634
6983e6b5
RB		 196352013-04-17 Richard Biener <rguenther@suse.de>
19636
19637 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19638 (vect_build_slp_tree): ... here.
19639 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19640 match. Remove special-casing of mismatched complex loads.
19641 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19642 re-try the match with swapped commutative operands.
19643 (vect_supported_load_permutation_p): Remove special-casing of
19644 mismatched complex loads.
19645 (vect_analyze_slp_instance): Adjust.
19646
ef23e6a2
RB		 196472013-04-17 Richard Biener <rguenther@suse.de>
19648
19649 PR rtl-optimization/56921
19650 * cfgloop.h (struct loop): Add simple_loop_desc member.
19651 (struct niter_desc): Mark with GTY(()).
19652 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19653 * loop-iv.c (get_simple_loop_desc): Likewise.
19654 (free_simple_loop_desc): Likewise.
19655
19656 Revert
19657 2013-04-16 Richard Biener <rguenther@suse.de>
19658
19659 PR rtl-optimization/56921
19660 * loop-init.c (pass_rtl_move_loop_invariants): Add
19661 TODO_do_not_ggc_collect to todo_flags_finish.
19662 (pass_rtl_unswitch): Same.
19663 (pass_rtl_unroll_and_peel_loops): Same.
19664 (pass_rtl_doloop): Same.
19665
fc6f94f5
EB		 196662013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19667
19668 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19669 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19670 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19671 references.
19672 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19673 * tree-streamer.c (record_common_node): Adjust reference in comment.
19674
10a88311
TG		 196752013-04-17 Terry Guo <terry.guo@arm.com>
19676
19677 * config/arm/cortex-m4.md: Add a new bypass.
19678
6d9b7208 196792013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 19680
19681 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19682 New pattern.
19683 (*subs_<optab><mode>_multp2): New pattern.
19684 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19685 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19686
6d9b7208 196872013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 19688
19689 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19690 (*subs_mul_imm_<mode>): New pattern.
19691
18a6701e
DE		 196922013-04-16 David Edelsohn <dje.gcc@gmail.com>
19693
19694 PR target/56948
19695 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19696 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19697 (vsx_movti_32bit): Change j->wa to O->wa.
19698
07c37b2f
RB		 196992013-04-16 Richard Biener <rguenther@suse.de>
19700
19701 PR rtl-optimization/56921
19702 * loop-init.c (pass_rtl_move_loop_invariants): Add
19703 TODO_do_not_ggc_collect to todo_flags_finish.
19704 (pass_rtl_unswitch): Same.
19705 (pass_rtl_unroll_and_peel_loops): Same.
19706 (pass_rtl_doloop): Same.
19707
0e0f87d4 197082013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 19709
19710 * config/arm/arm.c (emit_multi_reg_push): New declaration
19711 for an existing function.
19712 (arm_emit_strd_push): New function.
19713 (arm_expand_prologue): Used here.
19714 (arm_emit_ldrd_pop): New function.
19715 (arm_expand_epilogue): Used here.
19716 (arm_get_frame_offsets): Update condition.
19717 (arm_emit_multi_reg_pop): Add a special case for load of a single
19718 register with writeback.
19719
5e8e2af4
UB		 197202013-04-16 Uros Bizjak <ubizjak@gmail.com>
19721
19722 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19723 description.
19724
9fd9ccf7
RB		 197252013-04-16 Richard Biener <rguenther@suse.de>
19726
19727 PR tree-optimization/56756
19728 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19729 (first_mem_ref_loc): New.
19730 (execute_sm): Place the load temporarily before a previous
19731 access instead of in the latch edge to ensure its SSA dependencies
19732 are defined at points dominating the load.
19733
96fba521
SB		 197342013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19735
4c8af858
SB		 19736 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19737 correct fix by moving header and footer insn to the footer of
19738 the merged basic block. Clear BB_END of the merged-away block.
19739
96fba521
SB		 19740 PR middle-end/43631
19741 * emit-rtl.c (make_note_raw): New function.
19742 (link_insn_into_chain): New static inline function.
19743 (add_insn): Use it.
19744 (add_insn_before, add_insn_after): Factor insn chain linking code...
19745 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19746 using link_insn_into_chain.
19747 (note_outside_basic_block_p): New helper function for emit_note_after
19748 and emit_note_before.
19749 (emit_note_after): Use nobb variant of add_insn_after if the note
19750 should not be contained in a basic block.
19751 (emit_note_before): Use nobb variant of add_insn_before if the note
19752 should not be contained in a basic block.
19753 (emit_note_copy): Use make_note_raw.
19754 (emit_note): Likewise.
19755 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19756 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19757 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19758 the moved barrier the tail of the basic block it follows.
19759 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19760
7b8265ba
JJ		 197612013-04-15 Jakub Jelinek <jakub@redhat.com>
19762
19763 PR tree-optimization/56962
19764 * gimple-ssa-strength-reduction.c (record_increment): Only set
19765 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19766 either rhs1 or rhs2 is equal to c->base_expr.
19767
5185d248
RB		 197682013-04-15 Richard Biener <rguenther@suse.de>
19769
19770 PR tree-optimization/56933
19771 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19772 member.
19773 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19774 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19775 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19776 dependence check ...
19777 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19778 ... here.
19779 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19780 GROUP_READ_WRITE_DEPENDENCE.
19781
a24243a0
AK		 197822013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19783
19784 * emit-rtl.c (reset_all_used_flags): New function.
19785 (verify_rtl_sharing): Call reset_all_used_flags before and after
19786 performing the checks.
19787
1c50eada
KT		 197882013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19789
19790 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19791 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19792 * config/arm/constraints.md (De): New constraint.
19793 * config/arm/neon.md (anddi3_neon): Delete.
19794 (neon_vand<mode>): Expand to standard anddi3 pattern.
19795 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19796 Move earlier in the file.
19797 (neon_inv_logic_op2): Likewise.
19798 (arm_anddi_operand_neon): New predicate.
19799
e927b6ad
RO		 198002013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19801
19802 * configure.ac (gcc_cv_ld_as_needed): Set
19803 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19804 Use -z ignore, -z record on *-*-solaris2*.
19805 (HAVE_LD_AS_NEEDED): Update comment.
19806 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19807 * configure: Regenerate.
19808 * config.in: Regenerate.
19809 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19810 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19811 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19812 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19813 equivalents. Fix markup.
19814 * doc/tm.texi: Regenerate.
19815
e0ea8797
AH		 198162013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19817
19818 * config/i386/i386.opt: New option mstack-protector-guard=.
19819 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19820 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19821 TARGET_SSP_TLS_GUARD.
19822 * config/i386/i386.c (ix86_option_override_internal): Set
19823 ix86_stack_protector_guard.
19824 * config/i386/i386.md (stack_protect_set): Enable for
19825 TARGET_SSP_TLS_GUARD only.
19826 (stack_protect_set_<mode>): Ditto.
19827 (stack_protect_test): Ditto.
19828 (stack_protect_test_<mode>): Ditto.
19829 * doc/invoke.texi (i386 Option): Document.
19830
811b72f9
EB		 198312013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19832
19833 PR target/56890
19834 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19835 (S_MODES): Set H_MODE bit.
19836 (SF_MODES): Set only S_MODE and SF_MODE bits.
19837 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19838 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19839 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19840 <MODE_FLOAT>: Likewise.
19841
5529fdd6
JY		 198422013-04-15 Joey Ye <joey.ye@arm.com>
19843
19844 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19845
517b1da2
JY		 198462013-04-15 Joey Ye <joey.ye@arm.com>
19847
19848 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19849 for real far jump.
19850 (thumb_far_jump_used_p): Count instruction size and set
19851 far_jump_used.
19852
01007ae0
EB		 198532013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19854
19855 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19856 * resource.c (mark_target_live_regs): Reformat conditional block.
19857
c46f6580
SB		 198582013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19859
19860 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19861 notes, they are emitted only just before final.
19862 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19863
1f397f45
SB		 198642013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19865
19866 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19867 * cfgrtl.c (delete_insn): Call it here instead.
19868 * lra-spills.c (lra_final_code_change): Use delete_insn.
19869 * haifa-sched.c (sched_remove_insn): Likewise.
19870 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19871 returning to the nop pool.
19872 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19873 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19874
58a51369
SB		 198752013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19876
19877 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19878 * doc/tm.texi: Regenerated.
19879
33159866
UB		 198802013-04-12 Uros Bizjak <ubizjak@gmail.com>
19881
19882 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19883 QImode checks.
19884
226e378f
SB		 198852013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19886
19887 * df-core.c (df_find_def): Compare register numbers.
19888 (df_find_use): Likewise.
19889
fafb9b18
VM		 198902013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19891
19892 PR target/56903
19893 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19894 lra_in_progress for return.
19895
9a946fd6
GY		 198962013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19897
19898 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19899 define_insn into define_insn_and_split and emit movsicc patterns.
19900
41b83758
GY		 199012013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19902
19903 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19904
d3afd9aa
RB		 199052013-04-12 Richard Biener <rguenther@suse.de>
19906
19907 * tree-pass.h (TODO_do_not_ggc_collect): New.
19908 * passes.c (execute_one_ipa_transform_pass): Honor
19909 TODO_do_not_ggc_collect.
19910 (execute_one_pass): Likewise.
19911
19912 Revert
19913 2013-04-10 Richard Biener <rguenther@suse.de>
19914
19915 * passes.c (init_optimization_passes): Remove reload pass.
19916 * ira.c (do_reload): Merge into ...
19917 (ira): ... this.
19918 (rest_of_handle_reload): Remove.
19919 (pass_reload): Likewise.
19920 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19921 of reload for vzeroupper pass placement.
19922
06f9b387
JJ		 199232013-04-12 Jakub Jelinek <jakub@redhat.com>
19924
19925 PR tree-optimization/56918
19926 PR tree-optimization/56920
19927 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19928 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19929 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19930 use wide_mul_with_sign method.
19931
953094d2
RB		 199322013-04-12 Richard Biener <rguenther@suse.de>
19933
19934 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19935 not be considered a gimple constant.
19936
bb506982
MG		 199372013-04-12 Marc Glisse <marc.glisse@inria.fr>
19938
19939 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19940 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19941 operations.
19942
4b84d650
JJ		 199432013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19944 Jakub Jelinek <jakub@redhat.com>
19945
19946 * opts.c: Include diagnostic-color.h.
19947 (common_handle_option): Handle OPT_fdiagnostics_color_.
19948 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19949 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19950 (diagnostic-color.o): New.
19951 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19952 (diagnostic_color_rule): New enum.
19953 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19954 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19955 the location string.
19956 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19957 either NULL, or color kind.
19958 * diagnostic-color.c: New file.
19959 * diagnostic-color.h: New file.
19960 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19961 arguments.
19962 * doc/invoke.texi (-fdiagnostics-color): Document.
19963 * pretty-print.h (pp_show_color): Define.
19964 (struct pretty_print_info): Add show_color field.
19965 * diagnostic.c: Include diagnostic-color.h.
19966 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19967 macros. Colorize error:, warning: etc. strings and also the location
19968 string.
19969 (diagnostic_show_locus): Colorize the caret line.
19970 * pretty-print.c: Include diagnostic-color.h.
19971 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19972 inside of %< %> quotes or quoted through q format modifier.
19973
067a1e71
AK		 199742013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19975
33159866 19976 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 19977
33159866 199782013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 19979
19980 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19981 code in CC_NZ mode.
19982 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19983 pattern.
19984
7b55f98f
MP		 199852013-04-11 Marek Polacek <polacek@redhat.com>
19986
19987 PR tree-optimization/48184
33159866 19988 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 19989
966b587e
EB		 199902013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19991
19992 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19993 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19994 (skip_simple_arithmetic): Tidy up.
19995 * tree.h (skip_simple_constant_arithmetic): Declare.
19996
33159866 199972013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 19998
19999 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20000
1d42c1ec
RB		 200012013-04-11 Richard Biener <rguenther@suse.de>
20002
20003 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20004 generate vector constants.
20005
4ba5ea11
RB		 200062013-04-11 Richard Biener <rguenther@suse.de>
20007
20008 PR tree-optimization/56878
20009 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20010 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20011 New function.
20012 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20013 Prefer to align the DR with the most invariant base address.
20014
f0defe58
SKS		 200152013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20016
20017 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20018 comment.
20019
d07458be
JG		 200202013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20021
20022 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20023 floating-point vector comparisons against 0.
20024
146b8692
JJ		 200252013-04-11 Jakub Jelinek <jakub@redhat.com>
20026
20027 PR tree-optimization/56899
20028 * fold-const.c (extract_muldiv_1): Apply distributive law
20029 only if TYPE_OVERFLOW_WRAPS (ctype).
20030
b8578ff7
BC		 200312013-04-11 Bin Cheng <bin.cheng@arm.com>
20032
20033 PR target/56124
20034 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20035 loading has side effect.
20036
0ea8a6f9
SB		 200372013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20038
20039 * config/sparc/sparc.c: Include tree-pass.h.
20040 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20041 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20042 head of file. Change return type. Split off gate function.
20043 (sparc_gate_work_around_errata): New function.
20044 (pass_work_around_errata): New pass definition.
20045 (insert_pass_work_around_errata) New pass insert definition to
20046 insert pass_work_around_errata just after delayed-branch scheduling.
20047 (sparc_option_override): Insert the pass.
20048 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20049
42e37616
DM		 200502013-04-10 David S. Miller <davem@davemloft.net>
20051
89deeb3b
DM		 20052 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20053 or -mcpu=hypersparc.
20054
42e37616
DM		 20055 * target.def (cstore_mode): New hook.
20056 * target.h: Include insn-codes.h
20057 * targhooks.c: Likewise.
20058 (default_cstore_mode): New function.
20059 * targhooks.h: Declare it.
20060 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20061 * doc/tm.texi: Rebuild.
20062 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20063 target hook, rather than inspecting the insn_data.
20064 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20065 (TARGET_CSTORE_MODE): Redefine.
20066 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20067 result patterns.
20068 * config/sparc/predicates.md (cstore_result_operand): New special
20069 predicate.
20070 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20071 Use it for operand 0.
20072 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20073 (*snesi_special): Likewise.
20074 (*snesi_zero): Likewise.
20075 (*seqsi_zero): Likewise.
20076 (*sltu_insn): Likewise.
20077 (*sgeu_insn): Likewise.
20078 (*seqdi_special): Make operand 0 and comparison operation be of
20079 DImode.
20080 (*snedi_special): Likewise.
20081 (*snedi_special_vis3): Likewise.
20082 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20083 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20084 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20085 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20086 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20087 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20088 (*sltu_extend_sp64): Likewise.
20089 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20090 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20091 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20092 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20093 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20094
95ca411e
YZ		 200952013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20096
20097 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20098 (aarch64_start_file): Use the new function.
20099
6782438d 201002013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 20101 Jason Merrill <jason@redhat.com>
6782438d
SKS		 20102
20103 * common.opt: Add -gdwarf.
20104 * opts.c (common_handle_option): Handle it.
20105 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20106
bb313b93
RB		 201072013-04-10 Richard Biener <rguenther@suse.de>
20108
20109 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20110 (execute_one_ipa_transform_pass): But unconditionally here.
20111 (execute_one_pass): And here.
20112 (init_optimization_passes): Remove reload pass.
20113 * tree-pass.h (TODO_ggc_collect): Remove.
20114 (pass_reload): Likewise.
20115 * ira.c (do_reload): Merge into ...
20116 (ira): ... this.
20117 (rest_of_handle_reload): Remove.
20118 (pass_reload): Likewise.
20119 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20120 of reload for vzeroupper pass placement.
20121 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20122 and todo_flags_finish of all passes.
20123
793d9a16
RB		 201242013-04-10 Richard Biener <rguenther@suse.de>
20125
20126 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20127 first_const_oprnd field, rename first_def_type to first_op_type.
20128 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20129 (vect_get_and_check_slp_defs): Always use the type of the
20130 operand. Allow mixed vect_external_def, vect_constant_def types.
20131 (vect_get_constant_vectors): Handle mixed vect_external_def,
20132 vect_constant_def types.
20133
12211b99 201342013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 20135
20136 PR tree-optimization/55524
20137 * tree-ssa-math-opts.c
20138 (convert_mult_to_fma): Don't use an fms construct
20139 when we don't have an fms operation, but fnma, and it looks
20140 likely that we'll be able to use the latter.
20141
12211b99 201422013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 20143
20144 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20145 function.
20146 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20147 inline fail caused by overwritable functions.
20148
34ab4a5b
CJW		 201492013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20150
20151 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20152 unnecessary bits in the constant power of two case.
20153
abf9bfbc
RB		 201542013-04-10 Richard Biener <rguenther@suse.de>
20155
20156 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20157 broken code swapping operands.
20158 (vect_build_slp_tree): Do not compute load permutations here.
20159 (vect_analyze_slp_instance): Compute load permutations here,
20160 after building the SLP tree.
20161
f408477e
CB		 201622013-04-09 Christian Bruel <christian.bruel@st.com>
20163
20164 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20165 of next/prev_real_insn.
20166
5ac42672
JH		 201672013-04-09 Jan Hubicka <jh@suse.cz>
20168
abf9bfbc
RB		 20169 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20170 Drop aliased parameter.
5ac42672
JH		 20171 (function_and_variable_visibility): Do not handle alias pairs.
20172 * cgraph.c (varpool_externally_visible_p): Update prototype.
20173 * varpool.c (varpool_add_new_variable): Update.
20174
5017f1d2
KT		 201752013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20176
20177 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20178
48eecbee
SB		 201792013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20180
a949cf1c
SB		 20181 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20182
48eecbee
SB		 20183 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20184
75ef8e3d
MP		 201852013-04-09 Marek Polacek <polacek@redhat.com>
20186
20187 PR tree-optimization/48762
33159866 20188 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 20189
23847df4
RB		 201902013-04-09 Richard Biener <rguenther@suse.de>
20191
20192 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20193 dealing with cost.
20194 (vect_build_slp_tree): Likewise.
20195 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20196 calculating the cost of a SLP instance.
20197 (vect_analyze_slp_instance): Use it from here, after building
20198 the SLP tree.
20199
ea3a0fde
JJ		 202002013-04-09 Jakub Jelinek <jakub@redhat.com>
20201
20202 PR middle-end/56883
20203 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20204 expand_omp_for_static_chunk): Use simple_p = true in
20205 force_gimple_operand_gsi calls when assigning to addressable decls.
20206
a32dfe9d
JL		 202072013-04-09 Jeff Law <law@redhat.com>
20208
20209 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20210 when the boolean was created by converting a wider object which
20211 had a boolean range.
20212
d755c7ef
RB		 202132013-04-09 Richard Biener <rguenther@suse.de>
20214
20215 * tree-vectorizer.h (slp_void_p): Remove.
20216 (slp_tree): Typedef before _slp_tree declaration.
20217 (struct _slp_tree): Use a vector of slp_tree as children.
20218 (vect_get_place_in_interleaving_chain): Remove.
20219 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20220 Move ...
20221 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20222 and make static.
20223 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20224 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20225 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20226 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20227 Use slp_node instead of slp_void_p and adjust.
20228
3d741091
RB		 202292013-04-09 Richard Biener <rguenther@suse.de>
20230
20231 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20232 work that is not necessary.
20233
39307ba7
JJ		 202342013-04-09 Jakub Jelinek <jakub@redhat.com>
20235
20236 PR tree-optimization/56854
20237 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20238 forward into clobber stmts if it would change MEM_REF lhs into
20239 non-MEM_REF.
20240
343881fd
MK		 202412013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20242
20243 * tree.c (type_hash_lookup, type_hash_add): Make static.
20244 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20245
3922658a
RB		 202462013-04-09 Richard Biener <rguenther@suse.de>
20247
20248 * tree.h (unsave_expr_now): Remove.
20249 * tree-inline.c (mark_local_for_remap_r): Remove.
20250 (unsave_expr_1): Likewise.
20251 (unsave_r): Likewise.
20252 (unsave_expr_now): Likewise.
20253 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20254 (propagate_tree_value): Likewise.
20255
9fb6b620
SB		 202562013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20257
20258 * doc/rtl.texi (sequence): Rewrite documentation to match the
20259 current use of SEQUENCE rtl objects.
20260 * rtl.def (SEQUENCE): Likewise.
20261
20262 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20263 Update documentation.
20264 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20265 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20266
20267 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20268
8ddb5a29
TJ		 202692013-04-08 Teresa Johnson <tejohnson@google.com>
20270
20271 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20272 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 20273 methods.
8ddb5a29
TJ		 20274 (estimate_edge_size_and_time): Add comment to suggest using rounding
20275 methods.
20276 (estimate_node_size_and_time): Ditto.
20277 (remap_edge_change_prob): Use helper rounding divide methods.
20278 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20279 (gimple_mod_pow2_value_transform): Ditto.
20280 (gimple_mod_subtract_transform): Ditto.
20281 (gimple_ic_transform): Ditto.
20282 (gimple_stringops_transform): Ditto.
20283 * stmt.c (conditional_probability): Ditto.
20284 (emit_case_dispatch_table): Ditto.
20285 * lto-cgraph.c (merge_profile_summaries): Ditto.
20286 * tree-optimize.c (execute_fixup_cfg): Ditto.
20287 * cfgcleanup.c (try_forward_edges): Ditto.
20288 * cfgloopmanip.c (scale_loop_profile): Ditto.
20289 (loopify): Ditto.
20290 (duplicate_loop_to_header_edge): Ditto.
20291 (lv_adjust_loop_entry_edge): Ditto.
20292 * tree-vect-loop.c (vect_transform_loop): Ditto.
20293 * profile.c (compute_branch_probabilities): Ditto.
20294 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20295 * lto-streamer-in.c (input_cfg): Ditto.
20296 * gimple-streamer-in.c (input_bb): Ditto.
20297 * ipa-cp.c (update_profiling_info): Ditto.
20298 (update_specialized_profile): Ditto.
20299 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20300 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 20301 rounding methods.
8ddb5a29
TJ		 20302 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20303 (compute_trg_info): Ditto.
20304 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20305 (purge_dead_edges): Ditto.
20306 * loop-unswitch.c (unswitch_loop): Ditto.
20307 * cgraphclones.c (cgraph_clone_edge): Ditto.
20308 (cgraph_clone_node): Ditto.
20309 * tree-inline.c (copy_bb): Ditto.
20310 (copy_edges_for_bb): Ditto.
20311 (initialize_cfun): Ditto.
20312 (copy_cfg_body): Ditto.
20313 (expand_call_inline): Ditto.
20314
661e6bd7
KT		 203152013-04-08 Kai Tietz <ktietz@redhat.com>
20316
20317 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20318 TARGET_CYGWIN64 by TARGET_64BIT.
20319
105766f3
JR		 203202013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20321
20322 * config/epiphany/epiphany.md (GPR_1): New constant.
20323 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20324 * config/epiphany/epiphany.c (gen_compare_reg):
20325 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20326 is already in place.
20327 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20328 Don't require being called during rtl expansion; If y operlaps r0,
20329 return 0.
20330 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20331 (epiphany_expand_epilogue): Likewise.
20332
8afab237
JR		 20333 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20334 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20335 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20336
fa7707d6
JR		 20337 * config/epiphany/constraints.md (CnL): New constraint.
20338 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20339 * config/epiphany/predicates.md (add_operand): Allow 1024.
20340
5138e18d
JR		 20341 * config/epiphany/epiphany.md (logical_op): New code iterator.
20342 (op_mnc): New code attribute.
20343 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20344 (mov_f+1, mov_f+2): New peephole2 patterns.
20345
2ccc703d
JR		 20346 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20347 (cstoresi4): Also allow re-use of zero result when doing a NE
20348 comparison to a non-zero operand.
aefb0819 20349 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20350
093ac0a5
JR		 20351 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20352 Use gen_addsi3_i / gen_subsi3_i.
20353
f223bb13
JJ		 203542013-04-08 Jakub Jelinek <jakub@redhat.com>
20355
20356 PR c++/34949
20357 PR c++/50243
20358 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20359 contain anything but clobbers, at most one __builtin_stack_restore,
20360 optionally debug stmts and final resx, and if it has at least one
20361 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20362 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20363 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20364 which isn't defaut definition, remove them.
20365 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20366 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20367 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20368 with MEM_REF LHS with SSA_NAME address.
20369
4481581f
JL		 203702013-04-08 Jeff Law <law@redhat.com>
20371
20372 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20373
451dabda
RB		 203742013-04-08 Richard Biener <rguenther@suse.de>
20375
20376 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20377 extra newline.
20378 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20379 determined vector type.
20380 (vect_analyze_data_refs): Likewise.
20381 (vect_get_new_vect_var): Adjust.
20382 (vect_create_destination_var): Preserve SSA name versions.
20383 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20384 not dump anything here.
20385
3b088b47
JR		 203862013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20387
20388 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20389 Add member lr_slot_known.
20390 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20391 if necessary.
20392 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20393 Remove code that sets lr_slot_offset according to what a previous
20394 version of epiphany_emit_save_restore used to do.
20395 (epiphany_emit_save_restore): When doing an lr save or restore,
20396 set/verify lr_slot_known and lr_slot_offset.
20397
d8484d41
XQ		 203982013-04-08 Xinyu Qi <xyqi@marvell.com>
20399
33159866
UB		 20400 PR target/54338
20401 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 20402 in ALL_REGS.
20403
481be1c4
RB		 204042013-04-08 Richard Biener <rguenther@suse.de>
20405
20406 * alias.c (find_base_term): Fix thinko in previous change.
20407
401f3a81
JJ		 204082013-04-08 Jakub Jelinek <jakub@redhat.com>
20409
20410 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20411 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20412 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20413 if possible to compute val.
20414 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20415 For QImode integers don't require anything about precision. Use
20416 const_with_all_bytes_same to find out if the constant doesn't have
20417 repeated bytes in it.
20418
03ed99a8
AK		 204192013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20420
20421 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20422 within mode size.
20423
781b2e62
MP		 204242013-04-08 Marek Polacek <polacek@redhat.com>
20425
20426 PR rtl-optimization/48182
20427 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20428 value to 1.
20429
27e430a2
JDA		 204302013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20431
20432 PR target/55487
20433 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20434 nuses, make sure we have a label.
20435
4902aa64
BS		 204362013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20437
20438 PR target/56843
20439 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20440 (rs6000_emit_swdiv_low_precision): Remove.
20441 (rs6000_emit_swdiv): Rewrite to handle between one and four
20442 iterations of Newton-Raphson generally; modify required number of
20443 iterations for some cases.
20444 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20445
7bca81dc
SB		 204462013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20447
20448 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20449 set-but-unused variable.
20450
20451 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20452 basic blocks of released function bodies garbage-collectable.
20453
20454 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20455 (struct rtl_opt_pass): Add TODO_df_finish.
20456
20457 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20458
4542a38a
GY		 204592013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20460
20461 * config/arm/constraints.md (q): New constraint.
20462 * config/arm/ldrdstrd.md: New file.
20463 * config/arm/arm.md (ldrdstrd.md) New include.
20464 (arm_movdi): Use "q" instead of "r" constraint
20465 for double-word memory access.
20466 (movdf_soft_insn): Likewise.
20467 * config/arm/vfp.md (movdi_vfp): Likewise.
20468 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20469 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 20470 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20471 (mem_ok_for_ldrd_strd): Likewise.
20472 (output_move_double): Update assertion.
20473
2385b218
GY		 204742013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20475
20476 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20477
75fe1cb5
GY		 204782013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20479
20480 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20481 define_insn_and_split.
20482 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20483
dd660e8e
GY		 204842013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20485
20486 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20487 define_insn_and_split.
33159866 20488 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 20489 (shiftsi3_compare): New pattern.
20490 (rrx): New pattern.
20491 * config/arm/unspecs.md (UNSPEC_RRX): New.
20492
045e472c
GY		 204932013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20494
20495 * config/arm/arm.md (negdi_extendsidi): New pattern.
20496 (negdi_zero_extendsidi): Likewise.
20497
3f3bf1a8
GY		 204982013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20499
20500 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20501 define_insn_and_split.
20502 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20503 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20504
b0b49556
GY		 205052013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20506
20507 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20508 define_insn_and_split.
20509 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20510 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20511
d633dd84
GY		 205122013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20513
20514 * config/arm/arm.md (subsi3_carryin): New pattern.
20515 (subsi3_carryin_const): Likewise.
20516 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20517 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20518
f4499066
GY		 205192013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20520
20521 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20522
ceef6fd9
GY		 205232013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20524
20525 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20526 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20527
ddbdd8a7
KT		 205282013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20529
20530 * config/arm/arm.c (arm_expand_builtin): Change fcode
20531 type to unsigned int.
20532
8456d78a
RR		 205332013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20534
20535 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20536
526c230f
ILT		 205372013-04-04 Ian Lance Taylor <iant@google.com>
20538
33159866
UB		 20539 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20540 language standard.
526c230f 20541
3cfbe04d
SB		 205422013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20543
20544 PR middle-end/56729
20545 * df-scan.c (df_insn_delete): Disable failing assert.
20546
dfa3f8d0
KT		 205472013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20548
20549 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20550 New function prototype.
20551 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20552 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20553 (arm_builtin_vectorized_function): New function.
20554
f7837758
KT		 205552013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20556
20557 * config/arm/arm_neon_builtins.def: New file.
20558 * config/arm/arm.c (neon_builtin_data): Move contents to
20559 arm_neon_builtins.def.
20560 (enum arm_builtins): Include neon builtin definitions.
20561 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 20562 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 20563
39c1b6db
MP		 205642013-04-04 Marek Polacek <polacek@redhat.com>
20565
20566 PR tree-optimization/48186
20567 * predict.c (maybe_hot_frequency_p): Return false if
20568 HOT_BB_FREQUENCY_FRACTION is 0.
20569 (cgraph_maybe_hot_edge_p): Likewise.
20570
314f64eb
RB		 205712013-04-04 Richard Biener <rguenther@suse.de>
20572
20573 PR tree-optimization/56826
20574 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20575 more accurately.
20576
51a905b2
RB		 205772013-04-04 Richard Biener <rguenther@suse.de>
20578
20579 PR tree-optimization/56213
20580 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 20581 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 20582
f1bf4f3a
RB		 205832013-04-04 Richard Biener <rguenther@suse.de>
20584
20585 PR tree-optimization/56837
20586 * tree-loop-distribution.c (classify_partition): For non-zero
20587 values require that the value has the same precision as its
20588 mode to be useful as memset value.
20589
0bca7ded
NC		 205902013-04-03 Nick Clifton <nickc@redhat.com>
20591
33159866 20592 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 20593 (fmssf4): Use fmsf.s on E3V5 architectures.
20594 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20595 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20596
b4019227
JL		 205972013-04-03 Jeff Law <law@redhat.com>
20598
20599 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20600 (lra-eliminations.o): Likewise.
20601
f57ddb5b
TJ		 206022013-04-03 Teresa Johnson <tejohnson@google.com>
20603
20604 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 20605 compute_working_sets here from profile.c.
f57ddb5b 20606 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 20607 (gcov_working_set_t): Moved typedef here from basic-block.h
20608 (compute_working_set): Declare.
f57ddb5b
TJ		 20609 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20610 (get_working_sets): Renamed from compute_working_set,
33159866 20611 replace most of body with call to new compute_working_sets.
f57ddb5b 20612 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 20613 to get_working_sets.
20614 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 20615 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 20616 to get_working_sets.
f57ddb5b
TJ		 20617 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20618 * gcov-dump.c (dump_working_sets): New function.
20619
12211b99 206202013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 20621
20622 * hwint.c (sext_hwi, zext_hwi): New functions.
20623 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20624 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20625 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20626 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20627 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20628 (sext_hwi, zext_hwi): New functions.
0bca7ded 20629
be672e08
JL		 206302013-04-03 Jeff Law <law@redhat.com>
20631
20632 PR tree-optimization/56799
33159866
UB		 20633 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20634 back test for widening conversion erroneously dropped in prior change.
be672e08 20635
9d821fa5
KT		 206362013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20637
20638 PR target/56809
20639 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20640 instead of next_real_insn.
20641
71c581e7
MP		 206422013-04-03 Marek Polacek <polacek@redhat.com>
20643
20644 PR sanitizer/55702
33159866 20645 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 20646
4a32ef80
KT		 206472013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20648
20649 PR target/56809
20650 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20651 next_real_insn.
20652 (thumb1_output_casesi): Likewise.
20653 (thumb2_output_casesi): Likewise.
20654
1b2253d4
RB		 206552013-04-03 Richard Biener <rguenther@suse.de>
20656
20657 PR tree-optimization/56817
20658 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20659 Split out ...
20660 (tree_unroll_loops_completely_1): ... new function to manually
20661 walk the loop tree, properly defering outer loops of unrolled
20662 loops to later iterations.
20663
38000232
MG		 206642013-04-03 Marc Glisse <marc.glisse@inria.fr>
20665
20666 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20667 (vectorizable_load): Likewise.
20668 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20669 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20670
3fa35298
MG		 206712013-04-03 Marc Glisse <marc.glisse@inria.fr>
20672
20673 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20674 BIT_FIELD_REF.
20675
b3d45ff0
UW		 206762013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20677
20678 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20679
ec9202a8
BC		 206802013-04-03 Bin Cheng <bin.cheng@arm.com>
20681
20682 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20683
6805bd36
MG		 206842013-04-03 Marc Glisse <marc.glisse@inria.fr>
20685
20686 PR tree-optimization/56790
33159866
UB		 20687 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20688 folding.
6805bd36 20689
da694a77
MG		 206902013-04-03 Marc Glisse <marc.glisse@inria.fr>
20691
20692 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20693 Handle VEC_MERGE.
20694 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20695 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20696 equal arguments.
20697
4e7d7b3d
JJ		 206982013-04-03 Jakub Jelinek <jakub@redhat.com>
20699
20700 PR c/19449
20701 * tree.h (force_folding_builtin_constant_p): New decl.
20702 * builtins.c (force_folding_builtin_constant_p): New variable.
20703 (fold_builtin_constant_p): Fold immediately also if
20704 force_folding_builtin_constant_p.
20705
e6c9d234
RB		 207062013-04-03 Richard Biener <rguenther@suse.de>
20707
20708 PR tree-optimization/56812
20709 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20710 DRs of the same interleaving chain are independent.
20711
984d07dd
JM		 207122013-04-02 Jason Merrill <jason@redhat.com>
20713
20714 * gdbinit.in (pbb): Use debug fn.
20715
622849c9
LC		 207162013-04-02 Lawrence Crowl <crowl@google.com>
20717
20718 * sese.h (struct ivtype_map_elt_s): Remove unused.
20719 (extern debug_ivtype_map): Remove unused.
20720 (extern eq_ivtype_map_elts): Remove unused.
20721 * sese.c (debug_ivtype_map): Removed unused.
20722 (debug_ivtype_map_1): Removed unused.
20723 (debug_ivtype_elt): Remove unused.
20724 (eq_ivtype_map_elts): Remove unused.
20725
20726
82c0e1a0
KT		 207272013-04-02 Kai Tietz <ktietz@redhat.com>
20728
20729 PR target/52790
20730 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20731 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 20732 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20733 function.
82c0e1a0
KT		 20734 (legitimize_pe_coff_symbol): Likewise.
20735 (is_imported_p): New helper-function.
20736 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20737 for Windows x64 targets.
20738 (ix86_expand_prologue): Optimize for pe-coff targets.
20739 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20740 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20741 medium/large code-model.
20742 (legitimize_pic_address): Likewise.
20743 (legitimize_tls_address): Likewise.
20744 (ix86_expand_call): Likewise.
20745 (x86_output_mi_thunk): Likewise.
20746 (get_dllimport_decl): Add new beimport argument.
20747 (construct_plt_address): Don't assert for x64 pe-coff targets.
20748 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20749 targets.
20750 (SYMBOL_FLAG_STUBVAR): New macro.
20751 (SYMBOL_REF_STUBVAR_P): Likewise.
20752 * config/i386/winnt.c (stub_list): New structure.
20753 (stub_head): New local variable.
20754 (i386_pe_record_stub): New function.
20755 (i386_pe_file_end): Emit refptr-stubs.
20756
5d751b0c
JJ		 207572013-04-02 Jakub Jelinek <jakub@redhat.com>
20758
09bb4c99
JJ		 20759 PR rtl-optimization/56745
20760 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20761 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20762
a1d8947a
JJ		 20763 PR c++/34949
20764 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20765 and both of them are MEM_REFs, just compare first argument for
20766 equality and attempt to deal even with differing offsets.
20767
5d751b0c
JJ		 20768 PR c++/34949
20769 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20770 of gimple_clobber_p to be MEM_REF.
20771 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20772 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20773 after gimplification.
20774 * asan.c (get_mem_ref_of_assignment): Don't instrument
20775 gimple_clobber_p stmts.
20776 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20777 gimple_clobber_p stmt if they have MEM_REF lhs and
20778 are dead because of another gimple_clobber_p stmt.
20779 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20780 gimple_clobber_p stmts like debug stmts.
20781 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20782 that refer to unused VAR_DECLs or uninitialized values.
20783 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20784 gimple_clobber_p stmts if they refer to removed parameters.
20785 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20786 formatting.
20787
e594716a
UB		 207882013-04-02 Uros Bizjak <ubizjak@gmail.com>
20789
20790 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20791 using SWI48 mode attribute.
20792
7a80735b
WM		 207932013-04-02 Wei Mi <wmi@google.com>
20794
20795 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20796 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20797 *<rotate_insn><mode>3_mask in i386.md.
20798
f423a9e4
AI		 207992013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20800
20801 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20802
90eb75f2
RB		 208032013-04-02 Richard Biener <rguenther@suse.de>
20804
20805 PR tree-optimization/56778
20806 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20807 Runtime alias tests are not supported for gather loads.
20808 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20809 stmts referenced from SSA operands before updating SSA form.
20810
d8c69a92
IC		 208112013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20812 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20813
20814 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 20815 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20816 * config/arm/cortex-a53.md: New file.
20817 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20818 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20819 * config/arm/arm.c (arm_issue_rate): Likewise.
20820 * config/arm/arm-tune.md: Regenerate
20821 * config/arm/arm-tables.opt: Regenerate.
20822 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 20823
239eb04c
ZC		 208242013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20825
20826 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20827 non-static link.
20828
c902d3c8
SN		 208292013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20830
d8c69a92
IC		 20831 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20832 scalar load/store operations using B/H registers.
c902d3c8
SN		 20833 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20834
051d0e2f
SN		 208352013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20836
20837 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20838 scalar move.
20839 * config/aarch64/aarch64.c
20840 (aarch64_simd_scalar_immediate_valid_for_move): New.
20841 * config/aarch64/aarch64-protos.h
20842 (aarch64_simd_scalar_immediate_valid_for_move): New.
20843 * config/aarch64/constraints.md (Dh, Dq): New.
20844 * config/aarch64/iterators.md (hq): New.
20845
0ee1e3d9
EB		 208462013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20847
20848 * reorg.c (get_branch_condition): Deal with conditional returns.
20849 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20850
136fb3f7
WM		 208512013-04-01 Wei Mi <wmi@google.com>
20852
20853 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20854 Truncate operand 2 using %b asm operand modifier.
20855 (*<shift_insn><mode>3_mask): Ditto.
20856 (*<rotate_insn><mode>3_mask): Ditto.
20857
6388c738
SB		 208582013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20859
20860 PR middle-end/56798
20861 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20862
f7a4d826
KK		 208632013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20864
20865 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20866 of next_real_insn.
20867 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20868
4b943a49
LC		 208692013-03-30 Lawrence Crowl <crowl@google.com>
20870
20871 * dse.c (clear_alias_sets): Remove never set.
20872 (disqualified_clear_alias_sets): Remove never set.
20873 (clear_alias_mode_pool): Remove never set.
20874 (dse_step0): Remove condition that is never true.
20875 (canon_address): Remove condition that is never true.
20876 (dse_step7): Remove condition that is never true.
20877 (rest_of_handle_dse): Remove condition that is never true.
20878 (rest_of_handle_dse::did_global): Remove never read from above.
20879 (dse_step2_spill): Remove never called from above.
20880 (dse_step5_spill): Remove never called from above.
20881
39718607
SB		 208822013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20883
da5c6bde
SB		 20884 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20885 JUMP_TABLE_DATA changes.
20886 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20887 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20888 (Insns) <jump_table_data>: New entry.
20889 * doc/tm.texi: Regenerate.
20890
39718607
SB		 20891 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20892
20893 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20894 for table jump at the end of a basic block using tablejump_p.
20895 * targhooks.c (default_invalid_within_doloop): Likewise.
20896 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20897 target hook implementation that is identical to the default hook.
20898 (rs6000_invalid_within_doloop): Remove.
20899
20900 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20901 unused variable from tablejump_p call.
20902
20903 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20904 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20905 (INSN_DELETED_P): Likewise.
20906 (emit_jump_table_data): New prototype.
20907 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20908 after 4th as unused.
20909 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20910 * sched-vis.c (print_insn): Likewise.
20911 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20912 insn for compatibility with back ends that use next_active_insn to
20913 identify jump table data.
20914 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20915 (remove_insn): Likewise.
20916 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20917 to be emitted.
20918 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20919 (emit_jump_table_data): New function.
20920
20921 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20922 basic block, a JUMP_TABLE_DATA never is.
20923 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20924 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20925 off from code handling real insns.
20926 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20927 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20928 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20929 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20930 is not a NONDEBUG_INSN_P.
20931 * ira-costs.c (scan_one_insn): Likewise.
20932 * jump.c (mark_all_labels): Likewise.
20933 (mark_jump_label_1): Likewise.
20934 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20935 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20936 a basic block.
20937 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20938 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20939 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20940 JUMP_TABLE_DATA_P insns.
20941 (calculate_elim_costs_all_insns): Likewise.
20942 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20943 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20944 (delete_output_reload): Code style fixups.
20945 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20946 insn flags on this non-insn.
20947 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20948 as scheduling barriers, for pre-change compatibility.
20949 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20950 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20951
20952 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20953 redundant JUMP_TABLE_DATA_P test.
20954 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20955 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20956 (frv_for_each_packet): Likewise.
20957 * config/i386/i386.c (min_insn_size): Likewise.
20958 (ix86_avoid_jump_mispredicts): Likewise.
20959 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20960 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20961 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20962 (mips16_insn_length): Robustify.
20963 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20964 (mips16_split_long_branches): Likewise.
20965 * config/pa/pa.c (pa_combine_instructions): Likewise.
20966 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20967 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20968 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20969 as contributing to pool range lengths.
20970 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20971 Remove redundant JUMP_TABLE_DATA_P test.
20972 (sh_loop_align): Likewise.
20973 (split_branches): Likewise.
20974 (sh_insn_length_adjustment): Likewise.
20975 * config/spu/spu.c (get_branch_target): Likewise.
20976
0208f7da
JH		 209772013-03-29 Jan Hubicka <jh@suse.cz>
20978
20979 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20980 gcov streaming; stream hot bb threshold to ltrans.
20981 * predict.c (get_hot_bb_threshold): Break out from ....
20982 (maybe_hot_count_p): ... here.
20983 (set_hot_bb_threshold): New function.
20984 * lto-section-in.c (lto_section_name): Add profile.
20985 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20986 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20987 and data-streamer.h
20988 (histogram_entry): New structure.
20989 (histogram, histogram_pool): New global vars.
20990 (histogram_hash): New structure.
20991 (histogram_hash::hash): New method.
20992 (histogram_hash::equal): Likewise.
20993 (account_time_size): New function.
20994 (cmp_counts): New function.
20995 (dump_histogram): New function.
20996 (ipa_profile_generate_summary): New function.
20997 (ipa_profile_write_summary): New function.
20998 (ipa_profile_read_summary): New function.
20999 (ipa_profile): Decide on threshold.
e594716a
UB		 21000 (pass_ipa_profile): Add ipa_profile_write_summary and
21001 ipa_profile_read_summary.
0208f7da
JH		 21002 * Makefile.in (ipa.o): Update dependencies.
21003 * lto-streamer.h (LTO_section_ipa_profile): New section.
21004
5a6ccc94
GDR		 210052013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21006
21007 * tree.h (VAR_P): New.
21008
39385fa6
PC		 210092013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21010
21011 PR lto/56777
21012 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21013
34f0d87a
SB		 210142013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21015
21016 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21017 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21018 (control_flow_insn_p): Likewise.
21019 * cfgrtl.c (duplicate_insn_chain): Likewise.
21020 * final.c (get_attr_length_1): Likewise.
21021 (shorten_branches): Likewise.
21022 (final_scan_insn): Likewise.
21023 * function.c (instantiate_virtual_regs): Likewise.
21024 * gcse.c (insert_insn_end_basic_block): Likewise.
21025 * ira-costs.c (scan_one_insn): Likewise.
21026 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21027 * lra.c (check_rtl): Likewise.
21028 * reload1.c (elimination_costs_in_insn): Likewise.
21029 * reorg.c (follow_jumps): Likewise.
21030
21031 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21032 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21033 (thumb_far_jump_used_p): Likewise.
21034 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21035 (workaround_speculation): Likewise.
21036 (add_sched_insns_for_speculation): Likewise.
21037 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21038 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21039 (frv_for_each_packet): Likewise.
21040 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21041 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21042 (final_emit_insn_group_barriers): Likewise.
21043 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21044 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21045 (mips16_insn_length): Likewise.
21046 * config/pa/pa.c (pa_reorg): Likewise.
21047 (pa_combine_instructions): Likewise.
21048 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21049 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21050 (sh_reorg): Likewise.
21051 (split_branches): Likewise.
21052 * config/spu/spu.c (get_branch_target): Likewise.
21053
21054 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21055 JUMP_TABLE_DATA_P.
21056
4ac761b0
KY		 210572013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21058
39385fa6 21059 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 21060 Fix declaration name.
21061
58a49189
LC		 210622013-03-28 Lawrence Crowl <crowl@google.com>
21063
21064 * graphds.h (struct graph.indicies): Remove unused.
21065 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21066 (SCOP_ORIGINAL_PDDRS): Remove unused.
21067 * sese.h (extern insert_loop_close_phis): Removed unused.
21068 (extern insert_guard_phis): Removed unused.
21069 (extern ivtype_map_elt_info): Removed unused.
21070 (new_ivtype_map_elt): Removed unused.
21071 * sese.c (ivtype_map_elt_info): Removed unused.
21072
7b3b6ae4
LC		 210732013-03-28 Lawrence Crowl <crowl@google.com>
21074
21075 * Makefile.in: Add several missing include dependences.
21076 (DUMPFILE_H): New.
21077 (test-dump.o): New. This object is not added to any executable,
21078 but is present for ad-hoc testing.
21079 * bitmap.c
21080 (debug (const bitmap_head_def &)): New.
21081 (debug (const bitmap_head_def *)): New.
21082 * bitmap.h
21083 (extern debug (const bitmap_head_def &)): New.
21084 (extern debug (const bitmap_head_def *)): New.
21085 * cfg.c
21086 (debug (edge_def &)): New.
21087 (debug (edge_def *)): New.
21088 * cfghooks.c
21089 (debug (basic_block_def &)): New.
21090 (debug (basic_block_def *)): New.
21091 * dumpfile.h
21092 (dump_node (const_tree, int, FILE *)): Correct source file.
21093 * dwarf2out.c
21094 (debug (die_struct &)): New.
21095 (debug (die_struct *)): New.
21096 * dwarf2out.h
21097 (extern debug (die_struct &)): New.
21098 (extern debug (die_struct *)): New.
21099 * gimple-pretty-print.c
21100 (debug (gimple_statement_d &)): New.
21101 (debug (gimple_statement_d *)): New.
21102 * gimple-pretty-print.h
21103 (extern debug (gimple_statement_d &)): New.
21104 (extern debug (gimple_statement_d *)): New.
21105 * ira-build.c
21106 (debug (ira_allocno_copy &)): New.
21107 (debug (ira_allocno_copy *)): New.
21108 (debug (ira_allocno &)): New.
21109 (debug (ira_allocno *)): New.
21110 * ira-int.h
21111 (extern debug (ira_allocno_copy &)): New.
21112 (extern debug (ira_allocno_copy *)): New.
21113 (extern debug (ira_allocno &)): New.
21114 (extern debug (ira_allocno *)): New.
21115 * ira-lives.c
21116 (debug (live_range &)): New.
21117 (debug (live_range *)): New.
21118 * lra-int.h
21119 (debug (lra_live_range &)): New.
21120 (debug (lra_live_range *)): New.
21121 * lra-lives.c
21122 (debug (lra_live_range &)): New.
21123 (debug (lra_live_range *)): New.
21124 * omega.c
21125 (debug (omega_pb_d &)): New.
21126 (debug (omega_pb_d *)): New.
21127 * omega.h
21128 (extern debug (omega_pb_d &)): New.
21129 (extern debug (omega_pb_d *)): New.
21130 * print-rtl.c
21131 (debug (const rtx_def &)): New.
21132 (debug (const rtx_def *)): New.
21133 * print-tree.c
21134 (debug_tree (tree): Move within file.
21135 (debug_raw (const tree_node &)): New.
21136 (debug_raw (const tree_node *)): New.
21137 (dump_tree_via_hooks (const tree_node *, int)): New.
21138 (debug (const tree_node &)): New.
21139 (debug (const tree_node *)): New.
21140 (debug_verbose (const tree_node &)): New.
21141 (debug_verbose (const tree_node *)): New.
21142 (debug_head (const tree_node &)): New.
21143 (debug_head (const tree_node *)): New.
21144 (debug_body (const tree_node &)): New.
21145 (debug_body (const tree_node *)): New.
21146 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21147 (debug (vec<tree, va_gc> &)): New.
21148 (debug (vec<tree, va_gc> *)): New.
21149 * rtl.h
21150 (extern debug (const rtx_def &)): New.
21151 (extern debug (const rtx_def *)): New.
21152 * sbitmap.c
21153 (debug_raw (simple_bitmap_def &)): New.
21154 (debug_raw (simple_bitmap_def *)): New.
21155 (debug (simple_bitmap_def &)): New.
21156 (debug (simple_bitmap_def *)): New.
21157 * sbitmap.h
21158 (extern debug (simple_bitmap_def &)): New.
21159 (extern debug (simple_bitmap_def *)): New.
21160 (extern debug_raw (simple_bitmap_def &)): New.
21161 (extern debug_raw (simple_bitmap_def *)): New.
21162 * sel-sched-dump.c
21163 (debug (vinsn_def &)): New.
21164 (debug (vinsn_def *)): New.
21165 (debug_verbose (vinsn_def &)): New.
21166 (debug_verbose (vinsn_def *)): New.
21167 (debug (expr_def &)): New.
21168 (debug (expr_def *)): New.
21169 (debug_verbose (expr_def &)): New.
21170 (debug_verbose (expr_def *)): New.
21171 (debug (vec<rtx> &)): New.
21172 (debug (vec<rtx> *)): New.
21173 * sel-sched-dump.h
21174 (extern debug (vinsn_def &)): New.
21175 (extern debug (vinsn_def *)): New.
21176 (extern debug_verbose (vinsn_def &)): New.
21177 (extern debug_verbose (vinsn_def *)): New.
21178 (extern debug (expr_def &)): New.
21179 (extern debug (expr_def *)): New.
21180 (extern debug_verbose (expr_def &)): New.
21181 (extern debug_verbose (expr_def *)): New.
21182 (extern debug (vec<rtx> &)): New.
21183 (extern debug (vec<rtx> *)): New.
21184 * sel-sched-ir.h
21185 (_list_iter_cond_expr): Make inline instead of static.
21186 * sreal.c
21187 (debug (sreal &)): New.
21188 (debug (sreal *)): New.
21189 * sreal.h
21190 (extern debug (sreal &)): New.
21191 (extern debug (sreal *)): New.
21192 * tree.h
21193 (extern debug_raw (const tree_node &)): New.
21194 (extern debug_raw (const tree_node *)): New.
21195 (extern debug (const tree_node &)): New.
21196 (extern debug (const tree_node *)): New.
21197 (extern debug_verbose (const tree_node &)): New.
21198 (extern debug_verbose (const tree_node *)): New.
21199 (extern debug_head (const tree_node &)): New.
21200 (extern debug_head (const tree_node *)): New.
21201 (extern debug_body (const tree_node &)): New.
21202 (extern debug_body (const tree_node *)): New.
21203 (extern debug (vec<tree, va_gc> &)): New.
21204 (extern debug (vec<tree, va_gc> *)): New.
21205 * tree-cfg.c
21206 (debug (struct loop &)): New.
21207 (debug (struct loop *)): New.
21208 (debug_verbose (struct loop &)): New.
21209 (debug_verbose (struct loop *)): New.
21210 * tree-dump.c: Add header dependence.
21211 * tree-flow.h
21212 (extern debug (struct loop &)): New.
21213 (extern debug (struct loop *)): New.
21214 (extern debug_verbose (struct loop &)): New.
21215 (extern debug_verbose (struct loop *)): New.
21216 * tree-data-ref.c
21217 (debug (data_reference &)): New.
21218 (debug (data_reference *)): New.
21219 (debug (vec<data_reference_p> &)): New.
21220 (debug (vec<data_reference_p> *)): New.
21221 (debug (vec<ddr_p> &)): New.
21222 (debug (vec<ddr_p> *)): New.
21223 * tree-data-ref.h
21224 (extern debug (data_reference &)): New.
21225 (extern debug (data_reference *)): New.
21226 (extern debug (vec<data_reference_p> &)): New.
21227 (extern debug (vec<data_reference_p> *)): New.
21228 (extern debug (vec<ddr_p> &)): New.
21229 (extern debug (vec<ddr_p> *)): New.
21230 * tree-ssa-alias.c
21231 (debug (pt_solution &)): New.
21232 (debug (pt_solution *)): New.
21233 * tree-ssa-alias.h
21234 (extern debug (pt_solution &)): New.
21235 (extern debug (pt_solution *)): New.
21236 * tree-ssa-alias.c
21237 (debug (_var_map &)): New.
21238 (debug (_var_map *)): New.
21239 (debug (tree_live_info_d &)): New.
21240 (debug (tree_live_info_d *)): New.
21241 * tree-ssa-alias.h
21242 (extern debug (_var_map &)): New.
21243 (extern debug (_var_map *)): New.
21244 (extern debug (tree_live_info_d &)): New.
21245 (extern debug (tree_live_info_d *)): New.
21246
be77e1e5
JH		 212472013-03-28 Jan Hubicka <jh@suse.cz>
21248
21249 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21250
777e6976
IB		 212512013-03-28 Ian Bolton <ian.bolton@arm.com>
21252
21253 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21254 record only when desired or required.
21255
9cd347ae
UB		 212562013-03-28 Uros Bizjak <ubizjak@gmail.com>
21257
21258 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21259 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21260
33159866 212612013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 21262
21263 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21264 (*andsi3_compare0_uxtw): New pattern.
21265 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21266 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21267
89ab31c1
JH		 212682013-03-28 Jan Hubicka <jh@suse.cz>
21269
21270 * data-streamer-in.c (streamer_read_gcov_count): New function.
21271 * gimple-streamer-out.c: Include value-prof.h.
21272 (output_gimple_stmt): Output histogram.
21273 (output_bb): Use streamer_write_gcov_count.
21274 * value-prof.c: Include data-streamer.h
21275 (dump_histogram_value): Add HIST_TYPE_MAX.
21276 (stream_out_histogram_value): New function.
21277 (stream_in_histogram_value): New function.
21278 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21279 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21280 * data-streamer-out.c (streamer_write_gcov_count): New function.
21281 (streamer_write_gcov_count_stream): New function.
21282 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21283 (lto_output_node): Likewise.
21284 (input_node, input_edge): Likewise.
21285 * lto-streamer-out.c (output_cfg): Update streaming.
21286 * lto-streamer-in.c (input_cfg): Likewise.
21287 * data-streamer.h (streamer_write_gcov_count,
21288 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21289 * gimple-streamer-in.c: Include value-prof.h
21290 (input_gimple_stmt): Input histograms.
21291 (input_bb): Update profile streaming.
21292
e594716a 212932013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 21294
33159866 21295 * genmodes.c (emit_max_int): New function.
8697be17 21296 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 21297 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21298 Added doc.
8697be17 21299 * machmode.def: Fixed comment.
89ab31c1 21300
e594716a 213012013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 21302
21303 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 21304 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21305
74946978
MP		 213062013-03-28 Marek Polacek <polacek@redhat.com>
21307 Richard Biener <rguenther@suse.de>
21308
21309 PR tree-optimization/56695
21310 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21311 build signed result of a vector comparison.
21312 * tree-cfg.c (verify_gimple_comparison): Check that a result
21313 of a vector comparison has signed type.
21314
a64b9c26
RB		 213152013-03-28 Richard Biener <rguenther@suse.de>
21316
21317 PR tree-optimization/37021
21318 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21319 do not restrict gaps between groups.
21320 * tree-vect-stmts.c (vectorizable_load): Properly account for
21321 a gap between groups.
21322
a9dc2a2f
EB		 213232013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21324
21325 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21326 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21327 is not enabled.
21328
53cb97f9
GP		 213292013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21330
21331 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21332 * doc/extend.texi (Named Address Spaces): Ditto.
21333 (Variable Attributes): Ditto.
21334
b802ae5c
KT		 213352013-03-27 Kai Tietz <ktietz@redhat.com>
21336
eddae10a
KT		 21337 * config.build: Add support for cygwin x64 target.
21338 * config.gcc: Likewise.
21339 * config.host: Likewise.
21340 * configure.ac: Likewise
21341 * configure: Regenerated.
b802ae5c 21342
371e77e3 213432013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 21344
21345 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21346 * config/i386/t-cygwin-w64: New file.
21347 * config/i386/cygwin-w64.h: New file.
21348 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21349 and add support for x64-cygwin target.
21350 (CPP_SPEC): Likewise.
21351 (CXX_WRAP_SPEC_LIST): Undefine before define.
21352 (LIBGCJ_SONAME): Use 15 as version.
21353
f49b33cb
RB		 213542013-03-27 Richard Biener <rguenther@suse.de>
21355
21356 PR tree-optimization/56716
21357 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21358 dumping for ref nodes.
21359
b37a6ce5
MJ		 213602013-03-27 Martin Jambor <mjambor@suse.cz>
21361
21362 PR tree-optimization/55334
21363 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21364 restricted pointers to arrays.
21365
9469b9b2
GDR		 213662013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21367
21368 * Makefile.in (.SUFFIXES): Add .cc.
21369 (.c.o): Apply same recipe for implicit rule .cc.o.
21370
7d24f650
RB		 213712013-03-27 Richard Biener <rguenther@suse.de>
21372
21373 PR tree-optimization/37021
21374 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21375 REALPART/IMAGPART_EXPRs around the supported refs.
21376 * tree-ssa-structalias.c (find_func_aliases): Assume that
21377 floating-point values are not used to transfer pointers.
21378
2f251a05
AI		 213792013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21380
f43245d1
UB		 21381 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21382 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21383 * doc/tm.texi: Regenerate.
21384 * targhooks.h (default_has_ifunc_p): New.
21385 * targhooks.c (default_has_ifunc_p): Ditto.
21386 * config/linux-protos.h: New file.
21387 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21388 hook for linux which disables support of indirect functions in android.
21389 * config/linux-android.c: New file.
21390 * config/t-linux-android.c: Ditto.
21391 * config.gcc: Added new object file linux-android.o.
21392 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21393 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21394 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 21395 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21396 doesn't support indirect functions.
21397 * configure: Regenerate.
21398
78b4e425
BC		 213992013-03-27 Bin Cheng <bin.cheng@arm.com>
21400
21401 PR target/56102
21402 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21403 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21404 mult-word mode.
21405
a5ba7b92
AK		 214062013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21407
21408 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21409
be624986
TG		 214102013-03-27 Terry Guo <terry.guo@arm.com>
21411
21412 * config/arm/arm-cores.def: Added core cortex-r7.
21413 * config/arm/arm-tune.md: Regenerated.
21414 * config/arm/arm-tables.opt: Regenerated.
21415 * doc/invoke.texi: Added entry for core cortex-r7.
21416
0a514f47
WL		 214172013-03-27 Walter Lee <walt@tilera.com>
21418
21419 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21420 double-decrement of next_scratch_regno.
21421
9332b0d2
WL		 214222013-03-27 Walter Lee <walt@tilera.com>
21423
801d9b2a 21424 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 21425 input operands.
21426 (insn_v1mulus): Ditto.
21427 (insn_v2muls): Ditto.
21428
f54ea5dd
WL		 214292013-03-27 Walter Lee <walt@tilera.com>
21430
e594716a 21431 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 21432 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21433
bf60f4ca
WL		 214342013-03-27 Walter Lee <walt@tilera.com>
21435
e594716a 21436 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 21437 (*sibcall_value): Ditto.
21438
e3b51eeb
WL		 214392013-03-27 Walter Lee <walt@tilera.com>
21440
21441 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21442 (insn_mnz_v8qi): ... this ...
21443 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21444 vector equivalent.
21445 (insn_v<n>mnz): Replaced by ...
21446 (insn_v1mnz): ... this ...
21447 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21448 equivalent.
21449 (insn_mz_<mode>): Replaced by ...
21450 (insn_mz_v8qi): ... this ...
21451 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21452 vector equivalent.
21453 (insn_v<n>mz): Replaced by ...
21454 (insn_v1mz): ... this ...
21455 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21456 equivalent.
21457
4fc7b145
EB		 214582013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21459
21460 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21461
6f33817e
RM		 214622013-03-26 Roland McGrath <mcgrathr@google.com>
21463
21464 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21465 than fprintf with a non-constant, non-format string.
21466
93a4145b
UB		 214672013-03-26 Uros Bizjak <ubizjak@gmail.com>
21468
21469 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21470 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21471 operand 0 predicate.
21472 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21473 attribute. Use general_x64nomem_operand as operand 1 predicate.
21474 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21475 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21476 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21477 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21478 (mov<mode>_insv_1): Remove expander. Merge insn with
21479 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21480 Use general_x64nomem_operand as operand 1 predicate.
21481 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21482 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21483 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21484 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21485 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21486 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21487 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21488 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21489 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21490 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21491 (general_x64nomem_operand): Ditto.
21492
c6a9ed5a 214932013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 21494
21495 * config/rtems.opt: Add -pthread option.
21496
31b0a960
RB		 214972013-03-26 Richard Biener <rguenther@suse.de>
21498
93a4145b
UB		 21499 * alias.c (find_base_term): Avoid redundant and not used recursion.
21500 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 21501 (true_dependence_1): Compute and pass base terms to base_alias_check.
21502 (write_dependence_p): Likewise.
21503 (may_alias_p): Likewise.
21504
79517551
SN		 215052013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21506
21507 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21508 PC-relative load in SI modes and above only.
21509
a76213b9
XQ		 215102013-03-26 Xinyu Qi <xyqi@marvell.com>
21511
21512 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 21513 * config/arm/iwmmxt.md (WCGR0): Update.
21514 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21515
37ff9355
UB		 215162013-03-26 Uros Bizjak <ubizjak@gmail.com>
21517
21518 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21519 Use x64 and nox64 isa attributes.
21520
30b0317c
RB		 215212013-03-26 Richard Biener <rguenther@suse.de>
21522
21523 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21524 alignment computations and rely on get_object_alignment_1
21525 for the !TYPE_P case.
21526 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21527
cbcdb140
WL		 215282013-03-26 Walter Lee <walt@tilera.com>
21529
21530 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21531 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21532
dd884525 215332013-03-25 Jeff Law <law@redhat.com>
be672e08 21534
33159866
UB		 21535 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21536 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 21537 wrong version of prior patch.
21538
0051d3ec
WL		 215392013-03-25 Walter Lee <walt@tilera.com>
21540
21541 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21542 TILEGX_INSN_SHUFFLEBYTES1.
21543 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21544 shufflebytes1.
21545 (tilegx_builtins): Ditto.
21546 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21547
065a3605
WL		 215482013-03-25 Walter Lee <walt@tilera.com>
21549
21550 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 21551 (floatunssisf2): New pattern.
21552 (floatsidf2): New pattern.
21553 (floatunssidf2): New pattern.
065a3605 21554
5b2a3c82
WL		 215552013-03-25 Walter Lee <walt@tilera.com>
21556
21557 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21558 tests for constraint J, K, N, P.
21559
192ea533
WL		 215602013-03-25 Walter Lee <walt@tilera.com>
21561
21562 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21563 Use indirect/pcrel encoding.
21564 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21565 Ditto.
21566
855e0d0b
SE		 215672013-03-25 Steve Ellcey <sellcey@mips.com>
21568
21569 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21570 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21571 * config/mips/mips.c (mips_option_override): Set IMADD default.
21572 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21573 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21574 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21575 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21576 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21577
39e45653
JL		 215782013-03-25 Jeff Law <law@redhat.com>
21579
21580 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21581 slightly to avoid creating and folding useless trees. Simplify
21582 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21583
7482c470
UB		 215842013-03-25 Uros Bizjak <ubizjak@gmail.com>
21585
21586 * config/i386/i386.md (*zero_extendsidi2): Merge with
21587 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21588 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21589 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21590 targets. Clarify comment.
21591
4a53743e
MJ		 215922013-03-25 Martin Jambor <mjambor@suse.cz>
21593
21594 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21595 pass-through jump functions differently.
21596 (ipa_read_jump_function): Likewise. Also use setter functions to set
21597 up jump functions.
21598
162712de
MJ		 215992013-03-25 Martin Jambor <mjambor@suse.cz>
21600
21601 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21602 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21603 process it.
21604 (ipa_get_indirect_edge_target): New function.
21605 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21606 ipa_get_indirect_edge_target. Update all callers.
21607 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21608 ipa_get_indirect_edge_target_1 instead of calling
21609 ipa_get_indirect_edge_target.
21610 (create_specialized_node): Pass aggvlas to
21611 ipcp_discover_new_direct_edges.
21612
a5a4c20a
KT		 216132013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21614
21615 * config/arm/arm.md (f_sels, f_seld): New types.
21616 (*cmov<mode>): New pattern.
21617 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21618 predicate.
21619
f992b9fc
KT		 216202013-03-25 Kai Tietz <ktietz@redhat.com>
21621
21622 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21623 POSIX-printf for mingw-hosted builds.
21624
c16fd676
RB		 216252013-03-25 Richard Biener <rguenther@suse.de>
21626
21627 PR middle-end/56694
21628 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21629 must-not-throw stmt location.
21630
5ad29f12
KT		 216312013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21632
21633 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21634 Emit load-acquire versions when acq is true.
21635 (arm_emit_store_exclusive): Add rel parameter.
21636 Emit store-release versions when rel is true.
21637 (arm_split_compare_and_swap): Use acquire-release instructions
21638 instead.
21639 of barriers when appropriate.
21640 (arm_split_atomic_op): Likewise.
21641 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21642 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21643 (VUNSPEC_SLX): Likewise.
21644 (VUNSPEC_LDA): Likewise.
21645 (VUNSPEC_STL): Likewise.
21646 * config/arm/sync.md (atomic_load<mode>): New pattern.
21647 (atomic_store<mode>): Likewise.
21648 (arm_load_acquire_exclusive<mode>): Likewise.
21649 (arm_load_acquire_exclusivesi): Likewise.
21650 (arm_load_acquire_exclusivedi): Likewise.
21651 (arm_store_release_exclusive<mode>): Likewise.
21652
03a7dddb
CM		 216532013-03-25 Catherine Moore <clm@codesourcery.com>
21654
21655 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21656 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21657 * config/mip/predicates.md (lwsp_swsp_operand,
21658 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21659 sb16_operand, db4_operand, db7_operand, ib3_operand,
21660 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21661 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21662 andi16_operand): New predicates.
21663 * config/mips/mips.md (compression): New attribute.
21664 (enabled): New attribute.
21665 (length): Consider compression in computing length.
21666 (shift_compression): New code attribute.
21667 (*add<mode>3): New operands. Record compression.
21668 (sub<mode>3): Likewise.
21669 (one_cmpl<mode>2): Likewise.
21670 (*and<mode>3): Likewise.
21671 (*ior<mode>3): Likewise.
21672 (unnamed pattern for xor): Likewise.
21673 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21674 (*<optab><mode>3): Likewise.
21675 (*mov<mode>_internal: Likewise.
21676 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21677 (mips_unsigned_immediate_p): New.
21678 (umips_lwsp_swsp_address_p): New.
21679 (m16_based_address_p): New.
21680 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21681 (mips_unsigned_immediate_p): New prototype.
21682 (lwsp_swsp_address_p): New prototype.
21683 (m16_based_address_p): New prototype.
21684 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21685 (mips_signed_immediate_p): New function.
21686 (m16_based_address_p): New function.
21687 (lwsp_swsp_address_p): New function.
21688 (mips_print_operand_punctuation): Recognize short delay slot insns
21689 for microMIPS.add<mode>3"
21690
f35c297f 216912013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 21692
f35c297f
KT		 21693 PR target/56720
21694 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21695 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21696
051b9446
RB		 216972013-03-25 Richard Biener <rguenther@suse.de>
21698
21699 PR tree-optimization/56689
21700 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21701 any edge.
21702
374001cb
RB		 217032013-03-25 Richard Biener <rguenther@suse.de>
21704
21705 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21706 of bitmap.
21707 (memory_references): Likewise.
21708 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21709 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21710 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21711 (gather_mem_refs_in_loops): Fold into ...
21712 (analyze_memory_references): ... this. Move initialization
21713 to tree_ssa_lim_initialize.
21714 (fill_always_executed_in): Rename to ...
21715 (fill_always_executed_in_1): ... this.
21716 (fill_always_executed_in): Move contains_call computation to
21717 this new function from ...
21718 (tree_ssa_lim_initialize): ... here.
21719 (tree_ssa_lim): Call fill_always_executed_in.
21720
57895947
EB		 217212013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21722
21723 * postreload.c (reload_combine): Fix code detecting returns.
21724
1e1b18c1
EB		 217252013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21726
21727 * function.c (emit_use_return_register_into_block): On cc0 targets,
21728 do not emit the sequence between cc0 setter and user.
21729
9216baf1
KT		 217302013-03-25 Kai Tietz <ktietz@redhat.com>
21731
1e1b18c1
EB		 21732 * config/i386/predicates.md (local_symbolic_operand): Interpret
21733 dll-imported symbols as none-local.
9216baf1 21734
76421b44
RB		 217352013-03-25 Richard Biener <rguenther@suse.de>
21736
21737 * tree-ssa-loop-im.c (struct depend): Remove.
21738 (struct lim_aux_data): Make depends a vec of gimples.
21739 (free_lim_aux_data): Adjust.
21740 (add_dependency): Likewise.
21741 (set_level): Likewise.
21742
d154bfa2
RB		 217432013-03-25 Richard Biener <rguenther@suse.de>
21744
21745 PR middle-end/56434
21746 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21747 the pointer returned by calls with ECF_MALLOC set.
21748
2cd9804e
UB		 217492013-03-24 Uros Bizjak <ubizjak@gmail.com>
21750
acef5fe0
UB		 21751 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21752
217532013-03-24 Uros Bizjak <ubizjak@gmail.com>
21754
21755 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 21756 using MMXMODE mode iterator.
21757 (*move<mode>_internal): Merge with *movv2sf_internal and
21758 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21759
225ccc68
SB		 217602013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21761
85c0f02d
SB		 21762 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21763 (record_last_mem_set_info): Likewise.
21764
225ccc68
SB		 21765 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21766 of XNEWVEC followed by memset.
21767 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21768
b64925dc
SB		 217692013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21770
21771 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21772 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21773 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21774 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21775 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21776 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21777 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21778 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21779 BARRIER_P instead of GET_CODE.
21780
b9aaf52e
EB		 217812013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21782
21783 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21784 inaccuracy in the probing code.
21785
21786 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21787 (ctrapdi4): Likewise.
21788
66de4d7c
EB		 217892013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21790
21791 * calls.c (expand_call): Add missing guard to code handling return
21792 of non-BLKmode structures in MSB.
21793 * function.c (expand_function_end): Likewise.
21794
4f25c8fb
EB		 217952013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21796
21797 * combine.c (try_combine): Adjust comment. Do not add the set of
21798 insn #0 if the destination indirectly is set or dies in insn #2.
21799 Tidy up code to distribute a new note.
21800
23b7850d
UB		 218012013-03-22 Uros Bizjak <ubizjak@gmail.com>
21802
21803 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21804 also for alternatives 16 and 17.
21805
bed852cf
UB		 218062013-03-22 Uros Bizjak <ubizjak@gmail.com>
21807
21808 * config/i386/sse.md (*mov<mode>_internal): Merge with
21809 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21810 Emit insn template depending on type attribute. Use
21811 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21812 movd instead of movq mnemonic for interunit moves. Rewrite mode
21813 attribute calculation. Remove unit attribute calculation.
21814 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21815 Set prefix_data16 attribute for DImode ssemov types.
21816 Use Ym instead of y for SSE-MMX conversion alternatives.
21817 Reorder operand constraints.
21818
1e8a7937
SB		 218192013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21820
21821 * df.h (df_insn_delete): Adjust prototype.
21822 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21823 and let it decide whether mark the basic block dirty.
21824 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21825 * df-scan.c (df_insn_info_delete): New helper function, split
21826 off from df_insn_delete.
21827 (df_scan_free_bb_info): Use it.
21828 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21829 Likewise.
21830 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21831 that the insn is actually an insn and it has a non-NULL basic block.
21832 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21833
e14d094c
RB		 218342013-03-22 Richard Biener <rguenther@suse.de>
21835
21836 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21837 dep_ref members.
21838 (mem_ref_alloc): Do not allocate them.
21839 (refs_independent_p): Do not query or maintain a cache.
21840
e9cf7316
RB		 218412013-03-22 Richard Biener <rguenther@suse.de>
21842
21843 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21844 (gather_mem_refs_in_loops): Do not compute it.
21845 (analyze_memory_references): Do not allocate it.
21846 (tree_ssa_lim_finalize): Do not free it.
21847 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21848
f046e81b
RB		 218492013-03-22 Richard Biener <rguenther@suse.de>
21850
21851 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 21852
a05c0ddf
IB		 218532013-03-22 Ian Bolton <ian.bolton@arm.com>
21854
21855 * config/aarch64/aarch64.c (aarch64_print_operand): New
21856 format specifier for printing a constant in hex.
21857 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21858 format specifier for printing second operand.
21859
c00217fc
RB		 218602013-03-22 Richard Biener <rguenther@suse.de>
21861
21862 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21863 bitmaps.
21864 (gather_mem_refs_in_loops): Perform store accumulation here.
21865 (create_vop_ref_mapping_loop): Remove.
21866 (create_vop_ref_mapping): Likewise.
21867 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 21868 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 21869 (record_indep_loop): Remove.
21870 (record_dep_loop): New function.
21871 (ref_indep_loop_p_1): Adjust to only walk over references
21872 in the loop, not its subloops.
21873 (ref_indep_loop_p): Rename to ...
21874 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21875 maintaining a more fine-grained cache.
21876 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21877 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21878
15d19bf8
RB		 218792013-03-22 Richard Biener <rguenther@suse.de>
21880
21881 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21882 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21883 aggregate mem_ref_loc.
21884 (free_mem_ref_locs): Inline into ...
21885 (memref_free): ... this and adjust.
21886 (mem_ref_alloc): Adjust.
21887 (mem_ref_locs_alloc): Remove.
21888 (record_mem_ref_loc): Adjust.
21889 (get_all_locs_in_loop): Rewrite into ...
21890 (for_all_locs_in_loop): ... this iterator.
21891 (rewrite_mem_ref_loc): New functor.
21892 (rewrite_mem_refs): Use for_all_locs_in_loop.
21893 (sm_set_flag_if_changed): New functor.
21894 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21895 (ref_always_accessed): New functor.
21896 (ref_always_accessed_p): Use for_all_locs_in_loop.
21897
6f37411d
MG		 218982013-03-21 Marc Glisse <marc.glisse@inria.fr>
21899
21900 * tree-pass.h (PROP_gimple_lvec): New.
21901 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21902 (init_optimization_passes): Move pass_lower_vector.
21903 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21904 PROP_gimple_lvec.
21905 (pass_lower_vector): Provide PROP_gimple_lvec.
21906 (pass_lower_vector_ssa): Likewise.
21907 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21908
511ef689
MW		 219092013-03-21 Mark Wielaard <mjw@redhat.com>
21910
21911 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21912
cc1df30b
UB		 219132013-03-21 Uros Bizjak <ubizjak@gmail.com>
21914
21915 * config/i386/i386.md (*movdi_internal): Disparage slightly
21916 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21917 conversion alternatives.
21918
c845cfe1
JJ		 219192013-03-21 Jakub Jelinek <jakub@redhat.com>
21920
37e99116
JJ		 21921 PR middle-end/48087
21922 * diagnostic.def (DK_WERROR): New kind.
21923 * diagnostic.h (werrorcount): Define.
21924 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21925 promoted to DK_ERROR, increment DK_WERROR counter instead of
21926 DK_ERROR counter.
21927 * toplev.c (toplev_main): Call print_ignored_options even if
21928 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21929 even if just werrorcount is non-zero.
21930
c845cfe1
JJ		 21931 PR debug/55608
21932 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21933 on failure.
21934 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21935 (string_cst_pool_decl): New function.
21936 (optimize_one_addr_into_implicit_ptr): New function.
21937 (resolve_addr_in_expr): Optimize DWARF location expression
21938 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21939 which doesn't live in memory, but has DW_AT_location or
21940 DW_AT_const_value, or refers to a string literal, into
21941 DW_OP_GNU_implicit_pointer.
21942 (optimize_location_into_implicit_ptr): New function.
21943 (resolve_addr): If removing DW_AT_location of a variable because
21944 it was DW_OP_addr of address of the variable, but the variable doesn't
21945 live in memory, try to emit const value attribute for the initializer.
21946
08e0cda6
MG		 219472013-03-21 Marc Glisse <marc.glisse@inria.fr>
21948
21949 * tree.h (VECTOR_TYPE_P): New macro.
21950 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21951 TYPE_MODE): Use it.
21952 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21953 VEC_COND_EXPR cannot be lvalues.
21954 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21955
d08633b4
MG		 219562013-03-21 Marc Glisse <marc.glisse@inria.fr>
21957
21958 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21959 Restrict the transformation to equal modes.
21960
e6647190
RB		 219612013-03-21 Richard Biener <rguenther@suse.de>
21962
21963 PR tree-optimization/39326
21964 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21965 (MEM_ANALYZABLE): Adjust.
21966 (record_mem_ref_loc): Move bitmap ops ...
21967 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21968 unanalyzable refs, do not record locations for it.
21969 (analyze_memory_references): Allocate ref zero as shared
21970 unanalyzable ref.
21971 (refs_independent_p): Do not test for unanalyzed mems here.
21972 (ref_indep_loop_p_1): Special-case disambiguation against
21973 the unanalyzed ref.
cc1df30b 21974 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 21975
65074f54
CL		 219762013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21977
21978 * config/arm/arm-protos.h (tune_params): Add
21979 prefer_neon_for_64bits field.
21980 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21981 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21982 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21983 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21984 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21985 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21986 (arm_option_override): Handle -mneon-for-64bits new option.
21987 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21988 (prefer_neon_for_64bits): Declare new variable.
21989 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 21990 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 21991 (arch_enabled): Handle new arch types. Remove support for onlya8
21992 and nota8.
21993 (one_cmpldi2): Use new arch names.
bc5faa5b 21994 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 21995 * config/arm/arm.opt (mneon-for-64bits): Add option.
21996 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21997 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21998 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21999 of onlya8.
22000 * doc/invoke.texi (-mneon-for-64bits): Document.
22001
5a2d2a79
RB		 220022013-03-21 Richard Biener <rguenther@suse.de>
22003
22004 PR tree-optimization/39326
22005 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22006 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 22007 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 22008
5abe1e05
RB		 220092013-03-21 Richard Biener <rguenther@suse.de>
22010
22011 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22012 (vect_insert_into_interleaving_chain): Likewise.
22013 (vect_drs_dependent_in_basic_block): Inline ...
22014 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22015 split out from ...
22016 (vect_analyze_data_ref_dependence): ... here. Simplify.
22017 (vect_check_interleaving): Simplify.
22018 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22019 (vect_slp_analyze_data_ref_dependences): ... this new function.
22020 (dr_group_sort_cmp): New function.
22021 (vect_analyze_data_ref_accesses): Compute data-reference groups
22022 here instead of in vect_analyze_data_ref_dependence. Use
22023 a more efficient algorithm.
22024 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22025 vect_slp_analyze_data_ref_dependences. Call
22026 vect_analyze_data_ref_accesses earlier.
22027 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22028 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22029 (vect_slp_analyze_data_ref_dependences): New prototype.
22030
cad1735b
RB		 220312013-03-21 Richard Biener <rguenther@suse.de>
22032
22033 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22034 ref is stored in the loop.
22035 (find_refs_for_sm): Walk only over all stores.
22036 (store_motion_loop): Allocate from lim_bitmap_obstack.
22037 (store_motion): Likewise.
22038
141310ef
RB		 220392013-03-21 Richard Biener <rguenther@suse.de>
22040
22041 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22042 Update virtual SSA form.
22043
5022315a
RO		 220442013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22045
22046 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22047 * configure: Regenerate.
22048 * config.in: Regenerate.
22049 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22050 if !HAVE_LD_EH_FRAME_CIEV3.
22051
50d4421c
RB		 220522013-03-21 Richard Biener <rguenther@suse.de>
22053
22054 * tree-cfg.c (verify_expr_no_block): New function.
22055 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22056 nor DECL_VALUE_EXPR have locations with associated blocks.
22057 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22058 (clear_unused_block_pointer): Remove code dealing with
22059 blocks in DECL_DEBUG_EXPR locations.
22060
839b422f
RB		 220612013-03-21 Richard Biener <rguenther@suse.de>
22062
22063 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22064 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22065 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22066 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22067 instead of DECL_DEBUG_EXPR_IS_FROM.
22068 * gimplify.c (gimplify_modify_expr): Likewise.
22069 * tree-cfg.c (verify_expr_location_1): Likewise.
22070 * tree-complex.c (create_one_component_var): Likewise.
22071 * tree-sra.c (create_access_replacement): Likewise.
22072 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22073 (clear_unused_block_pointer): Likewise.
22074 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22075 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22076 * var-tracking.c (var_debug_decl): Likewise.
22077 (track_expr_p): Likewise.
22078 * tree-inline.c (add_local_variables): Likewise. Set
22079 DECL_HAS_DEBUG_EXPR_P after copying it.
22080 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22081 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22082
fe04878d
UB		 220832013-03-21 Uros Bizjak <ubizjak@gmail.com>
22084
22085 PR bootstrap/56656
22086 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22087 * configure: Regenerate.
22088 * config.in: Regenerate.
22089 * config/i386/i386.md (*movdf_internal): Use
22090 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22091 movd instead of movq mnemonic for interunit moves.
22092 (*movdi_internal): Ditto.
22093
bd059b26 220942013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 22095
22096 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22097 (abd<mode>_3): New pattern.
22098 (aba<mode>_3): New pattern.
22099 (fabd<mode>_3): New pattern.
22100
bd059b26 221012013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 22102
22103 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22104 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22105 occurrence of REGISTER_PREFIX as its empty string.
22106
4f1f78b9
JL		 221072013-03-20 Jeff Law <law@redhat.com>
22108
22109 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22110 addititional equivalences for equality comparisons between an SSA_NAME
22111 and a constant where the SSA_NAME was set from a widening conversion.
22112
327a1118
WL		 221132013-03-20 Walter Lee <walt@tilera.com>
22114
22115 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22116
813e0036
UB		 221172013-03-20 Uros Bizjak <ubizjak@gmail.com>
22118
22119 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22120 depending on type attribute.
22121 (*movti_internal): Ditto.
22122 (*movtf_internal): Ditto.
22123 (*movxf_internal): Ditto.
22124 (*movdf_internal): Ditto.
22125 (*movsf_internal): Ditto.
22126
7cf34aae
UB		 221272013-03-20 Uros Bizjak <ubizjak@gmail.com>
22128
22129 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22130 maybe_vex for sselog1 and ssemov types.
22131 (*movdi_internal): Reorder operand constraints.
22132 (*movsi_internal): Ditto. Set prefix attribute to
22133 maybe_vex for sselog1 and ssemov types.
22134 (*movtf_internal): Set prefix attribute to maybe_vex
22135 for sselog1 and ssemov types.
22136 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22137 DImode ssemov types. Reorder operand constraints.
22138 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22139 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22140 attribute for SImode ssemov types. Reorder operand constraints.
22141
19321415
MJ		 221422013-03-20 Martin Jambor <mjambor@suse.cz>
22143
22144 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22145 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22146
12211b99 221472013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 22148
22149 * config/rs6000/predicates.md (indexed_address, update_address_mem
22150 update_indexed_address_mem): New predicates.
22151 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22152 attribute for load/store instructions.
22153 * config/rs6000/dfp.md (movsd_store): Likewise.
22154 (movsd_load): Likewise.
22155 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22156 (unnamed HI->DI extend define_insn): Likewise.
22157 (unnamed SI->DI extend define_insn): Likewise.
22158 (unnamed QI->SI extend define_insn): Likewise.
22159 (unnamed QI->HI extend define_insn): Likewise.
22160 (unnamed HI->SI extend define_insn): Likewise.
22161 (unnamed HI->SI extend define_insn): Likewise.
22162 (extendsfdf2_fpr): Likewise.
22163 (movsi_internal1): Likewise.
22164 (movsi_internal1_single): Likewise.
22165 (movhi_internal): Likewise.
22166 (movqi_internal): Likewise.
22167 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22168 attribute for load/store instructions.
22169 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22170 instructions.
22171 (mov<mode>_softfloat): Likewise.
22172 (mov<mode>_hardfloat32): Likewise.
22173 (mov<mode>_hardfloat64): Likewise.
22174 (mov<mode>_softfloat64): Likewise.
22175 (movdi_internal32): Likewise.
22176 (movdi_internal64): Likewise.
22177 (probe_stack_<mode>): Likewise.
22178
221792013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22180
22181 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22182 floating point, and decimal floating point to reload iterator.
22183
22184 * config/rs6000/constraints.md (wl constraint): New constraints to
22185 return FLOAT_REGS if certain options are used to reduce the number
22186 of separate patterns that exist in the file.
22187 (wx constraint): Likewise.
22188 (wz constraint): Likewise.
22189
22190 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22191 -mdebug=reg, print wg, wl, wx, and wz constraints.
22192 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22193 Initialize the reload functions for 64-bit binary/decimal floating
22194 point types.
22195 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22196 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22197 create the buffer on the stack to overcome not having a 32-bit
22198 load and store.
22199 (rs6000_emit_move): Likewise.
22200 (rs6000_secondary_memory_needed_rtx): Likewise.
22201 (rs6000_alloc_sdmode_stack_slot): Likewise.
22202 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22203 via xxlxor, just like DFmode 0.0.
22204
22205 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22206 define as 1 if we are running on a power7 or newer.
22207 (enum r6000_reg_class_enum): Add new constraints.
22208
22209 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22210 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22211 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 22212 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 22213 (movsd splitter): Likewise.
22214 (movsd_hardfloat): Likewise.
22215 (movsd_softfloat): Likewise.
22216
22217 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22218 binary and decimal floating point moves.
22219 (fmove_ok): New attributes to combine binary and decimal floating
22220 point moves, and to combine power6x (mfpgpr) moves along normal
22221 floating moves.
22222 (real_value_to_target): Likewise.
22223 (f32_lr): Likewise.
22224 (f32_lm): Likewise.
22225 (f32_li): Likewise.
22226 (f32_sr): Likewise.
22227 (f32_sm): Likewise.
22228 (f32_si): Likewise.
22229 (movsf): Combine binary and decimal floating point moves. Combine
22230 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 22231 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 22232 (mov<mode> for SFmode/SDmode); Likewise.
22233 (SFmode/SDmode splitters): Likewise.
22234 (movsf_hardfloat): Likewise.
22235 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22236 (movsf_softfloat): Likewise.
22237 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22238
22239 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22240 wx and wz constraints.
22241
22242 * config/rs6000/constraints.md (wg constraint): New constraint to
22243 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22244
22245 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22246 constraint.
22247
22248 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22249 -mdebug=reg, print wg, wl, wx, and wz constraints.
22250 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22251 Initialize the reload functions for 64-bit binary/decimal floating
22252 point types.
22253 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22254 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22255 create the buffer on the stack to overcome not having a 32-bit
22256 load and store.
22257 (rs6000_emit_move): Likewise.
22258 (rs6000_secondary_memory_needed_rtx): Likewise.
22259 (rs6000_alloc_sdmode_stack_slot): Likewise.
22260 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22261 via xxlxor, just like DFmode 0.0.
22262
c6d5ff83
MM		 22263 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22264 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22265 with other moves by using conditional constraits (wg). Use LFIWZX
22266 and STFIWX for loading SDmode on power7.
22267 (movdd splitters): Likewise.
22268 (movdd_hardfloat32): Likewise.
22269 (movdd_softfloat32): Likewise.
22270 (movdd_hardfloat64_mfpgpr): Likewise.
22271 (movdd_hardfloat64): Likewise.
22272 (movdd_softfloat64): Likewise.
22273
22274 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22275 64-bit binary and decimal floating point moves.
22276 (FMOVE64X): Likewise.
22277 (movdf): Combine 64-bit binary and decimal floating point moves.
22278 Combine power6x (mfpgpr) moves with other moves by using
22279 conditional constraits (wg).
22280 (mov<mode> for DFmode/DDmode): Likewise.
22281 (DFmode/DDmode splitters): Likewise.
22282 (movdf_hardfloat32): Likewise.
22283 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22284 (movdf_softfloat32): Likewise.
22285 (movdf_hardfloat64_mfpgpr): Likewise.
22286 (movdf_hardfloat64): Likewise.
22287 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22288 (movdf_softfloat64): Likewise.
22289 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22290 (reload_<mode>_load): Move to later in the file so they aren't in
22291 the middle of the floating point move insns.
22292 (reload_<mode>_store): Likewise.
22293
22294 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22295 constraint.
22296
22297 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22298 constraint if -mdebug=reg.
bd059b26
UB		 22299 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22300 Enable using dd reload support if needed.
c6d5ff83
MM		 22301
22302 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22303 binary and decimal floating point moves in rs6000.md.
22304 (movtd_internal): Likewise.
22305
22306 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22307 decimal floating point moves.
22308 (movtf): Likewise.
22309 (movtf_internal): Likewise.
22310 (mov<mode>_internal, TDmode/TFmode): Likewise.
22311 (movtf_softfloat): Likewise.
22312 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22313
22314 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22315 movdi_internal64, using wg constraint for move direct operations.
22316 (movdi_internal64): Likewise.
22317
22318 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22319 MODES_TIEABLE_P for selected modes. Print the numerical value of
22320 the various virtual registers. Use GPR/FPR first/last values,
22321 instead of hard coding the register numbers. Print which modes
22322 have reload functions registered.
bd059b26
UB		 22323 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22324 settings before/after setting cpu, target and subtarget settings.
22325 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22326 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 22327 (rs6000_secondary_reload_fail): Likewise.
22328 (rs6000_secondary_reload_inner): Likewise.
22329
22330 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22331 macros for first/last GPR and FPR registers.
22332 (LAST_GPR_REGNO): Likewise.
22333 (FIRST_FPR_REGNO): Likewise.
22334 (LAST_FPR_REGNO): Likewise.
22335
22336 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22337 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22338 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22339 (vcond<mode><mode>): Likewise.
22340 (vcondu<mode><mode>): Likewise.
22341 (vector_gtu<mode>): Likewise.
22342 (vector_gte<mode>): Likewise.
22343 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22344 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 22345 (ior<mode>3): Likewise.
22346 (and<mode>3): Likewise.
22347 (one_cmpl<mode>2): Likewise.
22348 (nor<mode>3): Likewise.
22349 (andc<mode>3): Likewise.
22350
22351 * config/rs6000/constraints.md (wt constraint): New constraint
22352 that returns VSX_REGS if TImode is allowed in VSX registers.
22353
22354 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22355 constant under VSX.
22356
22357 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22358 similar to TImode, but it is restricted to being in the GPRs.
22359
22360 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22361 TImode to occupy a single VSX register.
22362
22363 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22364 -mvsx-timode for power7/power8.
22365 (power7 cpu): Likewise.
22366 (power8 cpu): Likewise.
22367
22368 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22369 sure that TFmode/TDmode take up two registers if they are ever
22370 allowed in the upper VSX registers.
22371 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22372 registers.
22373 (rs6000_init_hard_regno_mode_ok): Likewise.
22374 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22375 constraint. Print if LRA is turned on.
22376 (rs6000_option_override_internal): Give an error if -mvsx-timode
22377 and VSX is not enabled.
22378 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22379 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22380 to reg+offset addressing. Use PTImode when checking offset
22381 addresses for validity.
22382 (reg_offset_addressing_ok_p): Likewise.
22383 (rs6000_legitimate_offset_address_p): Likewise.
22384 (rs6000_legitimize_address): Likewise.
22385 (rs6000_legitimize_reload_address): Likewise.
22386 (rs6000_legitimate_address_p): Likewise.
22387 (rs6000_eliminate_indexed_memrefs): Likewise.
22388 (rs6000_emit_move): Likewise.
22389 (rs6000_secondary_reload): Likewise.
22390 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22391 reloads to fpr registers to continue to use reg+offset addressing,
22392 but 64-bit reloads to altivec registers need reg+reg addressing.
22393 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22394 it. Treat LO_SUM like a PLUS operation.
22395 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22396 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 22397 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22398 registers to share a register with a smaller sized type, since VSX
22399 puts scalars in the upper 64-bits.
22400 (print_operand): Add support for PTImode.
22401 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22402 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22403 registers, but don't have arithmetic support.
22404 (rs6000_memory_move_cost): Add test for VSX.
22405 (rs6000_opt_masks): Add -mvsx-timode.
22406
22407 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22408 for TImode.
22409 (VSs): Likewise.
22410 (VSr): Use wt constraint for TImode.
22411 (VSv): Drop TImode support.
22412 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22413 (vsx_movti_64bit): Likewise.
22414 (vsx_movti_32bit): Likewise.
22415 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22416 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22417 one '?' on the appropriate output constraint. Do not allow TImode
22418 logical operations on 32-bit systems.
22419 (vsx_ior<mode>3): Likewise.
22420 (vsx_xor<mode>3): Likewise.
22421 (vsx_one_cmpl<mode>2): Likewise.
22422 (vsx_nor<mode>3): Likewise.
22423 (vsx_andc<mode>3): Likewise.
22424 (vsx_concat_<mode>): Likewise.
22425 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22426
22427 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22428 OPTION_MASK_VSX_TIMODE.
22429 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22430 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22431
22432 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22433 (TI2 iterator): New iterator for TImode, PTImode.
22434 (wd mode attribute): Add values for vector types.
bd059b26
UB		 22435 (movti_string): Replace TI move operations with operations for TImode
22436 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 22437 (mov<mode>_string, TImode/PTImode): Likewise.
22438 (movti_ppc64): Likewise.
22439 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22440 (TI mode splitters): Likewise.
22441
22442 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22443 constraint.
22444
1fc5eced
MG		 224452013-03-20 Marc Glisse <marc.glisse@inria.fr>
22446
22447 PR tree-optimization/56355
22448 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22449 Also handle integers with undefined overflow.
22450
22c4c869
CM		 224512013-03-20 Catherine Moore <clm@codesourcery.com>
22452 Maciej W. Rozycki <macro@codesourcery.com>
22453 Tom de Vries <tom@codesourcery.com>
12211b99 22454 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 22455 Iain Sandoe <iain@codesourcery.com>
22456 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22457 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 22458
22459 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22460 Document new function attributes.
22c4c869
CM		 22461 * doc/invoke.texi (minterlink-compressed, mmicromips,
22462 m14k, m14ke, m14kec): Document new options.
22463 (minterlink-mips16): Update documentation.
22464 * doc/md.texi (ZC, ZD): Document new constraints.
22465 * configure.ac (gcc_cv_as_micromips): Check if linker
22466 supports the .set micromips directive.
22467 * configure: Regenerate.
22468 * config.in: Regenerate.
22469 * config/mips/mips-tables.opt: Regenerate.
22470 * config/mips/micromips.md: New file.
22471 * constraints.md (ZC, ZD): New constraints.
22472 * config/mips/predicates.md (movep_src_register): New predicate.
22473 (movep_src_operand): New predicate.
22474 (non_volatile_mem_operand): New predicate.
22475 * config/mips/mips.md (multimem): New type.
22476 (length): Differentiate between 17-bit and 18-bit branch offsets.
22477 (MOVEP1, MOVEP2): New mode iterator.
33159866 22478 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 22479 (mov_<load>r): Likewise.
22480 (mov_<store>l): Likewise.
22481 (mov_<store>r): Likewise.
22482 (*branch_equality<mode>_inverted): Add microMIPS support.
22483 (*branch_equality<mode>): Likewise.
22484 (*jump_absolute): Likewise.
22485 (indirect_jump_<mode>): Likewise.
22486 (tablejump_<mode>): Likewise.
22487 (<optab>_internal): Likewise.
22488 (sibcall_internal): Likewise.
22489 (sibcall_value_internal): Likewise.
22490 (prefetch): Use constraint ZD.
22491 * config/mips/mips.opt (minterlink-compressed): New option.
22492 (minterlink-mips16): Now an alias for minterlink-compressed.
22493 (mmicromips): New option.
22494 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22495 (compare_and_swap_12): Likewise.
22496 (sync_add<mode>): Likewise.
22497 (sync_<optab>_12): Likewise.
22498 (sync_old_<optab>_12): Likewise.
22499 (sync_new_<optab>_12): Likewise.
22500 (sync_nand_12): Likewise.
22501 (sync_old_nand_12): Likewise.
22502 (sync_new_nand_12): Likewise.
22503 (sync_sub<mode>): Likewise.
22504 (sync_old_add<mode>): Likewise.
22505 (sync_old_sub<mode>): Likewise.
22506 (sync_new_add<mode>): Likewise.
22507 (sync_new_sub<mode>): Likewise.
22508 (sync_<optab><mode>): Likewise.
22509 (sync_old_<optab><mode>): Likewise.
22510 (sync_new_<optab><mode>): Likewise.
22511 (sync_nand<mode>): Likewise.
22512 (sync_old_nand<mode>): Likewise.
22513 (sync_new_nand<mode>): Likewise.
22514 (sync_lock_test_and_set<mode>): Likewise.
22515 (test_and_set_12): Likewise.
22516 (atomic_compare_and_swap<mode>): Likewise.
22517 (atomic_exchange<mode>_llsc): Likewise.
22518 (atomic_fetch_add<mode>_llsc): Likewise.
22519 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22520 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22521 (umips_save_restore_pattern_p): Likewise.
22522 (umips_load_store_pair_p): Likewise.
22523 (umips_output_load_store_pair): Likewise.
22524 (umips_movep_target_p): Likewise.
22525 (umips_12bit_offset_address_p): Likewise.
22526 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22527 (mips_base_mips16): Rename this...
22528 (mips_base_compression_flags): ...to this. Update all uses.
22529 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22530 (mips_mips16_decl_p): Delete.
22531 (mips_nomips16_decl_p): Delete.
33159866 22532 (mips_get_compress_on_flags): New function.
22c4c869
CM		 22533 (mips_get_compress_off_flags): New function.
22534 (mips_get_compress_mode): New function.
22535 (mips_get_compress_on_name): New function.
22536 (mips_get_compress_off_name): New function.
22537 (mips_insert_attributes): Support multiple compression types.
22538 (mips_merge_decl_attributes): Likewise.
22539 (umips_12bit_offset_address_p): New function.
22540 (mips_start_function_definition): Emit .set micromips directive.
22541 (mips_call_may_need_jalx_p): New function.
22542 (mips_function_ok_for_sibcall): Add microMIPS support.
22543 (mips_print_operand_punctuation): Support short delay slots and
22544 compact jumps.
22545 (umips_swm_mask, umips_swm_encoding): New.
22546 (umips_build_save_restore): New function.
22547 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22548 (was_mips16_p): Remove.
22549 (old_compression_mode): New.
22550 (mips_set_compression_mode): New function.
22551 (mips_set_current_function): Add microMIPS support.
22552 (mips_option_override): Likewise.
22553 (umips_save_restore_pattern_p): New function.
22554 (umips_output_save_restore): New function.
22555 (umips_load_store_pair_p_1): New function.
22556 (umips_load_store_pair_p): New function.
22557 (umips_output_load_store_pair_1): New function.
22558 (umips_output_load_store_pair): New function.
22559 (umips_movep_target_p) New function.
22560 (mips_prepare_pch_save): Add microMIPS support.
22561 * config/mips/mips.h (TARGET_COMPRESSION): New.
22562 (TARGET_CPU_CPP_BUILTINS): Update macro
22563 to use new compression flags and to support microMIPS.
22564 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22565 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22566 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22567 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22568 (ASM_SPEC): Support mmicromips and mno-micromips.
22569 (M16STORE_REG_P): New macro.
22570 (MIPS_CALL): Support TARGET_MICROMIPS.
22571 (MICROMIPS_J): New macro.
22572 (mips_base_mips16): Rename this...
22573 (mips_base_compression_flags): ...to this.
22574 (UMIPS_12BIT_OFFSET_P): New macro.
22575 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22576 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 225772013-03-20 Richard Biener <rguenther@suse.de>
22578
22579 PR tree-optimization/56661
22580 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22581 the result does not have to be distinct.
22582
54714c68
RB		 225832013-03-20 Richard Biener <rguenther@suse.de>
22584
22585 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22586 remap_gimple_op_r.
22587
cca1130d 225882013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 22589 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 22590
22591 PR rtl-optimization/56605
22592 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22593
2a930377
UB		 225942013-03-20 Uros Bizjak <ubizjak@gmail.com>
22595
22596 PR bootstrap/56656
22597 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22598 that require movd instead of movq.
22599
d6d305fe
RB		 226002013-03-20 Richard Biener <rguenther@suse.de>
22601
22602 * tree-ssa-structalias.c (struct variable_info): Add pointer
22603 to the first field of an aggregate with sub-vars. Make
22604 this and the pointer to the next subfield its ID.
22605 (vi_next): New function.
22606 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22607 storedanything_id, integer_id): Increment by one.
22608 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22609 get_call_clobber_vi): Adjust.
22610 (solution_set_expand): Simplify and speedup.
22611 (solution_set_add): Inline into ...
22612 (set_union_with_increment): ... this. Adjust accordingly.
22613 (do_sd_constraint): Likewise.
22614 (do_ds_constraint): Likewise.
22615 (do_complex_constraint): Simplify.
22616 (build_pred_graph): Adjust.
22617 (solve_graph): Likewise. Simplify and speedup.
22618 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22619 get_constraint_for_component_ref, get_constraint_for_1,
22620 first_vi_for_offset, first_or_preceding_vi_for_offset,
22621 create_function_info_for, create_variable_info_for_1,
22622 create_variable_info_for, intra_create_variable_infos): Adjust.
22623 (init_base_vars): Push NULL for ID zero.
22624 (compute_points_to_sets): Adjust.
22625
a271b42d
RB		 226262013-03-20 Richard Biener <rguenther@suse.de>
22627
22628 * cfgloop.c (verify_loop_structure): Streamline and avoid
22629 ICEing on corrupt loop tree.
22630 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22631 loop tree.
22632
ebd65954
RB		 226332013-03-20 Richard Biener <rguenther@suse.de>
22634
22635 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22636 check whether an SSA update is needed.
22637
4547b7ee
RS		 226382013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22639
22640 * config/mips/constraints.md (T): Rename to...
22641 (Yf): ...this.
22642 (U): Rename to...
22643 (Yd): ...this.
22644 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22645 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22646
0b8cdc58
IB		 226472013-03-19 Ian Bolton <ian.bolton@arm.com>
22648
22649 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22650 (*subsi3_carryin_uxtw): Likewise.
22651
5977a10d
IB		 226522013-03-19 Ian Bolton <ian.bolton@arm.com>
22653
22654 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22655 (*rorsi3_insn_uxtw): Likewise.
22656
bd83ff2c
IB		 226572013-03-19 Ian Bolton <ian.bolton@arm.com>
22658
22659 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22660 (*extrsi5_insn_uxtw): Likewise.
22661
ba65123c
RB		 226622013-03-19 Richard Biener <rguenther@suse.de>
22663
22664 PR tree-optimization/56273
22665 * passes.c (init_optimization_passes): Move second VRP after DOM.
22666
2eac0476
UB		 226672013-03-19 Uros Bizjak <ubizjak@gmail.com>
22668
22669 * config/i386/i386.md (*movti_internal): Merge from
22670 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22671 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22672 nox64 isa attributes.
22673
6e55eda7
RB		 226742013-03-18 Richard Biener <rguenther@suse.de>
22675
22676 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22677 (unite): Likewise.
22678 (merge_node_constraints): Likewise.
22679 (build_succ_graph): Likewise.
22680 (valid_graph_edge): Inline into single caller.
22681 (unify_nodes): Likewise. Use bitmap_set_bit return value
22682 and cache varinfo.
22683 (scc_visit): Fix formatting and variable use.
22684 (do_sd_constraint): Use gcc_checking_assert.
22685 (do_ds_constraint): Likewise.
22686 (do_complex_constraint): Likewise.
22687 (condense_visit): Likewise. Cleanup.
22688 (dump_pred_graph): New function.
22689 (perform_var_substitution): Dump the pred-graph before
22690 variable substitution.
22691 (find_equivalent_node): Use gcc_checking_assert.
22692 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22693
4bdd44c4
RB		 226942013-03-18 Richard Biener <rguenther@suse.de>
22695
22696 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22697 Remove cond_expr_stmt_list argument and do not gimplify the
22698 built expression.
22699 (vect_loop_versioning): Adjust.
22700 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22701 Cleanup to use less temporaries.
22702 (vect_create_data_ref_ptr): Cleanup.
22703
38c56a5b
JJ		 227042013-03-18 Jakub Jelinek <jakub@redhat.com>
22705
22706 PR tree-optimization/56635
22707 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22708 require types_compatible_p types.
22709
20b2e6a0
NC		 227102013-03-18 Nick Clifton <nickc@redhat.com>
22711
a6178a25
NC		 22712 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22713 spurious backslash.
22714
20b2e6a0
NC		 22715 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22716 Add missing line to comment describing function.
22717
92608d0e
RB		 227182013-03-18 Richard Biener <rguenther@suse.de>
22719
22720 PR tree-optimization/56210
22721 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22722 Handle string / character search functions.
22723 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22724
31348d52
RB		 227252013-03-18 Richard Biener <rguenther@suse.de>
22726
22727 PR middle-end/56483
22728 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22729 and implement properly.
22730 * gimple.h (gimple_cond_single_var_p): Remove.
22731
fcac74a1
RB		 227322013-03-18 Richard Biener <rguenther@suse.de>
22733
22734 * tree-data-ref.h (find_data_references_in_loop): Declare.
22735 * tree-data-ref.c (get_references_in_stmt): Use a stack
22736 vector pre-allocated in the callers.
22737 (find_data_references_in_stmt): Adjust.
22738 (graphite_find_data_references_in_stmt): Likewise.
22739 (create_rdg_vertices): Likewise.
22740 (find_data_references_in_loop): Export.
22741 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22742 Compute dependences here...
22743 (vect_analyze_data_refs): ...not here. When we encounter
22744 a non-vectorizable data reference in basic-block vectorization
22745 truncate the data reference vector. Do not bother to
22746 fixup data-dependence information for gather loads.
22747 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22748 of data references, as reported.
22749
0d5a1b56
RB		 227502013-03-18 Richard Biener <rguenther@suse.de>
22751
22752 PR tree-optimization/3713
22753 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22754 has_constants and expr.
22755 (stmt_has_constants): Properly valueize SSA names when deciding
22756 whether the stmt has constants.
22757
789c34e3
RB		 227582013-03-18 Richard Biener <rguenther@suse.de>
22759
22760 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22761 whole function when there is nothing to do.
22762 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22763 * tree-vectorizer.c (vectorize_loops): Update virtual and
22764 loop-closed SSA once.
22765 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22766
076b4605
RB		 227672013-03-18 Richard Biener <rguenther@suse.de>
22768
22769 PR middle-end/56113
22770 * domwalk.c (bb_postorder): New global static.
22771 (cmp_bb_postorder): New function.
22772 (walk_dominator_tree): Replace scheme imposing an order for
22773 visiting dominator sons by one sorting them at the time they
22774 are pushed on the stack.
22775
bdb01696
RB		 227762013-03-18 Richard Biener <rguenther@suse.de>
22777
22778 PR tree-optimization/39326
22779 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22780 (struct mem_ref): Replace mem member with ao_ref typed member.
22781 (MEM_ANALYZABLE): Adjust.
22782 (memref_eq): Likewise.
22783 (mem_ref_alloc): Likewise.
22784 (gather_mem_refs_stmt): Likewise.
22785 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22786 (execute_sm_if_changed_flag_set): Adjust.
22787 (execute_sm): Likewise.
22788 (ref_always_accessed_p): Likewise.
22789 (refs_independent_p): Likewise.
22790 (can_sm_ref_p): Likewise.
22791
12d2dc5e
JJ		 227922013-03-18 Jakub Jelinek <jakub@redhat.com>
22793
22794 PR c/56566
22795 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22796 return 1 even for !unsignedp.
22797
286fb677
UB		 227982013-03-17 Uros Bizjak <ubizjak@gmail.com>
22799
22800 * config/i386/i386.md (isa): Add x64 and nox64.
22801 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22802 (*pushtf): Enable *roF alternative for x64 isa only.
22803 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22804 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22805 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22806 (*movtf_internal): Merge from *movtf_internal_rex64 and
22807 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22808 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22809 nox64 isa attributes.
22810 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22811 nox64 isa attributes.
22812 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22813
88b97037
UB		 228142013-03-17 Uros Bizjak <ubizjak@gmail.com>
22815
22816 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22817
9f4f1735
JJ		 228182013-03-17 Jakub Jelinek <jakub@redhat.com>
22819
22820 PR target/56640
22821 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22822 class names. Remove trailing comma after "ALL_REGS".
22823
48f4a6fa
JH		 228242013-03-16 Jan Hubicka <jh@suse.cz>
22825
22826 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22827 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22828 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22829 of cgraph_get_create_node.
22830 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22831
98e81123
JM		 228322013-03-16 Jason Merrill <jason@redhat.com>
22833
22834 PR debug/49090
22835 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22836 with DW_AT_default_value.
22837
8a298c04
JJ		 228382013-03-16 Jakub Jelinek <jakub@redhat.com>
22839
22840 * BASE-VER: Set to 4.9.0.
22841
4323afa0
AK		 228422013-03-14 Andi Kleen <ak@linux.intel.com>
22843
22844 PR target/56619
22845 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22846 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22847 Document _x* TSX intrinsics.
22848
b3c0d469
JJ		 228492013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22850 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 22851
22852 * configure.ac: Add MicroBlaze TLS support detection.
22853 * configure: Regenerate.
88b97037
UB		 22854 * config/microblaze/microblaze-protos.h
22855 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22856 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22857 * config/microblaze/microblaze.c (microblaze_address_type): Add
22858 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 22859 (microblaze_address_info): Add tls_reloc.
22860 (TARGET_HAVE_TLS): Define.
22861 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22862 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 22863 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22864 load_tls_operand, microblaze_call_tls_get_addr,
22865 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 22866 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22867 (get_base_reg): Use microblaze_tls_symbol_p.
22868 (microblaze_classify_address): Handle TLS.
88b97037
UB		 22869 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22870 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 22871 (microblaze_legitimize_address): Handle TLS.
22872 (microblaze_address_insns): Handle ADDRESS_TLS.
22873 (pic_address_needs_scratch): Handle TLS.
22874 (print_operand_address): Handle TLS.
22875 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22876 (microblaze_expand_move): Handle TLS.
88b97037
UB		 22877 (microblaze_legitimate_constant_p): Check
22878 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 22879 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22880 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22881 (PIC_OFFSET_TABLE_REGNUM): Set.
22882 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22883 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22884 (addsi3, movsi_internal2, movdf_internal): Update constraints
22885 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 22886 (move_operand): Redefine as move_src_operand,
22887 check microblaze_tls_referenced_p.
8cc9a5a5 22888
d803a491
IB		 228892013-03-14 Ian Bolton <ian.bolton@arm.com>
22890
22891 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22892 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22893
a8504f22
IB		 228942013-03-14 Ian Bolton <ian.bolton@arm.com>
22895
22896 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22897 CC mode for AND.
22898
df2dfaea
JJ		 228992013-03-14 Jakub Jelinek <jakub@redhat.com>
22900
fbd28bc3
JJ		 22901 PR tree-optimization/53265
22902 * common.opt (Waggressive-loop-optimizations): New option.
22903 * tree-ssa-loop-niter.c: Include tree-pass.h.
22904 (do_warn_aggressive_loop_optimizations): New function.
22905 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22906 if number_of_latch_executions returned constant.
22907 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22908 early. If number_of_latch_executions returned constant, set
22909 nb_iterations_upper_bound back to it.
22910 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22911 field.
22912 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22913 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22914
df2dfaea
JJ		 22915 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22916 (MULTILIB_OSDIRNAMES): Set.
22917 * genmultilib: If defaultosdirname doesn't start with :: , set
22918 defaultosdirname2 instead, clear it and emit two . multilib_raw
22919 entries instead of just one.
22920
ee0d2b68
KK		 229212013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22922
22923 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22924 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22925 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22926 (SUBTARGET_OVERRIDE_OPTIONS): New.
22927
decc676e
OE		 229282013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22929
22930 PR target/49880
22931 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22932 (musermode): Convert to Var(TARGET_USERMODE).
22933 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22934 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22935 * config/sh/sh.c (sh_option_override): Use
22936 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22937 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22938 condition.
22939 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22940 TARGET_SH4.
22941 (udivsi3_i4_single, divsi3_i4_single): Use
22942 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22943
f4b719c7
DK		 229442013-03-13 Dave Korn <dave.korn.cygwin@....>
22945
22946 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22947 default setting.
22948
c40eced0
RB		 229492013-03-13 Richard Biener <rguenther@suse.de>
22950
22951 PR tree-optimization/56608
22952 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22953 calls when vectorizing basic-blocks.
22954
1bfa5973
JJ		 229552013-03-13 Jakub Jelinek <jakub@redhat.com>
22956
22957 PR plugins/45078
22958 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22959 tm_file.
22960
6fcf5434
JJ		 229612013-03-12 Jakub Jelinek <jakub@redhat.com>
22962
22963 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22964
4f38fa8c
JH		 229652013-03-11 Jan Hubicka <jh@suse.cz>
22966
22967 PR lto/56557
22968 * lto-streamer-out.c (output_symbol_p): Skip references from
22969 constructors of external variables.
22970
c5c5ba89
JH		 229712013-03-11 Jan Hubicka <jh@suse.cz>
22972
22973 PR middle-end/56571
22974 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22975 from pseudos.
22976 * emit-rtl.c (verify_rtx_sharing): Likewise.
22977 (copy_insn_1): Likewise.
22978 * rtl.c (copy_rtx): Likewise.
22979
c2a939b1
GJL		 229802013-03-11 Georg-Johann Lay <avr@gjlay.de>
22981
22982 PR target/56591
22983 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22984 output_operand_lossage message.
88b97037 22985
3c3279fb
RE		 229862013-03-11 Richard Earnshaw <rearnsha@arm.com>
22987
22988 PR target/56470
22989 * arm.c (shift_op): Validate RTL pattern on the fly.
22990 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22991 the RTL.
22992
aef5ef9d
JDA		 229932013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22994
22995 PR target/56347
22996 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22997 new call patterns that clobber %fr12.
22998 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22999 split and postreload patterns.
23000 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23001 registers %fr12 and %fr12R as call used.
23002
456610d3
SB		 230032013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23004
23005 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23006 (canon_address, record_store, replace_read, check_mem_read_rtx,
23007 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23008 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23009 rest_of_handle_dse): Likewise.
23010
4b1baac8
RS		 230112013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23012
23013 PR middle-end/56524
23014 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23015 Add base_optabs.
23016 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23017 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23018 (save_optabs_if_changed): Replace with...
23019 (init_tree_optimization_optabs): ...this.
23020 * optabs.c (save_optabs_if_changed): Rename to...
23021 (init_tree_optimization_optabs): ...this. Take the optimization node
23022 as argument. Do nothing if the base optabs are already correct.
23023 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23024 to recompute optabs.
23025 * function.h (function): Remove optabs field.
23026 * function.c (invoke_set_current_function_hook): Call
23027 init_tree_optimization_optabs. Use the result to initialize
23028 this_fn_optabs.
23029
b7a78683
AH		 230302013-02-27 Aldy Hernandez <aldyh@redhat.com>
23031
23032 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23033 if GTMA_HAS_NO_INSTRUMENTATION.
23034 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23035 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23036 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23037 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23038 GTMA_HAS_NO_INSTRUMENTATION.
23039
6384c29b
JJ		 230402013-03-08 Jakub Jelinek <jakub@redhat.com>
23041
23042 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23043 libasan_preinit.o.
23044
ecd4f20a
MP		 230452013-03-08 Marek Polacek <polacek@redhat.com>
23046 Jakub Jelinek <jakub@redhat.com>
23047
23048 PR tree-optimization/56478
23049 * predict.c (is_comparison_with_loop_invariant_p): Change the
23050 type of loop_step to tree.
23051 (predict_loops): Adjust.
23052 (predict_iv_comparison): Perform the computations on double_ints.
23053
64366d35
RB		 230542013-03-08 Richard Biener <rguenther@suse.de>
23055
23056 PR tree-optimization/56570
23057 * tree-cfg.c (verify_expr_location_1): Verify locations for
23058 DECL_DEBUG_EXPR.
23059 * tree-sra.c (create_access_replacement): Strip locations
23060 from DECL_DEBUG_EXPRs.
23061
a9d5a059
RB		 230622013-03-08 Richard Biener <rguenther@suse.de>
23063
23064 * tree-inline.c (expand_call_inline): Do not associate
23065 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23066 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23067
b7aa4e9a
RB		 230682013-03-08 Richard Biener <rguenther@suse.de>
23069
23070 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23071 or block changes with -Og. Fix for location / block encoding
23072 changes and PHI arguments with locations.
23073
c4c2f9fa
SB		 230742013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23075
23076 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23077 for all counters.
23078 (struct output_info): Likewise.
23079 (register_overhead): Remove bad gcc_assert.
23080 (bitmap_find_bit): If there is only a single bitmap element, do not
23081 count a miss as a search.
23082 (print_statistics): Update for counter type changes.
23083 (dump_bitmap_statistics): Likewise. Print headers such that they
23084 are properly lined up with the printed counters.
23085
5bf6606a
JJ		 230862013-03-07 Jakub Jelinek <jakub@redhat.com>
23087
23088 PR tree-optimization/56559
23089 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23090 check that it has only a single use.
23091
2c653d46
RB		 230922013-03-07 Richard Biener <rguenther@suse.de>
23093
23094 * doc/invoke.texi (fwhole-program): Discourage use in combination
23095 with -flto.
23096
a72d8780
JJ		 230972013-03-06 Jakub Jelinek <jakub@redhat.com>
23098
01a454df
JJ		 23099 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23100
c1781047
JJ		 23101 PR tree-optimization/56539
23102 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23103 instead of GSI_CONTINUE_LINKING as last argument to
23104 force_gimple_operand_gsi. Adjust function comment.
23105
9772c47a
JJ		 23106 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23107 aarch64-cores.def.
23108
a72d8780
JJ		 23109 PR middle-end/56548
23110 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23111 promoted mode, convert the result back to the original mode.
23112
fa5556de
RB		 231132013-03-06 Richard Biener <rguenther@suse.de>
23114
23115 PR middle-end/56294
23116 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23117 (insert_updated_phi_nodes_compare_uids): New function.
23118 (update_ssa): Sort symbols_to_rename after UID before
23119 traversing it to insert PHI nodes.
23120
010403d1
RB		 231212013-03-06 Richard Biener <rguenther@suse.de>
23122
23123 PR middle-end/50494
23124 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23125 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23126
23127 Revert
23128 2013-02-13 Richard Biener <rguenther@suse.de>
23129
23130 PR lto/50494
23131 * varasm.c (output_constant_def_1): Get the decl representing
23132 the constant as argument.
23133 (output_constant_def): Wrap output_constant_def_1.
23134 (make_decl_rtl): Use output_constant_def_1 with the decl
23135 representing the constant.
23136 (build_constant_desc): Optionally re-use a decl already
23137 representing the constant.
23138 (tree_output_constant_def): Adjust.
23139
3c27ce4c
JY		 231402013-03-06 Joey Ye <joey.ye@arm.com>
23141
23142 PR lto/50293
23143 * gcc.c (convert_white_space): New function.
23144 (main): Handles white space in function name.
23145
8f6d1c86
OE		 231462013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23147
23148 PR target/56529
23149 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23150 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23151 to SH_DIV_CALL_TABLE for TARGET_SH2.
23152 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23153 list.
23154 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23155 call-table options.
23156
798a209f
SA		 231572013-03-05 Sterling Augustine <saugustine@google.com>
23158 Cary Coutant <ccoutant@google.com>
23159
23160 PR debug/55364
23161 * dwarf2out.c (resolve_addr): Don't call
23162 remove_loc_list_addr_table_entries a second time for the same
23163 expression.
23164
6cfa417f 231652013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 23166
6cfa417f
JJ		 23167 PR debug/56510
23168 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23169 (avoid_complex_debug_insns): New function.
23170 (expand_debug_locations): Call it.
23171
23172 PR rtl-optimization/56484
23173 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23174 lifetimes of hard registers on small register class machines.
23175
231762013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23177
23178 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 23179 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 23180 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 23181 fast_interrupt.
23182 (microblaze_fast_interrupt_function_p): New function.
23183 (microblaze_is_interrupt_handler): Rename to
23184 microblaze_is_interrupt_variant and add fast_interrupt check.
23185 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23186 (save_restore_insns): Likewise.
23187 (compute_frame_size): Likewise.
23188 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23189 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 23190 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23191 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 23192 microblaze_is_interrupt_variant.
23193
731aefac
KT		 231942013-03-05 Kai Tietz <ktietz@redhat.com>
23195
23196 * sdbout.c (sdbout_one_type): Switch to current function's section
23197 supporting cold/hot.
23198
a72d8780 231992013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 23200
23201 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23202 -mxl-reorder.
23203
334e71e8
JJ		 232042013-03-05 Jakub Jelinek <jakub@redhat.com>
23205
0b50e654
JJ		 23206 PR middle-end/56461
23207 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23208 if VALGRIND_GET_VBITS is defined, temporarily make object
23209 memory all defined, and restore previous valgrind addressability
23210 and definability afterwards. Free this_object at the end.
23211
4ccf8f43
JJ		 23212 PR middle-end/56461
23213 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23214 right before calling lra_create_live_ranges, also call it
23215 when clearing live_p. Only call lra_clear_live_ranges
23216 at the end if live_p.
23217
334e71e8
JJ		 23218 PR middle-end/56461
23219 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23220
9ca966ca
RB		 232212013-03-05 Richard Biener <rguenther@suse.de>
23222
23223 PR tree-optimization/56521
23224 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23225 value-id.
23226
d273b176
SB		 232272013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23228
23229 PR c++/55135
23230 * except.h (remove_unreachable_eh_regions): New prototype.
23231 * except.c (remove_eh_handler_splicer): New function, split out
23232 of remove_eh_handler.
23233 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23234 warning about running it on many EH regions one at a time.
23235 (remove_unreachable_eh_regions_worker): New function, walk the
23236 EH tree in depth-first order and remove non-marked regions.
23237 (remove_unreachable_eh_regions): New function.
23238 * tree-eh.c (mark_reachable_handlers): New function, split out
23239 from remove_unreachable_handlers.
23240 (remove_unreachable_handlers): Use mark_reachable_handlers and
23241 remove_unreachable_eh_regions.
23242 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23243 and remove_unreachable_eh_regions.
23244
5e7f1aef
RB		 232452013-03-05 Richard Biener <rguenther@suse.de>
23246
23247 PR middle-end/56525
23248 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23249 not freeing them until the end.
23250
f276b762
AK		 232512013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23252
23253 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23254
12a43ab8
RB		 232552013-03-05 Richard Biener <rguenther@suse.de>
23256
23257 PR tree-optimization/56270
23258 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23259 of loads after scheduling an SLP instance.
23260
93675444
JJ		 232612013-03-05 Jakub Jelinek <jakub@redhat.com>
23262
db4138e3
JJ		 23263 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23264 tic6x.exp.
23265 (check_gcc_parallelize): Run guality.exp as a separate job from
23266 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23267 struct-layout-1.exp with stackalign.exp.
23268
dd3d1ec0
JJ		 23269 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23270
b8d381a3
JJ		 23271 PR middle-end/56461
23272 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23273 load_index sbitmap even if some bit in it isn't set.
23274
b4f9786b
JJ		 23275 PR middle-end/56461
23276 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23277 (discover_iteration_bound_by_body_walk): Change queues to
23278 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23279 spelling in comment. Call safe_push on queues[bound_index] directly.
23280 Release queues[queue_index] in every iteration unconditionally.
23281 Release bounds vector.
23282
93675444
JJ		 23283 PR middle-end/56461
23284 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23285 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23286 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23287 inner_phis vector.
23288
3e492e9c
RB		 232892013-03-05 Richard Biener <rguenther@suse.de>
23290
23291 PR lto/56515
23292 * tree-inline.c (remap_blocks_to_null): New function.
23293 (expand_call_inline): When expanding a call stmt without
23294 an associated block inline remap all callee blocks to NULL.
23295
a5d1569a
JJ		 232962013-03-05 Jakub Jelinek <jakub@redhat.com>
23297
86efb5cd
JJ		 23298 PR rtl-optimization/56494
23299 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23300 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23301 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23302
85f5dbea
JJ		 23303 PR middle-end/56461
23304 * sel-sched-ir.c (free_sched_pools): Release
23305 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23306 if succs_info_pool.max_top isn't -1.
23307
a5d1569a
JJ		 23308 PR bootstrap/56509
23309 * opts.c (opts_obstack, opts_concat): Moved to...
23310 * opts-common.c (opts_obstack, opts_concat): ... here.
23311
4432aa6c
JJ		 233122013-03-04 Jakub Jelinek <jakub@redhat.com>
23313
23314 PR middle-end/56461
23315 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23316
f3a1fb91
MJ		 233172013-03-04 Martin Jambor <mjambor@suse.cz>
23318
23319 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23320 all appropriate places.
23321
22110e6c
EB		 233222013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23323
23324 PR tree-optimization/56424
23325 * ipa-split.c (split_function): Do not set the RSO flag if result is
23326 not by reference and its type is a register type.
23327
a72d8780 233282013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23329
88b97037 23330 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23331 (microblaze_legitimate_pic_operand): Likewise
88b97037 23332 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 23333 new function microblaze_legitimate_pic_operand
23334 * config/microblaze/microblaze-protos.h
23335 (microblaze_legitimate_pic_operand): Declare.
23336
a72d8780 233372013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23338
a72d8780 23339 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23340 New predicate for supported rtx code types.
a72d8780 23341 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 23342 call_insn_simple_operand predicate.
23343
541d9ac8
JJ		 233442013-03-04 Jakub Jelinek <jakub@redhat.com>
23345
5eb010bc
JJ		 23346 PR middle-end/56461
23347 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23348 partitions.ordered_remove.
23349
30862efc
JJ		 23350 PR middle-end/56461
23351 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23352 vec_oprnds0.create (1) for modifier == NONE.
23353
8930f723
JJ		 23354 PR middle-end/56461
23355 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23356 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23357 vec_oprnds1 right before pushing anything to it for
23358 scalar_shift_arg.
23359
541d9ac8
JJ		 23360 PR middle-end/56461
23361 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23362 set nbbs to 0 instead of having separate code path.
23363 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23364 instead of false as last argument if returning NULL.
23365
7aa7f2e3
SL		 233662013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23367
23368 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23369 the attribute is now called "target" instead of "option".
23370 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23371 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23372 attribute/pragma name for TARGET_OPTION_VALID_P and
23373 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23374 * doc/tm.texi: Regenerated.
23375
8930f723 233762013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23377
8930f723 23378 * config/microblaze/microblaze.c:
cb8a1637 23379 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 23380 * config/microblaze/microblaze.h: Add -mxl-reorder to
23381 DRIVER_SELF_SPECS.
23382 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23383 instructions emitted if TARGET_REORDER.
88b97037
UB		 23384 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23385 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23386 separately.
cb8a1637 23387
4dc1d68c
XDL		 233882013-03-01 Xinliang David Li <davidxl@google.com>
23389
23390 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23391 walk length.
23392
689eaba3
JJ		 233932013-03-01 Jakub Jelinek <jakub@redhat.com>
23394
9d676bf6
JJ		 23395 PR middle-end/56461
23396 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23397 vector even when returning true. Fix up function comment formatting.
23398
caff7edf
JJ		 23399 PR middle-end/56461
23400 * ira-build.c (ira_loop_nodes_count): New variable.
23401 (create_loop_tree_nodes): Initialize it.
23402 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23403
b6b9227d
JJ		 23404 PR middle-end/56461
23405 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23406 method on dr_chain and result_chain.
23407 * tree-vect-stmts.c (vectorizable_store): Only call
23408 result_chain.create if j == 0.
23409
689eaba3
JJ		 23410 PR middle-end/56461
23411 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23412 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23413 before overwriting it.
23414
40bf31ed
TB		 234152013-03-01 Tobias Burnus <burnus@net-b.de>
23416
23417 * doc/extended.texi (C Extensions): Change order in @menu
23418 to match @node.
23419 (Other MIPS Built-in Functions): Move last MIPS entry before
23420 "picoChip Built-in Functions".
23421 (SH Built-in Functions): Move after RX Built-in Functions.
23422 * doc/gcc.texi (Introduction): Change order in @menu
23423 to match @node.
23424 * doc/md.texi (Constraints): Ditto.
23425 * gty.texi (Type Information): Ditto.
23426 (User-provided marking routines for template types): Make
23427 subsection.
23428 * doc/invoke.texi (AArch64 Options): Move before
23429 "Adapteva Epiphany Options".
23430
e664c61c
KS		 234312013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23432 Jakub Jelinek <jakub@redhat.com>
23433
23434 PR sanitizer/56454
23435 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23436 no_address_safety_analysis attribute.
23437 * doc/extend.texi (no_address_safety_attribute): Rename to
23438 no_sanitize_address attribute, mention no_address_safety_analysis
23439 attribute as deprecated alias.
23440
37b5ec8f
JJ		 234412013-02-28 Jakub Jelinek <jakub@redhat.com>
23442
23443 PR middle-end/56461
23444 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23445 type to vec<vec<tree> > *.
23446 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23447 to be vec<tree> instead of vec<tree> *, set vec_defs
23448 to vNULL and call vec_defs.create (number_of_vects), adjust other
23449 uses of vec_defs.
23450 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23451 vectorizable_condition): Adjust vect_get_slp_defs callers.
23452
ba96cdfb
JG		 234532013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23454
23455 * config/aarch64/aarch64.c
23456 (aarch64_float_const_representable): Remove unused variable.
23457
6f549691
JG		 234582013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23459
23460 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23461
af55e82d
JG		 234622013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23463
23464 * config/aarch64/aarch64-builtins.c
23465 (aarch64_init_simd_builtins): Make static.
23466
1df3f464
JG		 234672013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23468
23469 * config/aarch64/aarch64.c
23470 (aarch64_simd_make_constant): Make static.
23471
f8f42513
MJ		 234722013-02-28 Martin Jambor <mjambor@suse.cz>
23473
23474 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23475 with no initialization to the RHS of debug statements.
23476
b48b3fc4
MJ		 234772013-02-28 Martin Jambor <mjambor@suse.cz>
23478
23479 PR tree-optimization/56294
23480 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23481 Adjust dumping.
23482 (get_access_replacement): Do not call create_access_replacement.
23483 Assert a replacement exists.
23484 (get_repl_default_def_ssa_name): Create the replacement declaration
23485 itself.
23486
c3ae224c
RR		 234872013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23488
23489 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23490 final_end_function.
23491
45fa0eef
MP		 234922013-02-28 Marek Polacek <polacek@redhat.com>
23493
23494 PR rtl-optimization/56466
23495 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23496 if we're changing a loop.
23497 (peel_loops_completely): Likewise.
23498
502c067d
PC		 234992013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23500
23501 PR c++/55813
23502 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23503
ff24afc8
GJL		 235042013-02-28 Georg-Johann Lay <avr@gjlay.de>
23505
23506 PR target/56445
23507 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23508 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23509 INTX_FTYPE_FX, FX_FTYPE_INTX.
23510 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23511
c7b1fc1b
GJL		 235122013-02-28 Georg-Johann Lay <avr@gjlay.de>
23513
23514 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23515 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23516 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23517 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23518 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23519 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23520 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23521 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23522 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23523 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23524 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23525 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23526 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23527 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23528 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23529 (avrxmega6): Increase max flash segments from 5 to 6.
23530 * config/avr/t-multilib: Regenerate.
23531 * config/avr/avr-tables.opt: Regenerate.
23532 * doc/avr-mmcu.texi: Regenerate.
23533
4a0e3cfe
GJL		 235342013-02-28 Georg-Johann Lay <avr@gjlay.de>
23535
23536 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23537 (avr_device_to_arch): Rename to avr_device_to_ld.
23538 (avr_device_to_as): New prototype.
23539 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23540 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23541 * config/avr/driver-avr.c (avr_device_to_as): New.
23542 (avr_device_to_arch): Rename to avr_device_to_ld.
23543
97785e52
JJ		 235442013-02-27 Jakub Jelinek <jakub@redhat.com>
23545
3f292312
JJ		 23546 PR middle-end/56461
23547 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23548 method on dr_chain and result_chain.
23549
a344216b
JJ		 23550 PR middle-end/56461
23551 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23552 pointer_set_destroy on not_executed_last_iteration.
23553
f121ad02 23554 PR middle-end/56461
88b97037 23555 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 23556
307f83a3
JJ		 23557 PR middle-end/56461
23558 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23559 FOR_EACH_DEFINED_FUNCTION when freeing state.
23560
e19624ee
JJ		 23561 PR middle-end/56461
23562 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23563 pool_free.
23564 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23565 overwriting it.
23566
90e709fd
JJ		 23567 PR middle-end/56461
23568 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23569 known_aggs[i].items and release known_aggs vector.
23570
97785e52
JJ		 23571 PR middle-end/56461
23572 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23573
235742013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 23575
97785e52
JJ		 23576 * config/microblaze/microblaze.c (microblaze_emit_compare):
23577 Use xor for EQ/NE comparisions.
23578 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23579 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 23580
6fa5e0ed
JJ		 235812013-02-27 Jakub Jelinek <jakub@redhat.com>
23582
23583 PR middle-end/56461
23584 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23585 vector.
23586 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23587 vec_safe_push, always update *slot.
23588 (redirect_edge_var_map_clear): Use vec_free.
23589 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23590 (free_var_map_entry): Use vec_free.
23591 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23592 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23593
436a956a
AB		 235942013-02-27 Andrey Belevantsev <abel@ispras.ru>
23595
23596 PR middle-end/45472
436a956a
AB		 23597 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23598 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 23599 Reorder tests for speculativeness in the logical and operator.
23600
f5c2caca
JJ		 236012013-02-27 Jakub Jelinek <jakub@redhat.com>
23602
0fcb564b
JJ		 23603 * incpath.c (add_standard_paths): Use reconcat instead of concat
23604 where appropriate and avoid leaking memory.
23605
dc357798
JJ		 23606 * opts.h: Include obstack.h.
23607 (opts_concat): New prototype.
23608 (opts_obstack): New declaration.
23609 * opts.c (opts_concat): New function.
23610 (opts_obstack): New variable.
23611 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23612 (finish_options): Use opts_concat instead of concat
23613 and XOBNEWVEC instead of XNEWVEC.
23614 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23615 generate_option): Likewise.
23616 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23617 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23618
f5c2caca
JJ		 23619 PR target/56455
23620 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23621 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23622
d0163673
JJ		 236232013-02-26 Jakub Jelinek <jakub@redhat.com>
23624
23625 PR middle-end/56461
23626 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23627
b5ad2b8e
JR		 236282013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23629
23630 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23631 (arm_block_move_unaligned_straight): Likewise.
23632 (arm_adjust_block_mem): Likewise.
23633
9b639e2c
JR		 236342013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23635
23636 PR target/48901
23637 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23638 temp, cond and label.
23639 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23640
e13a0ccb
JR		 23641 PR target/52500
23642 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23643 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23644
fbe4f171
JR		 23645 PR target/52501
23646 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23647 of prologue/epilogue functions.
23648
ae006543
JR		 23649 PR target/52550
23650 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23651 Remove unused variable cfa_offset.
23652 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23653
c81369fa
JR		 23654 PR target/54639
23655 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23656 type promotion to unsigned.
23657
f8a8fea7
JR		 23658 PR target/54640
23659 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23660 for HOST_WIDE_INT of 32 bit / same size as int.
23661 (arm_block_move_unaligned_straight): Likewise.
23662 (arm_adjust_block_mem): Likewise.
23663
f8be5169
JR		 23664 PR target/54662
23665 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23666 ALL_CFLAGS.
23667
f1ad3354
MP		 236682013-02-26 Marek Polacek <polacek@redhat.com>
23669
23670 PR tree-optimization/56426
88b97037 23671 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 23672
a6af0f42
RB		 236732013-02-26 Richard Biener <rguenther@suse.de>
23674
23675 PR target/56444
23676 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23677 unused variable loops.
23678
aca43c6c
JJ		 236792013-02-26 Jakub Jelinek <jakub@redhat.com>
23680
cecbe5d9
JJ		 23681 PR tree-optimization/56448
23682 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23683 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23684 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23685 later operands of the references, or even first operand for
23686 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23687
aca43c6c
JJ		 23688 PR tree-optimization/56443
23689 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23690 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23691 to type_for_mode langhook.
23692
53e2e141
MT		 236932013-02-25 Matt Turner <mattst88@gmail.com>
23694
23695 * doc/invoke.texi: Document r4700.
23696
259ee451
RB		 236972013-02-25 Richard Biener <rguenther@suse.de>
23698
23699 PR tree-optimization/56175
23700 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23701 split out from ...
23702 (simplify_bitwise_binary): ... here. Also guard the conversion
23703 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23704
6d65e8f1
CM		 237052013-02-25 Catherine Moore <clm@codesourcery.com>
23706
23707 Revert:
23708 2013-02-24 Catherine Moore <clm@codesourcery.com>
23709 Maciej W. Rozycki <macro@codesourcery.com>
23710 Tom de Vries <tom@codesourcery.com>
a72d8780 23711 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 23712 Iain Sandoe <iain@codesourcery.com>
23713 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23714 Chao-ying Fu <fu@mips.com>
88b97037 23715
6d65e8f1 23716 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23717 Document new function attributes.
6d65e8f1
CM		 23718 * doc/invoke.texi (minterlink-compressed, mmicromips,
23719 m14k, m14ke, m14kec): Document new options.
23720 (minterlink-mips16): Update documentation.
23721 * doc/md.texi (ZC, ZD): Document new constraints.
23722 * configure.ac (gcc_cv_as_micromips): Check if linker
23723 supports the .set micromips directive.
23724 * configure: Regenerate.
23725 * config.in: Regenerate.
23726 * config/mips/mips-tables.opt: Regenerate.
23727 * config/mips/micromips.md: New file.
23728 * constraints.md (ZC, AD): New constraints.
23729 * config/mips/predicates.md (movep_src_register): New predicate.
23730 (movep_src_operand): New predicate.
23731 (non_volatile_mem_operand): New predicate.
23732 * config/mips/mips.md (multimem): New type.
23733 (length): Differentiate between 17-bit and 18-bit branch offsets.
23734 (MOVEP1, MOVEP2): New mode iterator.
23735 (mov_<load>l): Use ZC constraint.
23736 (mov_<load>r): Likewise.
23737 (mov_<store>l): Likewise.
23738 (mov_<store>r): Likewise.
23739 (*branch_equality<mode>_inverted): Add microMIPS support.
23740 (*branch_equality<mode>): Likewise.
23741 (*jump_absolute): Likewise.
23742 (indirect_jump_<mode>): Likewise.
23743 (tablejump_<mode>): Likewise.
23744 (<optab>_internal): Likewise.
23745 (sibcall_internal): Likewise.
23746 (sibcall_value_internal): Likewise.
23747 (prefetch): Use constraint ZD.
23748 * config/mips/mips.opt (minterlink-compressed): New option.
23749 (minterlink-mips16): Now an alias for minterlink-compressed.
23750 (mmicromips): New option.
23751 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23752 (compare_and_swap_12): Likewise.
23753 (sync_add<mode>): Likewise.
23754 (sync_<optab>_12): Likewise.
23755 (sync_old_<optab>_12): Likewise.
23756 (sync_new_<optab>_12): Likewise.
23757 (sync_nand_12): Likewise.
23758 (sync_old_nand_12): Likewise.
23759 (sync_new_nand_12): Likewise.
23760 (sync_sub<mode>): Likewise.
23761 (sync_old_add<mode>): Likewise.
23762 (sync_old_sub<mode>): Likewise.
23763 (sync_new_add<mode>): Likewise.
23764 (sync_new_sub<mode>): Likewise.
23765 (sync_<optab><mode>): Likewise.
23766 (sync_old_<optab><mode>): Likewise.
23767 (sync_new_<optab><mode>): Likewise.
23768 (sync_nand<mode>): Likewise.
23769 (sync_old_nand<mode>): Likewise.
23770 (sync_new_nand<mode>): Likewise.
23771 (sync_lock_test_and_set<mode>): Likewise.
23772 (test_and_set_12): Likewise.
23773 (atomic_compare_and_swap<mode>): Likewise.
23774 (atomic_exchange<mode>_llsc): Likewise.
23775 (atomic_fetch_add<mode>_llsc): Likewise.
23776 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23777 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23778 (umips_save_restore_pattern_p): Likewise.
23779 (umips_load_store_pair_p): Likewise.
23780 (umips_output_load_store_pair): Likewise.
23781 (umips_movep_target_p): Likewise.
23782 (umips_12bit_offset_address_p): Likewise.
23783 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23784 (mips_base_mips16): Rename this...
23785 (mips_base_compression_flags): ...to this. Update all uses.
23786 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23787 (mips_mips16_decl_p): Delete.
23788 (mips_nomips16_decl_p): Delete.
23789 (mips_get_compress_on_flags): New function.
23790 (mips_get_compress_off_flags): New function.
23791 (mips_get_compress_mode): New function.
23792 (mips_get_compress_on_name): New function.
23793 (mips_get_compress_off_name): New function.
23794 (mips_insert_attributes): Support multiple compression types.
23795 (mips_merge_decl_attributes): Likewise.
23796 (umips_12bit_offset_address_p): New function.
23797 (mips_start_function_definition): Emit .set micromips directive.
23798 (mips_call_may_need_jalx_p): New function.
23799 (mips_function_ok_for_sibcall): Add microMIPS support.
23800 (mips_print_operand_punctuation): Support short delay slots and
23801 compact jumps.
23802 (umips_swm_mask, umips_swm_encoding): New.
23803 (umips_build_save_restore): New function.
23804 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23805 (was_mips16_p): Remove.
23806 (old_compression_mode): New.
23807 (mips_set_compression_mode): New function.
23808 (mips_set_current_function): Add microMIPS support.
23809 (mips_option_override): Likewise.
23810 (umips_save_restore_pattern_p): New function.
23811 (umips_output_save_restore): New function.
23812 (umips_load_store_pair_p_1): New function.
23813 (umips_load_store_pair_p): New function.
23814 (umips_output_load_store_pair_1): New function.
23815 (umips_output_load_store_pair): New function.
23816 (umips_movep_target_p) New function.
23817 (mips_prepare_pch_save): Add microMIPS support.
23818 * config/mips/mips.h (TARGET_COMPRESSION): New.
23819 (TARGET_CPU_CPP_BUILTINS): Update macro
23820 to use new compression flags and to support microMIPS.
23821 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23822 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23823 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23824 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23825 (ASM_SPEC): Support mmicromips and mno-micromips.
23826 (M16STORE_REG_P): New macro.
23827 (MIPS_CALL): Support TARGET_MICROMIPS.
23828 (MICROMIPS_J): New macro.
23829 (mips_base_mips16): Rename this...
23830 (mips_base_compression_flags): ...to this.
23831 (UMIPS_12BIT_OFFSET_P): New macro.
23832 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23833 (MULTILIB_DIRNAMES): Likewise.
23834
5e5df392
TV		 238352013-02-25 Tom de Vries <tom@codesourcery.com>
23836
23837 PR rtl-optimization/56131
23838 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23839 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23840 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23841
3fdb53c1
TB		 238422013-02-25 Tobias Burnus <burnus@net-b.de>
23843
23844 * doc/invoke.texi (-fsanitize=): Move from optimization
23845 to debugging options.
23846
ed358aea
AB		 238472013-02-25 Andrey Belevantsev <abel@ispras.ru>
23848
23849 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23850
f45e9053
AB		 238512013-02-25 Andrey Belevantsev <abel@ispras.ru>
23852 Alexander Monakov <amonakov@ispras.ru>
23853
23854 PR middle-end/56077
23855 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23856 flush pending lists also on non-jumps. Adjust comment.
23857
6941b508
CM		 238582013-02-24 Catherine Moore <clm@codesourcery.com>
23859 Maciej W. Rozycki <macro@codesourcery.com>
23860 Tom de Vries <tom@codesourcery.com>
a72d8780 23861 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 23862 Iain Sandoe <iain@codesourcery.com>
23863 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23864 Chao-ying Fu <fu@mips.com>
6941b508
CM		 23865
23866 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23867 Document new function attributes.
6941b508
CM		 23868 * doc/invoke.texi (minterlink-compressed, mmicromips,
23869 m14k, m14ke, m14kec): Document new options.
23870 (minterlink-mips16): Update documentation.
23871 * doc/md.texi (ZC, ZD): Document new constraints.
23872 * configure.ac (gcc_cv_as_micromips): Check if linker
23873 supports the .set micromips directive.
23874 * configure: Regenerate.
23875 * config.in: Regenerate.
23876 * config/mips/mips-tables.opt: Regenerate.
23877 * config/mips/micromips.md: New file.
23878 * constraints.md (ZC, AD): New constraints.
23879 * config/mips/predicates.md (movep_src_register): New predicate.
23880 (movep_src_operand): New predicate.
23881 (non_volatile_mem_operand): New predicate.
23882 * config/mips/mips.md (multimem): New type.
23883 (length): Differentiate between 17-bit and 18-bit branch offsets.
23884 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 23885 (mov_<load>l): Use ZC constraint.
6941b508
CM		 23886 (mov_<load>r): Likewise.
23887 (mov_<store>l): Likewise.
23888 (mov_<store>r): Likewise.
23889 (*branch_equality<mode>_inverted): Add microMIPS support.
23890 (*branch_equality<mode>): Likewise.
23891 (*jump_absolute): Likewise.
23892 (indirect_jump_<mode>): Likewise.
23893 (tablejump_<mode>): Likewise.
23894 (<optab>_internal): Likewise.
23895 (sibcall_internal): Likewise.
23896 (sibcall_value_internal): Likewise.
23897 (prefetch): Use constraint ZD.
23898 * config/mips/mips.opt (minterlink-compressed): New option.
23899 (minterlink-mips16): Now an alias for minterlink-compressed.
23900 (mmicromips): New option.
23901 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23902 (compare_and_swap_12): Likewise.
23903 (sync_add<mode>): Likewise.
23904 (sync_<optab>_12): Likewise.
23905 (sync_old_<optab>_12): Likewise.
23906 (sync_new_<optab>_12): Likewise.
23907 (sync_nand_12): Likewise.
23908 (sync_old_nand_12): Likewise.
23909 (sync_new_nand_12): Likewise.
23910 (sync_sub<mode>): Likewise.
23911 (sync_old_add<mode>): Likewise.
23912 (sync_old_sub<mode>): Likewise.
23913 (sync_new_add<mode>): Likewise.
23914 (sync_new_sub<mode>): Likewise.
23915 (sync_<optab><mode>): Likewise.
23916 (sync_old_<optab><mode>): Likewise.
23917 (sync_new_<optab><mode>): Likewise.
23918 (sync_nand<mode>): Likewise.
23919 (sync_old_nand<mode>): Likewise.
23920 (sync_new_nand<mode>): Likewise.
23921 (sync_lock_test_and_set<mode>): Likewise.
23922 (test_and_set_12): Likewise.
23923 (atomic_compare_and_swap<mode>): Likewise.
23924 (atomic_exchange<mode>_llsc): Likewise.
23925 (atomic_fetch_add<mode>_llsc): Likewise.
23926 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23927 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23928 (umips_save_restore_pattern_p): Likewise.
23929 (umips_load_store_pair_p): Likewise.
23930 (umips_output_load_store_pair): Likewise.
23931 (umips_movep_target_p): Likewise.
23932 (umips_12bit_offset_address_p): Likewise.
23933 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23934 (mips_base_mips16): Rename this...
23935 (mips_base_compression_flags): ...to this. Update all uses.
23936 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23937 (mips_mips16_decl_p): Delete.
23938 (mips_nomips16_decl_p): Delete.
3fdb53c1 23939 (mips_get_compress_on_flags): New function.
6941b508
CM		 23940 (mips_get_compress_off_flags): New function.
23941 (mips_get_compress_mode): New function.
23942 (mips_get_compress_on_name): New function.
23943 (mips_get_compress_off_name): New function.
23944 (mips_insert_attributes): Support multiple compression types.
23945 (mips_merge_decl_attributes): Likewise.
23946 (umips_12bit_offset_address_p): New function.
23947 (mips_start_function_definition): Emit .set micromips directive.
23948 (mips_call_may_need_jalx_p): New function.
23949 (mips_function_ok_for_sibcall): Add microMIPS support.
23950 (mips_print_operand_punctuation): Support short delay slots and
23951 compact jumps.
23952 (umips_swm_mask, umips_swm_encoding): New.
23953 (umips_build_save_restore): New function.
23954 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23955 (was_mips16_p): Remove.
23956 (old_compression_mode): New.
23957 (mips_set_compression_mode): New function.
23958 (mips_set_current_function): Add microMIPS support.
23959 (mips_option_override): Likewise.
23960 (umips_save_restore_pattern_p): New function.
23961 (umips_output_save_restore): New function.
23962 (umips_load_store_pair_p_1): New function.
23963 (umips_load_store_pair_p): New function.
23964 (umips_output_load_store_pair_1): New function.
23965 (umips_output_load_store_pair): New function.
23966 (umips_movep_target_p) New function.
23967 (mips_prepare_pch_save): Add microMIPS support.
23968 * config/mips/mips.h (TARGET_COMPRESSION): New.
23969 (TARGET_CPU_CPP_BUILTINS): Update macro
23970 to use new compression flags and to support microMIPS.
23971 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23972 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23973 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23974 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23975 (ASM_SPEC): Support mmicromips and mno-micromips.
23976 (M16STORE_REG_P): New macro.
23977 (MIPS_CALL): Support TARGET_MICROMIPS.
23978 (MICROMIPS_J): New macro.
23979 (mips_base_mips16): Rename this...
23980 (mips_base_compression_flags): ...to this.
23981 (UMIPS_12BIT_OFFSET_P): New macro.
23982 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23983 (MULTILIB_DIRNAMES): Likewise.
23984
cdebbc6d
JJ		 239852013-02-24 Jakub Jelinek <jakub@redhat.com>
23986
23987 PR target/52555
23988 * target-globals.c (save_target_globals): For init_reg_sets and
23989 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23990
18c63565
JG		 239912013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23992
23993 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23994 * config/aarch64/t-aarch64
23995 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23996
2194f7a2
VM		 239972013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23998
23999 PR inline-asm/56148
24000 * lra-constraints.c (process_alt_operands): Reload operand
24001 conflicting with earlier clobber only if no more other conflicting
24002 operands.
24003
7d613735
JJ		 240042013-02-22 Jakub Jelinek <jakub@redhat.com>
24005
24006 PR sanitizer/56393
24007 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24008 if not linking a shared library.
24009
ac8d93a7
SL		 240102013-02-22 Seth LaForge <sethml@google.com>
24011
24012 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24013
e0237780
GY		 240142013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24015
3fdb53c1
TB		 24016 * config/arm/arm.md (split for extendsidi): Update condition.
24017 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24018 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24019 (qhs_zextenddi_cstr): Likewise.
e0237780 24020
d7fde18c
JJ		 240212013-02-21 Jakub Jelinek <jakub@redhat.com>
24022
be63b77d
JJ		 24023 PR middle-end/56420
24024 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24025 avoid signed wrapping.
24026 (expand_mult): Handle properly multiplication by
24027 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24028 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24029 in the compiler if coeff is HOST_WIDE_INT_MIN.
24030 (expand_divmod): Don't make ext_op1 static, change it's type to
24031 uhwi. Avoid undefined behavior in -INTVAL (op1).
24032
d7fde18c
JJ		 24033 PR rtl-optimization/50339
24034 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24035 field.
24036 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24037 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24038 into splitting_ashiftrt field.
24039 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24040 ASHIFTRT.
24041 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24042 choices.
24043
6aad4455
AH		 240442013-02-20 Aldy Hernandez <aldyh@redhat.com>
24045
24046 PR middle-end/56108
24047 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24048 are sure to go irrevocable.
24049
38fe784d
HPN		 240502013-02-21 Hans-Peter Nilsson <hp@axis.com>
24051
24052 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24053 scalars are valid operands.
24054
0fd44da3
MJ		 240552013-02-21 Martin Jambor <mjambor@suse.cz>
24056
24057 PR tree-optimization/56310
24058 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24059 only matching indices and non-negative final offsets.
24060 (intersect_aggregates_with_edge): Pass src_idx to
24061 agg_replacements_to_vector. Pass src_idx insstead of index to
24062 intersect_with_agg_replacements.
24063
7a92038b
MJ		 240642013-02-21 Martin Jambor <mjambor@suse.cz>
24065
24066 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24067 instead of hard-wired defaults.
24068
c0da9c37
MR		 240692013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24070
24071 * doc/invoke.texi (MIPS Options): Update documentation of the
24072 floating-point multiply-accumulate instruction restrictions.
24073
d247ea0c 240742013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 24075
24076 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24077 asan_shadow_offset on x86_64 linux.
24078
22deefcb
RB		 240792013-02-21 Richard Biener <rguenther@suse.de>
24080
24081 PR tree-optimization/56415
24082 Revert
24083 2013-02-11 Richard Biener <rguenther@suse.de>
24084
24085 PR tree-optimization/56273
24086 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24087 first VRP run.
24088
7bcc6e75
JJ		 240892013-02-21 Jakub Jelinek <jakub@redhat.com>
24090
a0ad148f
JJ		 24091 PR bootstrap/56258
24092 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24093 instead of @itemx.
24094
7bcc6e75
JJ		 24095 PR inline-asm/56405
24096 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24097 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24098
a0a7b611
JH		 240992013-02-20 Jan Hubicka <jh@suse.cz>
24100
24101 PR tree-optimization/56265
88b97037
UB		 24102 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24103 when target is referenced for first time.
a0a7b611 24104
c0e50f72
RB		 241052013-02-20 Richard Biener <rguenther@suse.de>
24106
24107 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24108 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24109 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24110 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24111 not return anything.
24112 (rename_ssa_copies): Do not remove unused locals.
24113 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 24114 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 24115 * passes.c (execute_function_todo): Do not schedule unused locals
24116 removal if cleanup_tree_cfg did something.
24117 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24118 about the number of removed locals.
24119
a52ca739
RB		 241202013-02-20 Richard Biener <rguenther@suse.de>
24121
24122 PR tree-optimization/56398
88b97037 24123 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 24124
ec9c9d1b
MJ		 241252013-02-20 Martin Jambor <mjambor@suse.cz>
24126
24127 PR tree-optimization/55334
24128 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24129 restricted pointers to arrays.
24130
e91c8ed6 241312013-02-20 Richard Biener <rguenther@suse.de>
88b97037 24132 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 24133
24134 PR tree-optimization/56396
24135 * tree-ssa-ccp.c (n_const_val): New static variable.
24136 (get_value): Return NULL for SSA names we don't have a lattice
24137 entry for.
24138 (ccp_initialize): Initialize n_const_val.
24139 * tree-ssa-copy.c (n_copy_of): New static variable.
24140 (init_copy_prop): Initialize n_copy_of.
24141 (get_value): Return NULL_TREE for SSA names we don't have a
24142 lattice entry for.
24143
3d916479
MJ		 241442013-02-20 Martin Jambor <mjambor@suse.cz>
24145
24146 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24147
71a86758
RB		 241482013-02-20 Richard Biener <rguenther@suse.de>
24149
24150 * genpreds.c (write_lookup_constraint): Do not compare first
24151 letter of the constraint again.
24152
79836a12
RB		 241532013-02-20 Richard Biener <rguenther@suse.de>
24154
24155 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24156 and ceil_log2.
24157 (get_use_iv_cost): Terminate hashtable walk when coming across
24158 an empty entry.
24159
bbe4fb2c
IZ		 241602013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24161
24162 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24163 reassociation for avx2 targets.
24164
a72d8780 241652012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 24166
c77f83d5
EI		 24167 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24168 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 24169 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 24170 version and TARGET_PATTERN_COMPARE check
24171 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 24172
a72d8780 241732012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 24174
a72d8780 24175 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 24176 function before branching.
24177
242387fa
AB		 241782012-02-19 Andrey Belevantsev <abel@ispras.ru>
24179
24180 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24181 DUMP_INSN_RTX_UID.
24182 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24183
b4979ab9
AB		 241842012-02-19 Andrey Belevantsev <abel@ispras.ru>
24185
24186 PR middle-end/55889
b4979ab9
AB		 24187 * sel-sched.c: Include ira.h.
24188 (implicit_clobber_conflict_p): New function.
24189 (moveup_expr): Use it.
88b97037 24190 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 24191
24d63016
RB		 241922013-02-19 Richard Biener <rguenther@suse.de>
24193
24194 PR tree-optimization/56384
24195 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24196 (vn_hash_type): Split out from ...
24197 (vn_hash_constant_with_type): ... here.
24198 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24199 (vn_phi_eq): Compare types from vn_phi_s structure.
24200 (vn_phi_lookup): Populate vn_phi_s type.
24201 (vn_phi_insert): Likewise.
24202
a475fd3d
JJ		 242032013-02-19 Jakub Jelinek <jakub@redhat.com>
24204
47cc28f5
JJ		 24205 PR tree-optimization/56350
24206 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24207 if haven't found reduction or nested cycle operand, rather than
24208 asserting we must find it.
24209
a475fd3d
JJ		 24210 PR tree-optimization/56381
24211 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24212 to fold_build3.
24213
242142013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 24215 Jakub Jelinek <jakub@redhat.com>
24216
24217 PR target/52555
24218 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24219 (swap_optab_enable): Same.
24220 (init_all_optabs): Use argument instead of global.
88b97037 24221 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 24222 * expr.h (init_all_optabs): Add argument to prototype.
24223 (TREE_OPTIMIZATION_OPTABS): New.
24224 (save_optabs_if_changed): Protoize.
24225 * optabs.h: Declare this_fn_optabs.
24226 * optabs.c (save_optabs_if_changed): New.
24227 Declare this_fn_optabs.
24228 (init_optabs): Add argument to init_all_optabs() call.
24229 * function.c (invoke_set_current_function_hook): Handle per
24230 function optabs.
24231 * function.h (struct function): New field optabs.
24232 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24233 optimization_current_node has changed.
24234 * target-globals.h (save_target_globals_default_opts): Protoize.
24235 * target-globals.c (save_target_globals_default_opts): New.
24236
3f587ca3
JDA		 242372013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24238
24239 PR target/56347
24240 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24241 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 24242
24243 PR target/56214
24244 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 24245 and HImode, require all displacements to be an integer multiple of
24246 their mode size.
ceaca33e
JDA		 24247 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24248 only allow QImode and HImode when reload is in progress and strict is
24249 true. Likewise for symbolic addresses. Use base14_operand to check
24250 displacements in REG+BASE addresses.
24251
fe0b4796
RB		 242522013-02-18 Richard Biener <rguenther@suse.de>
24253
24254 PR tree-optimization/56366
24255 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24256 handle sign-conversion of outer-loop initial induction value.
24257
6aaf596b
RB		 242582013-02-18 Richard Biener <rguenther@suse.de>
24259
73db8ff1 24260 PR middle-end/56349
6aaf596b
RB		 24261 * cfghooks.c (merge_blocks): If we merge a latch into another
24262 block adjust references to it.
24263 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24264 (verify_loop_structure): Verify that a recorded latch is in fact
24265 a latch.
24266
5e97dfb6
RB		 242672013-02-18 Richard Biener <rguenther@suse.de>
24268
24269 PR tree-optimization/56321
24270 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24271 order SSA name release and virtual operand unlinking.
24272
825527e8
EI		 242732013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24274
24275 * config/microblaze/microblaze.md (save_stack_block): Define.
24276 (restore_stack_block): Likewise.
24277
debd11d9
EI		 242782013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24279
24280 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24281 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24282 * config/microblaze/microblaze.c (microblaze_option_override):
24283 Bail out early for PIC modes when target does not support PIC.
24284
8ec77be0
EI		 242852013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24286
88b97037 24287 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 24288 Replace with a microblaze version.
24289 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 24290 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 24291 microblaze.
24292
b41288b3
JJ		 242932013-02-16 Jakub Jelinek <jakub@redhat.com>
24294 Dodji Seketeli <dodji@redhat.com>
24295
24296 PR asan/56330
88b97037 24297 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 24298 (instrument_mem_region_access): Do not forget to always put
24299 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24300 0) statement, even for cases where either 'base' or 'base + len'
24301 are not instrumented -- because they have been previously
24302 instrumented. Simplify the logic by putting all the statements
24303 instrument 'base + len' inside a sequence, and then insert that
24304 sequence right before the current insertion point. Then, to
24305 instrument 'base + len', just get an iterator on that statement.
24306 And do not forget to update the pointer to iterator the function
24307 received as argument.
24308
47918951
VM		 243092013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24310
24311 PR rtl-optimization/56348
24312 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24313
3fb7c699
SB		 243142013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24315
24316 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24317 (clean_graph_dump_file): Pass base to start_graph_dump.
24318
576fe41a
RH		 243192013-02-14 Richard Henderson <rth@redhat.com>
24320
24321 PR target/55941
24322 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24323
bb3accfa
SB		 243242013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24325
24326 * collect2-aix.h: Define F_LOADONLY.
24327
03143140
RB		 243282013-02-14 Richard Biener <rguenther@suse.de>
24329
24330 PR lto/50494
24331 * varasm.c (output_constant_def_1): Get the decl representing
24332 the constant as argument.
24333 (output_constant_def): Wrap output_constant_def_1.
24334 (make_decl_rtl): Use output_constant_def_1 with the decl
24335 representing the constant.
24336 (build_constant_desc): Optionally re-use a decl already
24337 representing the constant.
24338 (tree_output_constant_def): Adjust.
24339
8fb06726
DS		 243402013-02-14 Dodji Seketeli <dodji@redhat.com>
24341
24342 Fix an asan crash
24343 * asan.c (instrument_builtin_call): Really put the length of the
24344 second source argument into src1_len.
24345
c4bfe8bf
JJ		 243462013-02-13 Jakub Jelinek <jakub@redhat.com>
24347
24348 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24349 argument. If it is false, don't create edge from then_bb to
24350 fallthru_bb.
24351 (insert_if_then_before_iter): Pass true to it.
24352 (build_check_stmt): Pass false to it.
24353 (transform_statements): Flush hash table only on extended basic
24354 block boundaries, rather than at the beginning of every bb.
24355 Don't flush hash table on nonfreeing_call_p calls.
24356 * tree-flow.h (nonfreeing_call_p): New prototype.
24357 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24358
7afe2801
DM		 243592013-02-13 David S. Miller <davem@davemloft.net>
24360
24361 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24362
70cc3288
VM		 243632013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24364
24365 PR target/56184
24366 * ira.c (max_regno_before_ira): Move from ...
24367 (ira): ... here.
24368 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24369 vec_safe_length.
24370
6422242b
JJ		 243712013-02-13 Jakub Jelinek <jakub@redhat.com>
24372
24373 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24374
976a81ee
RB		 243752013-02-13 Richard Biener <rguenther@suse.de>
24376
24377 PR lto/56295
24378 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24379 globals in MEM_REFs.
24380
c1874a87
RB		 243812013-02-13 Richard Biener <rguenther@suse.de>
24382
24383 * loop-init.c (loop_optimizer_init): Clear loop state when
24384 re-initializing preserved loops.
24385 * loop-unswitch.c (unswitch_single_loop): Return whether
24386 we unswitched the loop. Do not verify loop state here.
88b97037 24387 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24388
b4ab7d34
KS		 243892013-02-13 Kostya Serebryany <kcc@google.com>
24390
88b97037
UB		 24391 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24392 on x86_64 linux.
b4ab7d34
KS		 24393 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24394
bdcbe80c
DS		 243952013-02-12 Dodji Seketeli <dodji@redhat.com>
24396
24397 Avoid instrumenting duplicated memory access in the same basic block
24398 * Makefile.in (asan.o): Add new dependency on hash-table.h
24399 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24400 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24401 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24402 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24403 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24404 (get_mem_ref_of_assignment): New functions.
24405 (get_mem_refs_of_builtin_call): Extract from
24406 instrument_builtin_call and tweak a little bit to make it fit with
24407 the new signature.
24408 (instrument_builtin_call): Use the new
24409 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24410 of is_gimple_builtin_call.
24411 (instrument_derefs, instrument_mem_region_access): Insert the
24412 instrumented memory reference into the hash table.
24413 (maybe_instrument_assignment): Renamed instrument_assignment into
24414 this, and change it to advance the iterator when instrumentation
24415 actually happened and return true in that case. This makes it
24416 homogeneous with maybe_instrument_assignment, and thus give a
24417 chance to callers to be more 'regular'.
24418 (transform_statements): Clear the memory reference hash table
24419 whenever we enter a new BB, when we cross a function call, or when
24420 we are done transforming statements. Use
24421 maybe_instrument_assignment instead of instrumentation. No more
24422 need to special case maybe_instrument_assignment and advance the
24423 iterator after calling it; it's now handled just like
24424 maybe_instrument_call. Update comment.
24425
4861a1f7
RB		 244262013-02-13 Richard Biener <rguenther@suse.de>
24427
24428 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24429 Fix loop discovery code.
24430
1c86bd80
VM		 244312013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24432
24433 PR inline-asm/56148
24434 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24435 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 24436 the operand is not reloaded. Prefer to reload conflicting operand
24437 if earlyclobber and matching operands are the same.
24438
7cbda518
RB		 244392013-02-12 Richard Biener <rguenther@suse.de>
24440
24441 PR lto/56297
24442 * lto-streamer-out.c (write_symbol): Do not output symbols
24443 for hard register variables.
24444
e68a4ef6
GJL		 244452013-02-12 Georg-Johann Lay <avr@gjlay.de>
24446
24447 PR target/54222
24448 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24449 (umulsidi3_insn, mulsidi3_insn): New insns.
24450
a72d8780 244512013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 24452
24453 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24454 (struct tune_params): Add vec_costs field.
24455 * config/arm/arm.c (arm_builtin_vectorization_cost)
24456 (arm_add_stmt_cost): New functions.
24457 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24458 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24459 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24460 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24461 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24462 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24463 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24464
43320568
RB		 244652013-02-12 Richard Biener <rguenther@suse.de>
24466
24467 PR lto/56295
24468 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24469 decls again if possible.
24470
a011aa39
RB		 244712013-02-12 Richard Biener <rguenther@suse.de>
24472
24473 PR middle-end/56288
24474 * tree-ssa.c (verify_ssa_name): Fix check, move
24475 SSA_NAME_IN_FREE_LIST check up.
24476
6da26889
JJ		 244772013-02-12 Jakub Jelinek <jakub@redhat.com>
24478 Steven Bosscher <steven@gcc.gnu.org>
24479
24480 PR rtl-optimization/56151
24481 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24482 equal to op0 or op1, and last_insn pattern is CODE operation
24483 with MEM dest and one of the operands matches that MEM.
24484
f80e0faf
ST		 244852013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24486
24487 * doc/extend.texi: Document Function Multiversioning and "default"
24488 parameter string to target attribute.
24489 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24490 target attribute parameter is "default".
24491 (ix86_compare_version_priority): Remove checks for target attribute.
24492 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24493 Remove check for target attribute equal to NULL. Add assert.
24494 (ix86_generate_version_dispatcher_body): Change error to sorry.
24495
6c59ffd1
IS		 244962013-02-11 Iain Sandoe <iain@codesourcery.com>
24497 Jack Howarth <howarth@bromo.med.uc.edu>
24498 Patrick Marlier <patrick.marlier@gmail.com>
24499
24500 PR libitm/55693
24501 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24502 define ENDFILE_SPEC as TM_DESTRUCTOR.
24503 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24504
a72d8780 245052013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 24506 Jack Howarth <howarth@bromo.med.uc.edu>
24507 Jakub Jelinek <jakub@redhat.com>
24508
24509 PR sanitizer/55617
24510 * config/darwin.c (cdtor_record): Rename ctor_record.
24511 (sort_cdtor_records): Rename sort_ctor_records.
24512 (finalize_dtors): New routine to sort destructors by
24513 priority before use in assemble_integer.
24514 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24515
b63fe007
UB		 245162013-02-11 Uros Bizjak <ubizjak@gmail.com>
24517
24518 PR rtl-optimization/56275
24519 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24520 offset is non-negative and less than cmode size before
24521 calling simplify_subreg.
24522
8e89b5b5
RB		 245232013-02-11 Richard Biener <rguenther@suse.de>
24524
24525 PR tree-optimization/56264
24526 * cfgloop.h (fix_loop_structure): Adjust prototype.
24527 * loop-init.c (fix_loop_structure): Return the number of
24528 newly discovered loops.
24529 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24530 are discovered, do a full loop-closed SSA rewrite.
24531
b4a4b56d
RB		 245322013-02-11 Richard Biener <rguenther@suse.de>
24533
24534 PR tree-optimization/56273
24535 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24536 first VRP run.
24537 (check_array_ref): Fix missing newline in dumps.
24538 (search_for_addr_array): Likewise.
24539
0c885229
DE		 245402013-02-09 David Edelsohn <dje.gcc@gmail.com>
24541
24542 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24543
59ac9a55
JJ		 245442013-02-09 Jakub Jelinek <jakub@redhat.com>
24545
24546 PR target/56256
24547 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24548
25bb0bb5
VM		 245492013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24550
24551 PR rtl-optimization/56246
0c885229 24552 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 24553 reload pseudo.
24554 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24555 constraints are satisfied.
24556
a698cc03
JL		 245572013-02-08 Jeff Law <law@redhat.com>
24558
24559 PR debug/53948
24560 * emit-rtl.c (reg_is_parm_p): New function.
24561 * regs.h (reg_is_parm_p): New prototype.
24562 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24563 callee-clobbered registers.
24564
e1122ddd
MM		 245652013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24566
24567 PR target/56043
24568 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24569 If there is no implicit builtin declaration, just return NULL.
24570
19c5f6e6
UB		 245712013-02-08 Uros Bizjak <ubizjak@gmail.com>
24572
24573 * config/i386/sse.md (FMAMODEM): New mode iterator.
24574 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24575 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24576
2480f2ca 245772013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 24578
2480f2ca
UB		 24579 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24580 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24581 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24582
245832013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24584
24585 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24586 (microblaze*-*-elf): Likewise.
24587 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24588 LINK_SPEC.
24589 * config/microblaze/microblaze-c.c: Add builtin defines for
24590 _LITTLE_ENDIAN and _BIG_ENDIAN.
24591 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24592 add to TARGET_DEFAULT flags.
76ef61fb 24593 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 24594 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24595 * config/microblaze/microblaze.md: Update extendsidi2 and
24596 movdi_internal instructions to use low-order / high-order reg
24597 print_operands.
24598 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24599 options and inversemask / mask of LITTLE_ENDIAN.
24600 * config/microblaze/t-microblaze: Expand multilib options to
24601 include mlittle-endian (le) and update exceptions patterns.
24602
600a5961
JJ		 246032013-02-08 Jakub Jelinek <jakub@redhat.com>
24604
5df81313
JJ		 24605 PR rtl-optimization/56195
24606 * lra-constraints.c (get_reload_reg): Don't reuse regs
24607 if they have smaller mode than requested, if they have
24608 wider mode than requested, try to return a SUBREG.
24609
600a5961
JJ		 24610 PR tree-optimization/56250
24611 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24612 if type is unsigned and code isn't MULT_EXPR.
24613
ff544649
GJL		 246142013-02-08 Georg-Johann Lay <avr@gjlay.de>
24615
24616 PR tree-optimization/56064
24617 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24618 bits according to mode.
24619 * fixed-value.h (fixed_from_double_int)
24620 (const_fixed_from_double_int): Adjust comments.
24621
e45cde98
RB		 246222013-02-08 Richard Biener <rguenther@suse.de>
24623
24624 PR lto/56231
24625 * lto-streamer.h (struct data_in): Remove current_file, current_line
24626 and current_col members.
24627 * lto-streamer-out.c (lto_output_location): Stream changed bits
24628 en-block for efficiency.
24629 * lto-streamer-in.c (clear_line_info): Remove.
24630 (lto_input_location): Cache current file, line and column
24631 globally via local statics. Read changed bits en-block.
24632 (input_function): Do not call clear_line_info.
24633 (lto_read_body): Likewise.
24634 (lto_input_toplevel_asms): Likewise.
24635
c1ca73d8
MM		 246362013-02-08 Michael Matz <matz@suse.de>
24637
24638 PR tree-optimization/52448
24639 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24640 (nt_call_phase): New static.
24641 (add_or_mark_expr): Only mark accesses with newer phase than any
24642 call seen.
24643 (nonfreeing_call_p): New.
24644 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24645 (nt_fini_block): Keep blocks marked as visited.
24646 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24647
57548aa2
RB		 246482013-02-08 Richard Biener <rguenther@suse.de>
24649
24650 * ira.c (ira): Free broken dominator information.
24651
8e10366f
UB		 246522013-02-08 Uros Bizjak <ubizjak@gmail.com>
24653
24654 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24655
f64fb0fa
MP		 246562013-02-08 Marek Polacek <polacek@redhat.com>
24657
8e10366f 24658 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 24659
0375167b
RB		 246602013-02-08 Richard Biener <rguenther@suse.de>
24661
24662 PR middle-end/56181
24663 * cfgloop.h (flow_loops_find): Adjust.
24664 (bb_loop_header_p): Declare.
24665 * cfgloop.c (bb_loop_header_p): New function split out from ...
24666 (flow_loops_find): ... here. Adjust function signature,
24667 support incremental loop structure update.
24668 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24669 * cfgloopmanip.c (fix_loop_structure): Move ...
24670 * loop-init.c (fix_loop_structure): ... here.
24671 (apply_loop_flags): Split out from ...
24672 (loop_optimizer_init): ... here.
24673 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24674 in incremental mode, only remove dead loops here.
24675
85d768f3
GJL		 246762013-02-08 Georg-Johann Lay <avr@gjlay.de>
24677
24678 PR target/54222
24679 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24680 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24681 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24682 (*round<mode>3.libgcc): New insns for fixed-modes.
24683 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24684 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24685 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24686 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24687 implementations. Define to __builtin_avr_absFX,
24688 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24689 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24690 __builtin_avr_countlsFX, respectively.
24691 * config/avr/avr-c.c (target.h): Include it.
24692 (enum avr_builtin_id): New enum.
24693 (avr_resolve_overloaded_builtin): New static function.
24694 (avr_register_target_pragmas): Use it to set
24695 targetm.resolve_overloaded_builtin.
24696 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24697 tree nodes used by DEF_BUILTIN.
24698 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24699 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24700 <AVR_BUILTIN_xxBITS>: Same.
24701
661bc682
RB		 247022013-02-08 Richard Biener <rguenther@suse.de>
24703
24704 * cfgloop.c (verify_loop_structure): Properly handle
24705 a loop exiting to another loop header.
24706 * ira-int.h (ira_loops): Remove.
24707 * ira.c (ira_loops): Remove.
24708 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24709 (do_reload): Use loop_optimizer_finalize.
24710 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24711 number_of_loops to access the loop tree.
24712 (more_one_region_p): Likewise.
24713 (finish_loop_tree_nodes): Likewise.
24714 (rebuild_regno_allocno_maps): Likewise.
24715 (mark_loops_for_removal): Likewise.
24716 (mark_all_loops_for_removal): Likewise.
24717 (remove_unnecessary_regions): Likewise.
24718 (ira_build): Likewise.
24719 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24720
0d5049b2
RB		 247212013-02-08 Richard Biener <rguenther@suse.de>
24722
24723 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24724 * ipa-pure-const.c (analyze_function): Avoid calling
24725 mark_irreducible_loops twice.
8e10366f 24726 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 24727
499f32e8
DM		 247282013-02-07 David S. Miller <davem@davemloft.net>
24729
24730 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24731 on 'reg'.
24732 * var-tracking.c (vt_add_function_parameter): Test the presence of
24733 HAVE_window_save properly and do not remap argument registers when
24734 we have a leaf function.
24735
6edc3e32
UB		 247362013-02-07 Uros Bizjak <ubizjak@gmail.com>
24737
24738 PR bootstrap/56227
24739 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24740 instead of "ll".
24741 * config/i386/i386.c (ix86_print_operand): Ditto.
24742
5306401f
VM		 247432013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24744
6edc3e32 24745 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 24746
027ece11
VM		 247472013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24748
24749 PR rtl-optimization/56225
24750 * lra-constraints.c (process_alt_operands): Check that reload hard
24751 reg can hold value for strict_low_part.
24752
f980dfdb
JJ		 247532013-02-07 Jakub Jelinek <jakub@redhat.com>
24754
24755 PR debug/56154
24756 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24757 dwarf2out_end_function.
24758 (in_first_function_p, maybe_at_text_label_p,
24759 first_loclabel_num_not_at_text_label): New variables.
24760 (dwarf2out_var_location): In the first function find out
6edc3e32 24761 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 24762 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24763 functions.
24764
73dd3123
EB		 247652013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24766
24767 PR rtl-optimization/56178
24768 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24769 SUBREG of a register. Tidy up related block of code.
24770 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24771 note if the source is a register or a SUBREG of a register.
24772
e3936f47
JJ		 247732013-02-07 Jakub Jelinek <jakub@redhat.com>
24774
24775 PR target/56228
24776 * config/rs6000/rs6000.md (ptrm): New mode attr.
24777 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24778 call_value_indirect_aix<pttrsize>,
24779 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24780 m in constraints.
24781
d96d674b
MH		 247822013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24783
24784 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24785 if -bnortl. Convert to strcmp and strncmp.
24786
921f2dee
AM		 247872013-02-07 Alan Modra <amodra@gmail.com>
24788
24789 PR target/54009
24790 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24791 addresses won't wrap when offsetting.
24792 (rs6000_secondary_reload): Provide secondary reloads needed for
24793 wrapping LO_SUM addresses.
24794
d09c7dba
TS		 247952013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24796
24797 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24798 MACH, just __MACH__.
24799
a44bbd48
RB		 248002013-02-06 Richard Biener <rguenther@suse.de>
24801
24802 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24803 instead of calling fix_loop_structure.
24804
6d840d99
JJ		 248052013-02-06 Jakub Jelinek <jakub@redhat.com>
24806
24807 PR middle-end/56217
24808 * omp-low.c (use_pointer_for_field): Return false if
24809 lower_send_shared_vars doesn't generate any copy-out code.
24810
0f33baa9
TV		 248112013-02-06 Tom de Vries <tom@codesourcery.com>
24812
24813 PR rtl-optimization/56131
24814 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24815 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24816 of the label is NULL. Add comment.
24817
11f1e3ab
JJ		 248182013-02-05 Jakub Jelinek <jakub@redhat.com>
24819
31502f9f
JJ		 24820 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24821
a74db9bd
JJ		 24822 PR sanitizer/55374
24823 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24824 (STATIC_LIBTSAN_LIBS): Likewise.
24825 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24826 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24827 is defined, don't add anything else beyond that.
24828 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24829 (LINK_COMMAND_SPEC): Use them.
24830
11f1e3ab
JJ		 24831 PR tree-optimization/56205
24832 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24833 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24834 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24835
14ac6aa2
RB		 248362013-02-05 Richard Biener <rguenther@suse.de>
24837
24838 PR tree-optimization/53342
24839 PR tree-optimization/53185
24840 * tree-vectorizer.h (vect_check_strided_load): Remove.
24841 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24842 not disallow peeling for vectorized strided loads.
24843 (vect_check_strided_load): Make static and simplify.
24844 (vect_analyze_data_refs): Adjust.
24845 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24846 correctly when vectorizing strided loads.
24847
6f22445a
RB		 248482013-02-05 Richard Biener <rguenther@suse.de>
24849
24850 * doc/install.texi: Refer to ISL, not PPL.
24851
39f9719e
JH		 248522013-02-05 Jan Hubicka <jh@suse.cz>
24853
ec4224ac
JH		 24854 PR tree-optimization/55789
24855 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24856
248572013-02-05 Jan Hubicka <jh@suse.cz>
24858
24859 PR tree-optimization/55789
39f9719e
JH		 24860 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24861 the dead call anyway.
24862
956d3b33
EB		 248632013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24864
24865 PR sanitizer/55374
24866 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24867
248682013-02-04 Alexander Potapenko <glider@google.com>
24869 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 24870 Jakub Jelinek <jakub@redhat.com>
24871
24872 PR sanitizer/55617
24873 * config/darwin.c (sort_ctor_records): Stabilized qsort
24874 on constructor priority by using original position.
24875 (finalize_ctors): New routine to sort constructors by
24876 priority before use in assemble_integer.
24877 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24878
7ac3af38
JJ		 248792013-02-04 Jakub Jelinek <jakub@redhat.com>
24880
24881 PR libstdc++/54314
24882 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24883 about visibility on artificial decls.
24884 * config/sol2.c (solaris_assemble_visibility): Likewise.
24885
152689dc
KT		 248862013-02-04 Kai Tietz <ktietz@redhat.com>
24887
24888 PR target/56186
24889 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24890 argument and improve checking of return-argument types for 16-byte
24891 modes.
24892 (ix86_function_value_1): Add additional valtype argument on call
24893 of function_value_64.
24894 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24895 handling infunction_value_64 function.
24896
2b5987b5
MGD		 248972013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24898
7ac3af38 24899 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 24900
14597080
RB		 249012013-02-04 Richard Biener <rguenther@suse.de>
24902
24903 PR tree-optimization/56188
24904 * tree-ssa-structalias.c (label_visit): Consider case with
24905 initially non-empty points-to set.
24906 (perform_var_substitution): Dump node mapping and clean up.
24907
ed73881e
RG		 249082013-02-04 Richard Guenther <rguenther@suse.de>
24909
24910 PR lto/56168
24911 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24912 node prevail as last resort.
24913 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 24914 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 24915
139a0707
RB		 249162013-02-04 Richard Biener <rguenther@suse.de>
24917
24918 PR tree-optimization/56113
24919 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24920 Merge into ...
24921 (equiv_class_lookup_or_add): ... this.
24922 (label_visit): Adjust and fix error in previous patch.
24923 (perform_var_substitution): Adjust.
24924
50fe8924
OE		 249252013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24926
24927 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24928 * config/sh/sh4-300.md: Likewise.
24929 * config/sh/sh4a.md: Likewise.
24930 * config/sh/constraints.md: Likewise.
24931 * config/sh/sh.md: Likewise.
24932 * config/sh/netbsd-elf.h: Likewise.
24933 * config/sh/predicates.md: Likewise.
24934 * config/sh/sh-protos.h: Likewise.
24935 * config/sh/ushmedia.h: Likewise.
24936 * config/sh/linux.h: Likewise.
24937 * config/sh/sh.c: Likewise.
24938 * config/sh/superh.h: Likewise.
24939 * config/sh/elf.h: Likewise.
24940 * config/sh/sh4.md: Likewise.
24941 * config/sh/sh.h: Likewise.
24942
1a04ac2b
JDA		 249432013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24944
24945 * config/pa/constraints.md: Adjust unused letters. Change "T"
24946 constraint to match_test floating_point_store_memory_operand().
24947 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24948 (base14_operand): New.
24949 (floating_point_store_memory_operand): New.
24950 (integer_store_memory_operand): Revise to use base14_operand and
24951 reg_plus_base_memory_operand.
24952 (move_dest_operand): Allow symbolic_memory_operands.
24953 (symbolic_memory_operand): Check for LO_SOM.
24954 (symbolic_operand): Change default case to break.
24955 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24956 CONST_DOUBLE values to be reloaded by putting them into memory when
24957 the destination is a floating point register.
24958 (movdf): Remove code to handle CONST_DOUBLE.
24959 (movsf): Likewise.
24960 (reload_indf_r1): New.
24961 (reload_insf_r1): New.
24962 Consistently use "Q" and "T" constraints with integer and floating
24963 point move instructions, respectively.
24964 (movdi): Remove FAIL.
24965 Change predicate for source operand unamed DImode move from
24966 general_operand to move_src_operand.
24967 (umulsidi3): Change predicate for destination operand to
24968 register_operand.
24969 Likewise for similar unamed patterns.
24970 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24971 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24972 (hppa_legitimize_address): Simplify mask calculation.
24973 (pa_emit_move_sequence): Revised handling of secondary reloads from
24974 REG+D addresses for floating point loads and stores. Directly handle
24975 loading CONST0_RTX (mode) to a floating point register.
24976 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24977 to floating point registers. Don't restrict secondary reloads to
24978 floating point registers to integer modes. Revise some comments and
24979 cleanup some code.
24980 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24981 (pa_legitimate_address_p): New.
24982 (pa_legitimize_reload_address): New.
24983 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24984 (STRICT_REG_OK_FOR_BASE_P): New.
24985 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24986 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24987
3bf65591
DE		 249882013-02-03 David Edelsohn <dje.gcc@gmail.com>
24989 Andrew Dixie <andrewd@gentrack.com>
24990
24991 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24992 flag set.
24993
c0a8a3e6
RS		 249942013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24995
24996 * expmed.c (extract_bit_field_1): Pass the full width of the
24997 structure to get_best_reg_extraction_insn.
24998
99113dff
DE		 249992013-02-01 David Edelsohn <dje.gcc@gmail.com>
25000
25001 PR target/54601
25002 * configure.ac (use_cxa_atexit): Add AIX.
25003 * configure: Regenerate.
25004
25005 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25006
ddd84654
JJ		 250072013-02-01 Jakub Jelinek <jakub@redhat.com>
25008
25009 PR debug/54793
25010 * final.c (need_profile_function): New variable.
25011 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25012 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25013 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25014 notes, targetm.asm_out.function_prologue doesn't emit anything,
25015 HAVE_prologue and profiler should be emitted before prologue,
25016 set need_profile_function instead of emitting it.
25017 (final_scan_insn): If need_profile_function, emit
25018 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25019 NOTE_INSN_FUNCTION_BEG note.
25020
250212013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 25022
25023 * config/rs6000/rs6000.md (smulditi3): New.
25024 (umulditi3): New.
25025
ff2a9d88
RH		 25026 * config/alpha/alpha.md (umulditi3): New.
25027
14d52b90
DE		 250282013-02-01 David Edelsohn <dje.gcc@gmail.com>
25029
25030 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25031 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25032
8c7ca45c
RB		 250332013-02-01 Richard Biener <rguenther@suse.de>
25034
25035 PR tree-optimization/56113
25036 * tree-ssa-structalias.c (label_visit): Reduce work for
25037 single-predecessor nodes.
25038
9f419393
EB		 250392013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25040
25041 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25042 range isn't testing for zero.
25043
c5f4be84
SB		 250442013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25045
25046 PR middle-end/56113
25047 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25048
dbdbd982
NC		 250492013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25050 Nick Clifton <nickc@redhat.com>
25051
25052 * config/v850/constraints.md (Q): Define as a memory constraint.
25053 * config/v850/predicates.md (label_ref_operand): New predicate.
25054 (e3v5_shift_operand): New predicate.
25055 (ior_operator): New predicate.
25056 * config/v850/t-v850: Add e3v5 multilib.
25057 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25058 (v850_gen_movdi): Prototype.
25059 * config/v850/v850.c: Add support for e3v5 architecture.
25060 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25061 TARGET_V850E_UP.
25062 (construct_save_jarl): Add e3v5 long JARL support.
25063 (v850_adjust_insn_length): New function. Adjust length of call
25064 insns when using e3v5 instructions.
25065 (v850_gen_movdi): New function: Generate instructions to move a
25066 DImode value.
25067 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25068 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25069 (TARGET_USE_FPU): Enable for e3v5.
25070 (CONST_OK_FOR_W): New macro.
25071 (ADJUST_INSN_LENGTH): Define.
25072 * config/v850/v850.md (UNSPEC_LOOP): Define.
25073 (attr cpu): Add v850e3v5.
25074 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25075 (movdi): New pattern.
25076 (movdi_internal): New pattern.
25077 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25078 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25079 (cstoresf4): Likewise.
25080 (cstoredf4): Likewise.
25081 (insv): New pattern.
25082 (rotlso3_a): New pattern.
25083 (rotlsi3_b): New pattern
25084 (rotlsi3_v850e3v5): New pattern.
25085 (doloop_begin): New pattern.
25086 (fix_loop_counter): New pattern.
25087 (doloop_end): New pattern.
25088 (branch_normal): Add e3v5 long branch support.
25089 (branch_invert): Likewise.
25090 (branch_z_normal): Likewise.
25091 (branch_z_invert): Likewise.
25092 (branch_nz_normal): Likewise.
25093 (branch_nz_invert): Likewise.
25094 (call_internal_short): Add e3v5 register-indirect JARL support.
25095 (call_internal_long): Likewise.
25096 (call_value_internal_short): Likewise.
25097 (call_value_internal_long): Likewise.
25098 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25099 (mloop): New option.
25100 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 25101 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 25102
73861a41
PK		 251032013-01-31 Paul Koning <ni1d@arrl.net>
25104
25105 PR debug/55059
25106 PR debug/54508
25107 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25108 children if parent is a class.
25109 (prune_unused_types_prune): Don't add DW_AT_declaration.
25110
e44978dc
RB		 251112013-01-31 Richard Biener <rguenther@suse.de>
25112
25113 PR tree-optimization/56157
25114 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25115 match up operand with SLP child.
25116
90b10dec
JM		 251172013-01-31 Jason Merrill <jason@redhat.com>
25118
04d2dadd 25119 PR debug/54410
90b10dec
JM		 25120 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25121 parameters the first time.
25122 (gen_scheduled_generic_parms_dies): Check completeness here.
25123
6e616110
RB		 251242013-01-31 Richard Biener <rguenther@suse.de>
25125
25126 PR middle-end/53073
25127 * common.opt (faggressive-loop-optimizations): New flag,
25128 enabled by default.
25129 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25130 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25131 infer_loop_bounds_from_undefined by it.
25132
636f59cf
RB		 251332013-01-31 Richard Biener <rguenther@suse.de>
25134
25135 PR tree-optimization/56150
25136 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25137 visit virtual operands.
25138 (find_uses_to_rename_bb): Likewise.
25139
251402013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 25141
25142 PR tree-optimization/56150
25143 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25144 mixed store non-store stmts.
25145
32887460
JJ		 251462013-01-30 Jakub Jelinek <jakub@redhat.com>
25147
e60e09a0
JJ		 25148 PR sanitizer/55374
25149 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25150 LIBASAN_EARLY_SPEC is defined.
25151 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25152 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25153 before %o.
25154 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25155
32887460
JJ		 25156 PR c++/55742
25157 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25158 invalid args instead of ICEing on it.
25159 (ix86_valid_target_attribute_tree): Return error_mark_node if
25160 ix86_valid_target_attribute_inner_p failed.
25161 (ix86_valid_target_attribute_p): Return false only if
25162 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25163 target("default") attribute.
25164 (sorted_attr_string): Change argument from const char * to tree,
25165 merge in all target attribute arguments rather than just one.
25166 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25167 instead of free. Avoid using strcat.
25168 (ix86_mangle_function_version_assembler_name): Mangle
25169 target("default") as if no target attribute is present. Adjust
25170 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25171 instead of xmalloc and XDELETEVEC instead of free.
25172 (ix86_function_versions): Don't return true if one of the decls
25173 doesn't have target attribute. If they don't and one of the decls
25174 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25175 sorted_attr_string caller. Use XDELETEVEC instead of free.
25176 (ix86_supports_function_versions): Remove.
25177 (make_name): Fix up formatting.
25178 (make_dispatcher_decl): Remove resolver_name and its initialization.
25179 Avoid leaking memory.
25180 (is_function_default_version): Return true if there is
25181 target("default") attribute rather than no target attribute at all.
25182 (make_resolver_func): Avoid leaking memory.
25183 (ix86_generate_version_dispatcher_body): Likewise.
25184 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25185 * target.def (supports_function_versions): Remove.
25186 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25187 * doc/tm.texi: Regenerated.
25188
73cca0cc
VM		 251892013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25190
25191 PR rtl-optimization/56144
25192 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25193 for values with side effects.
25194
11452e7b
RB		 251952013-01-30 Richard Biener <rguenther@suse.de>
25196
25197 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25198 (sparseset_pop): Likewise.
25199 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25200 to be able to use quick_push in the worker loop.
25201
01cb1ef5
MP		 252022013-01-30 Marek Polacek <polacek@redhat.com>
25203
25204 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25205
e892936e
RB		 252062013-01-30 Richard Biener <rguenther@suse.de>
25207
25208 PR lto/56147
6edc3e32 25209 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 25210
cc06c01d
GJL		 252112013-01-30 Georg-Johann Lay <avr@gjlay.de>
25212
25213 PR tree-optimization/56064
25214 * fixed-value.c (fixed_from_double_int): New function.
25215 * fixed-value.h (fixed_from_double_int): New prototype.
25216 (const_fixed_from_double_int): New static inline function.
25217 * fold-const.c (native_interpret_fixed): New static function.
25218 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25219 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25220 (native_encode_fixed): New static function.
25221 (native_encode_expr) <FIXED_CST>: Use it.
25222 (native_interpret_int): Move double_int worker code to...
25223 * double-int.c (double_int::from_buffer): ...this new static method.
25224 * double-int.h (double_int::from_buffer): Prototype it.
25225
d394a308
RB		 252262013-01-30 Richard Biener <rguenther@suse.de>
25227
25228 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25229 New pointer-map and obstack.
25230 (init_alias_vars): Allocate pointer-map and obstack.
25231 (delete_points_to_sets): Free them.
25232 (find_what_var_points_to): Cache result.
25233 (find_what_p_points_to): Adjust for changed interface of
25234 find_what_var_points_to.
25235 (compute_points_to_sets): Likewise.
25236 (ipa_pta_execute): Likewise.
25237
20804d96
RO		 252382013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25239
25240 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25241 * configure: Regenerate.
25242 * config.in: Regenerate.
25243 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25244 #nobits/#progbits if supported.
25245
a7ad88a2
OE		 252462013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25247
25248 PR target/56121
25249 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25250 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25251 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25252
7af79f92
GY		 252532013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25254
91bfca59
OE		 25255 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25256 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 25257
753bcf7b
GY		 252582013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25259
91bfca59
OE		 25260 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25261 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 25262
8cbc2ea8
GY		 252632013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25264
91bfca59
OE		 25265 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25266 declaration.
25267 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25268 * config/arm/cortex-a7.md: New bypasses using
25269 arm_mac_accumulator_is_result.
8cbc2ea8 25270
697a3325
GY		 252712013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25272
91bfca59 25273 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 25274 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 25275 (cortex_a7_fpfmad): New reservation.
25276 (cortex_a7_fpmacs): Use ffmas and update required units.
25277 (cortex_a7_fpmuld): Update required units and latency.
25278 (cortex_a7_fpmacd): Likewise.
25279 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25280 (cortex_a7_neon). Likewise.
25281 (bypass) Update participating units.
697a3325 25282
29637783
GY		 252832013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25284
91bfca59
OE		 25285 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25286 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25287 from fmac to ffma.
25288 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25289 (vfp_fmul): Use ffmad.
25290 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25291 (cortex_r4_fmacd): Use ffmad.
25292 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25293 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25294 (cortex_a9_fmacd): Use ffmad.
25295 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25296 (cortex_a8_vfp_macd): Use ffmad.
25297 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25298 (cortex_a5_fpmacd): Use ffmad.
25299 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25300 (cortex_a15_vfp_macd): Use ffmad.
25301 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 25302
2ee8a2d5
JM		 253032013-01-29 Jason Merrill <jason@redhat.com>
25304
25305 PR libstdc++/54314
25306 * varasm.c (default_assemble_visibility): Don't warn about
25307 visibility on artificial decls.
25308
78d087bc
RB		 253092013-01-29 Richard Biener <rguenther@suse.de>
25310
25311 PR tree-optimization/56113
25312 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25313 the bitmap leader.
25314 (label_visit): Free duplicate bitmaps and record the leader instead.
25315 (perform_var_substitution): Adjust.
25316
83ba4d6f
RB		 253172013-01-29 Richard Biener <rguenther@suse.de>
25318
25319 PR tree-optimization/55270
25320 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25321 the CFG, schedule loops for fixup.
25322
66dfe13f
NC		 253232013-01-29 Nick Clifton <nickc@redhat.com>
25324
25325 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25326 SP_REG.
25327
5a579c3b
LE		 253282013-01-28 Leif Ekblad <leif@rdos.net>
25329
25330 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25331 * config/i386/i386.h (TARGET_RDOS): New macro.
25332 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25333 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25334 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25335 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25336 DEFAULT_LARGE_SECTION_THRESHOLD.
25337 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25338 * config/i386/rdos.h: New file.
25339 * config/i386/rdos64.h: New file.
25340
51e44392
BS		 253412013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25342
25343 PR other/54814
25344 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25345 TEST_HARD_REG_BIT.
25346
db1fb332
JJ		 253472013-01-28 Jakub Jelinek <jakub@redhat.com>
25348
25349 PR rtl-optimization/56117
25350 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25351 call cselib_lookup_from_insn on the MEM before calling
25352 add_insn_mem_dependence.
25353
16917761
RB		 253542013-01-28 Richard Biener <rguenther@suse.de>
25355
25356 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25357 to a stmt that didn't have one.
25358 (copy_phis_for_bb): Likewise for PHI arguments.
25359 (copy_debug_stmt): Likewise for debug stmts.
25360
b9fc0497
RB		 253612013-01-28 Richard Biener <rguenther@suse.de>
25362
25363 PR tree-optimization/56034
6edc3e32 25364 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 25365 (partition_builtin_p): Adjust.
25366 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25367 it is the last partition.
25368 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25369 up the vertex for the definition.
25370 (classify_partition): Classify whether a partition is a
25371 PKIND_REDUCTION, thus has uses outside of the loop.
25372 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25373 Merge all PKIND_REDUCTION partitions into the last partition.
25374 (tree_loop_distribution): Seed partitions from reductions as well.
25375
aa710d25
JJ		 253762013-01-28 Jakub Jelinek <jakub@redhat.com>
25377
0bfbca58
JJ		 25378 PR tree-optimization/56125
25379 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25380 pow(x,c) into sqrt(x) * powi(x, n/2) or
25381 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25382 optimizing for size.
25383 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25384 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25385 integer.
25386
aa710d25
JJ		 25387 PR tree-optimization/56094
25388 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25389 to UNKNOWN_LOCATION while gimplifying expr.
25390
77dc5297
UB		 253912013-01-27 Uros Bizjak <ubizjak@gmail.com>
25392
25393 PR target/56114
25394 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25395 operand 0 in movabs insn template for -masm=intel asm alternative.
25396 (*movabs<mode>_2): Ditto for operand 1.
25397
0bfbca58 253982013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 25399
25400 PR target/54663
25401 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25402 of microblaze-c.o
25403
0bfbca58 254042013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 25405
25406 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25407 tm_file.
25408
0bfbca58 254092013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 25410
25411 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25412 Undef to avoid warning.
25413
478f60f9
MH		 254142013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25415
25416 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25417 * configure: Regenerate.
25418
d7fa6ee2
JJ		 254192013-01-25 Jakub Jelinek <jakub@redhat.com>
25420
25421 PR tree-optimization/56098
25422 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25423 for stmts with volatile ops.
25424 (cond_store_replacement): Don't optimize if assign has volatile ops.
25425 (cond_if_else_store_replacement_1): Don't optimize if either
25426 then_assign or else_assign have volatile ops.
25427 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25428 volatile ops.
25429
f8fe87bd
GJL		 254302013-01-25 Georg-Johann Lay <avr@gjlay.de>
25431
25432 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25433
28398d0d
GJL		 254342013-01-25 Georg-Johann Lay <avr@gjlay.de>
25435
25436 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25437 missing ':' in asm example.
25438
b7d7d917
TB		 254392013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25440
25441 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25442 entries into lane and laneq entries.
77dc5297
UB		 25443 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25444 Remove AdvSIMD scalar modes.
b7d7d917
TB		 25445 (aarch64_sq<r>dmulh_laneq<mode>): New.
25446 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25447 modes.
25448 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25449 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25450 * config/aarch64/iterators.md (VCOND): New.
25451 (VCONQ): New.
25452
556f9906
GJL		 254532013-01-25 Georg-Johann Lay <avr@gjlay.de>
25454
25455 PR target/54222
25456 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25457 Add NULL LIBNAME argument to existing definitions.
25458 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25459 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25460 * config/avr/avr.c (DEF_BUILTIN): Same.
25461 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25462 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25463 implementation is available (DECL_ASSEMBLER_NAME is set).
25464 (avr_fold_absfx): New static function.
25465 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25466 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25467 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25468 AVR_BUILTIN_ABSLLK.
25469 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25470 (abshk, absk, abslk, absllk): Provide as static inline functions.
25471
1f546bbb
MP		 254722013-01-25 Marek Polacek <polacek@redhat.com>
25473
25474 PR tree-optimization/56035
25475 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25476
7e184bd7
UB		 254772012-01-24 Uros Bizjak <ubizjak@gmail.com>
25478
25479 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25480 (*movtf_internal_rex64): Add (!o,C) alternative
25481 (*movxf_internal_rex64): Ditto.
25482 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25483
f5ba49ac
SM		 254842013-01-24 Shenghou Ma <minux.ma@gmail.com>
25485
25486 * doc/invoke.texi: fix typo.
25487 * doc/objc.texi: fix typo.
25488
a6343728
RS		 254892013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25490
25491 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25492 for the first two alternatives.
25493
cd030c07
DN		 254942013-01-24 Diego Novillo <dnovillo@google.com>
25495
77dc5297 25496 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 25497 (ggc-zone.o): Remove.
25498 * configure.ac: Remove option --with-gc.
25499 * configure: Re-generate.
25500 * doc/install.texi: Remove documentation for --with-gc.
25501 * gengtype.c (write_enum_defn): Remove. Update all users.
25502 (write_Types_process_field): Remove generation of gt_e_* argument.
25503 (output_type_enum): Remove. Update all users.
25504 (write_enum_defn): Remove. Update all users.
25505 (enum alloc_zone): Remove. Update all users.
77dc5297 25506 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 25507 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25508 Update all callers.
25509 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25510 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 25511 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25512 Update all users.
aaf1e810 25513 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 25514 (struct alloc_zone): Remove.
25515 (ggc_internal_alloc_zone_stat): Remove.
25516 (ggc_internal_cleared_alloc_zone_stat): Remove.
25517 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 25518 (ggc_pch_count_object): Remove last argument. Update all users.
25519 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 25520 (struct alloc_zone): Remove.
25521 * ggc-zone.c: Remove.
77dc5297 25522 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 25523 (struct alloc_zone): Remove.
25524 (ggc_alloc_typed_stat): Remove.
25525 (ggc_alloc_typed): Remove.
25526 (ggc_splay_alloc): Remove first argument.
25527 (rtl_zone): Remove. Update all users.
25528 (tree_zone): Remove. Update all users.
25529 (tree_id_zone): Remove. Update all users.
25530 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25531 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25532 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25533 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25534
a861ffa4
GJL		 255352013-01-24 Georg-Johann Lay <avr@gjlay.de>
25536
25537 * config/avr/avr.c (avr_out_fract): Make register numbers that
25538 might be outside of source operand signed.
25539
593c0ddd
UB		 255402013-01-24 Uros Bizjak <ubizjak@gmail.com>
25541
25542 * config/i386/constraints.md (Yf): New constraint.
25543 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25544 of f constraint to conditionaly disable x87 register preferences.
25545 (*movdf_internal): Ditto.
25546 (*movsf_internal): Ditto.
25547
e86c0101
SB		 255482013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25549
25550 PR inline-asm/55934
25551 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25552 that have operands with impossible constraints.
25553 Add a FIXME for a speed-up opportunity.
25554 * lra-constraints.c (process_alt_operands): Verify that a class
25555 selected from constraints on asms is valid for the operand mode.
25556 (curr_insn_transform): Remove incorrect comment.
25557
f6fee35f
DE		 255582013-01-23 David Edelsohn <dje.gcc@gmail.com>
25559
25560 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25561 TOC operand is a valid symbol ref in the constant pool.
25562
255632013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 25564
aaf1e810 25565 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 25566
dc62d7d1
GJL		 255672013-01-23 Georg-Johann Lay <avr@gjlay.de>
25568
25569 PR target/54222
25570 * config/avr/stdfix.h: New file.
25571 * t-avr (stdfix-gcc.h): New rule to build it.
25572 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25573
2660d12d
KS		 255742013-01-23 Kostya Serebryany <kcc@google.com>
25575
77dc5297
UB		 25576 * config/darwin.h: remove dependency on
25577 CoreFoundation (asan on Mac OS).
2660d12d 25578
a70418fc
JJ		 255792013-01-23 Jakub Jelinek <jakub@redhat.com>
25580
25581 PR target/49069
25582 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25583 instead of cmpdi_operand for first comparison operand.
25584 Don't assert that comparison operands aren't both constants.
25585
47876a2a
JW		 255862013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25587
25588 * doc/install.texi (Downloading the Source): Update references to
25589 downloading separate components.
25590
255912013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 25592
25593 * doc/extend.texi (__int128): Improve grammar.
25594
47876a2a 255952013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 25596
25597 PR target/56028
25598 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25599 alternative to (o,r).
25600 (*movdi_internal_rex64): Remove (!o,n) alternative.
25601 (DImode immediate->memory splitter): Remove.
25602 (DImode immediate->memory peephole2): Remove.
25603 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25604 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25605 alternative to (!o,*r).
25606 (*movtf_internal_sse): New pattern.
25607 (*movxf_internal_rex64): New pattern.
25608 (*movxf_internal): Disable for TARGET_64BIT.
25609 (*movdf_internal_rex64): Remove (!o,F) alternative.
25610
3a984f10
JJ		 256112013-01-22 Jakub Jelinek <jakub@redhat.com>
25612
502498d5
JJ		 25613 PR middle-end/56074
25614 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25615 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25616 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25617 stmt locations where LOCATION_LOCUS of the stmt location is
25618 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25619
3a984f10
JJ		 25620 PR target/55686
25621 * config/i386/i386.md (UNSPEC_STOS): New.
25622 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25623 *strsetqi_1): Add UNSPEC_STOS.
25624
fa817f7f
PC		 256252013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25626
25627 PR c++/56067
25628 * doc/invoke.texi: Remove left over -Wsynth example.
25629
8f498c1b
JJ		 256302013-01-21 Jakub Jelinek <jakub@redhat.com>
25631
25632 PR tree-optimization/56051
25633 * fold-const.c (fold_binary_loc): Don't fold
25634 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25635 a narrowing conversion, or widening conversion from signed
25636 to unsigned.
25637
47876a2a 256382013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 25639
25640 PR rtl-optimization/56023
25641 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25642 dependent on debug instruction.
25643
5b9db1bc
MJ		 256442013-01-21 Martin Jambor <mjambor@suse.cz>
25645
25646 PR middle-end/56022
25647 * function.c (allocate_struct_function): Call
25648 invoke_set_current_function_hook earlier.
25649
e8bb7d68
JJ		 256502013-01-21 Jakub Jelinek <jakub@redhat.com>
25651
25652 * reload1.c (init_reload): Only initialize reload_obstack
25653 during the first call.
25654
616a4e32
MP		 256552013-01-21 Marek Polacek <polacek@redhat.com>
25656
25657 * cfgloop.c (verify_loop_structure): Fix up grammar.
25658
4401981b
YHH		 256592013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25660
25661 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25662 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25663
8e87740b
RR		 256642013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25665
25666 PR target/56058
25667 * config/arm/marvell-pj4.md: Update copyright year.
25668 Fix up use of alu to alu_reg and simple_alu_imm.
25669
47876a2a 256702013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 25671
25672 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25673
89d56d79
VM		 256742013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25675
25676 PR target/55433
25677 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 25678 insn for secondary memory move when memory mode should be different.
89d56d79 25679
fe603553
JDA		 256802013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25681
25682 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25683 atomic_storedi_1): New patterns.
25684
01284895
VK		 256852013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25686
25687 btver2 pipeline descriptions.
25688 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25689 descriptions.
25690 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 25691 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 25692 type attributes.
25693 * config/i386/btver2.md: New file describing btver2 pipelines.
25694
5630e3e1
JL		 256952013-01-19 Andrew Pinski <apinski@cavium.com>
25696
25697 PR tree-optimization/52631
25698 * tree-ssa-sccvn (visit_use): Before looking up the original
25699 statement, try looking up the simplified expression.
25700
650ae806
AG		 257012013-01-19 Anthony Green <green@moxielogic.com>
25702
25703 * config/moxie/moxie.c (moxie_expand_prologue): Set
25704 current_function_static_stack_size.
25705
e300ec2d
JJ		 257062013-01-18 Jakub Jelinek <jakub@redhat.com>
25707
25708 PR tree-optimization/56029
25709 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25710 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25711
a3d7af04
SS		 257122013-01-18 Sharad Singhai <singhai@google.com>
25713
25714 PR tree-optimization/55995
25715 * dumpfile.c (dump_loc): Print location only if available.
25716 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25717
66aa7879
VM		 257182013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25719
25720 PR target/55433
25721 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25722 secondary memory move.
25723 (inherit_reload_reg): Use rclass instead of cl for
25724 check_secondary_memory_needed_p.
25725
3f0fee7b
JJ		 257262013-01-18 Jakub Jelinek <jakub@redhat.com>
25727
25728 PR middle-end/56015
25729 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 25730 the case where writing real complex part of target modifies op1.
3f0fee7b 25731
70c67693
JG		 257322013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25733
25734 * config/aarch64/aarch64-simd.md
25735 (aarch64_vcond_internal<mode>): Handle unordered cases.
25736 * config/aarch64/iterators.md (v_cmp_result): New.
25737
df8de9b3
YHH		 257382013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25739 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25740
25741 * config/arm/marvell-pj4.md: New file.
25742 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25743 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25744 (generic_vfp): Likewise.
25745 * config/arm/arm-cores.def: Add marvell-pj4.
25746 * config/arm/arm-tune.md: Regenerate.
25747 * config/arm/arm-tables.opt: Regenerate.
25748 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25749 * doc/invoke.texi: Document marvell-pj4.
25750
be30c356
TB		 257512013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25752
25753 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25754
0bfbca58 257552013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 25756
25757 PR debug/54114
25758 PR debug/54402
25759 PR debug/49888
25760 * var-tracking.c (negative_power_of_two_p): New.
25761 (global_get_addr_cache, local_get_addr_cache): New.
25762 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25763 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25764 heading comment.
25765 (vt_stack_offset_p): Remove.
25766 (vt_canon_true_dep): Always canonicalize loc's address.
25767 (clobber_overlapping_mems): Make sure we have a MEM.
25768 (local_get_addr_clear_given_value): New.
25769 (val_reset): Clear local cached entries.
25770 (compute_bb_dataflow): Create and release the local cache.
25771 Disable duplicate MEMs clobbering.
25772 (emit_notes_in_bb): Clobber MEMs likewise.
25773 (vt_emit_notes): Create and release the local cache.
25774 (vt_initialize, vt_finalize): Create and release the global
25775 cache, respectively.
1f6bc337 25776 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 25777
0bfbca58 257782013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 25779
25780 PR libmudflap/53359
25781 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25782 not found in the symtab.
25783
0bfbca58 257842013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 25785
c350ba53 25786 PR debug/56006
3aa03517
AO		 25787 PR rtl-optimization/55547
25788 PR rtl-optimization/53827
25789 PR debug/53671
25790 PR debug/49888
25791 * alias.c (offset_overlap_p): New, factored out of...
25792 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25793 the conservative special case for symbolic constants. Don't
25794 adjust zero sizes on alignment.
25795
c664546f
JL		 257962013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25797
25798 PR rtl-optimization/52573
25799 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25800 REG_UNUSED for the same register.
25801
1bd3f750
MP		 258022013-01-17 Richard Biener <rguenther@suse.de>
25803 Marek Polacek <polacek@redhat.com>
25804
25805 PR rtl-optimization/55833
25806 * loop-unswitch.c (unswitch_loops): Move loop verification...
25807 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25808 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25809 Set it to true when we're removing a loop from hierarchy tree in
25810 an irreducible region.
25811 (fix_bb_placements): Adjust caller.
25812 (fix_loop_placements): Likewise.
25813
e52a8b71
GJL		 258142013-01-17 Georg-Johann Lay <avr@gjlay.de>
25815
25816 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25817 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25818 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25819 Remove ID. Adjust comments.
25820 * config/avr/avr-c.c (avr_builtin_name): Remove.
25821 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25822 * config/avr/avr.c (avr_tolower): New static function.
25823 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25824 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25825 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25826 default expansion.
25827
8386a7ea
JH		 258282013-01-17 Jan Hubicka <jh@suse.cz>
25829
610fb637 25830 PR tree-optimization/55273
8386a7ea
JH		 25831 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25832
47876a2a 258332013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 25834
25835 PR target/55981
25836 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25837 store through atomic_store<mode>_1.
25838 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25839
8222c37e
MJ		 258402013-01-17 Martin Jambor <mjambor@suse.cz>
25841
25842 PR tree-optimizations/55264
25843 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25844 for virtual methods.
25845 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25846 virtual methods before inlining is over.
25847 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25848 virtual functions.
25849 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25850 non-virtual.
25851
79f01c76
VM		 258522013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25853
25854 PR rtl-optimization/56005
25855 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25856 pending reads for prefetch.
25857
d0b6bb1b
IB		 258582013-01-16 Ian Bolton <ian.bolton@arm.com>
25859
aaf1e810 25860 * config/aarch64/aarch64.md
d0b6bb1b
IB		 25861 (*cstoresi_neg_uxtw): New pattern.
25862 (*cmovsi_insn_uxtw): New pattern.
25863 (*<optab>si3_uxtw): New pattern.
25864 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25865 (*<optab>si3_insn_uxtw): New pattern.
25866 (*bswapsi2_uxtw): New pattern.
25867
cb9cf03b
RB		 258682013-01-16 Richard Biener <rguenther@suse.de>
25869
25870 * tree-inline.c (tree_function_versioning): Remove set but
25871 never used variable.
25872
2cfc56b9
RB		 258732013-01-16 Richard Biener <rguenther@suse.de>
25874
25875 PR tree-optimization/55964
25876 * tree-flow.h (rename_variables_in_loop): Remove.
25877 (rename_variables_in_bb): Likewise.
25878 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25879 (copy_loop_before): Adjust and delete update-ssa status.
25880 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 25881 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 25882 (rename_variables_in_loop): Remove.
25883 (slpeel_update_phis_for_duplicate_loop): Likewise.
25884 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25885 use available cfg machinery instead of duplicating it.
25886 Update PHI nodes and perform poor-mans SSA update here.
25887 (slpeel_tree_peel_loop_to_edge): Adjust.
25888
c25a0c60
RB		 258892013-01-16 Richard Biener <rguenther@suse.de>
25890
25891 PR tree-optimization/54767
25892 PR tree-optimization/53465
25893 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25894 (vrp_visit_phi_node): For PHI arguments coming via backedges
25895 drop all symbolical range information.
25896 (execute_vrp): Compute backedges.
25897
04b535af
RB		 258982013-01-16 Richard Biener <rguenther@suse.de>
25899
25900 * doc/install.texi: Update CLooG and ISL requirements to
25901 0.18.0 and 0.11.1.
25902
8b0a1e0b
CB		 259032013-01-16 Christian Bruel <christian.bruel@st.com>
25904
25905 PR target/55301
25906 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25907 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25908 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25909
259102013-01-16 DJ Delorie <dj@redhat.com>
25911
25912 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25913 (UNSPECV_SP_SWITCH_E): New.
25914 (sp_switch_1): Change to an unspec.
25915 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25916 replace $r15.
25917
259182013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 25919
25920 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25921 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25922 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25923 (expand_mem_thread_fence): Ditto.
25924 (expand_mem_signal_fence): Ditto.
25925 (expand_atomic_load): Ditto.
25926 (expand_atomic_store): Ditto.
25927
0bfbca58 259282013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 25929
25930 PR rtl-optimization/55547
25931 PR rtl-optimization/53827
25932 PR debug/53671
25933 PR debug/49888
25934 * alias.c (memrefs_conflict_p): Set sizes to negative after
25935 AND adjustments.
25936
305e3ac1
JJ		 259372013-01-15 Jakub Jelinek <jakub@redhat.com>
25938
25939 PR target/55940
25940 * function.c (thread_prologue_and_epilogue_insns): Always
25941 add crtl->drap_reg to set_up_by_prologue.set, even if
25942 stack_realign_drap is false.
25943
f78ac4f2
JBG		 259442013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25945
25946 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25947 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25948 *call): Fix indention.
25949
a78a8cc4
TV		 259502013-01-15 Tom de Vries <tom@codesourcery.com>
25951
25952 PR target/55876
25953 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25954 Update comment.
25955
0e80383f
VM		 259562013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25957
305e3ac1 25958 PR rtl-optimization/55153
0e80383f
VM		 25959 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25960
207b5956
MJ		 259612013-01-15 Martin Jambor <mjambor@suse.cz>
25962
25963 PR tree-optimization/55920
25964 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25965 accesses as grp_to_be_debug_replaced.
25966
a7818b54
JJ		 259672013-01-15 Jakub Jelinek <jakub@redhat.com>
25968
25969 PR tree-optimization/55920
25970 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25971 there is non-useless type conversion needed from debug rhs to lhs,
25972 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25973
b0fe107e
JM		 259742013-01-15 Joseph Myers <joseph@codesourcery.com>
25975 Mikael Pettersson <mikpe@it.uu.se>
25976
25977 PR target/43961
25978 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25979 Thumb.
25980 (ASM_OUTPUT_CASE_LABEL): Remove.
25981 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25982 * final.c (shorten_branches): Update alignment of labels before
25983 jump tables if CASE_VECTOR_SHORTEN_MODE.
25984
34ab62ee
RB		 259852013-01-15 Richard Biener <rguenther@suse.de>
25986
25987 PR bootstrap/55961
25988 * system.h: Do not include gmp.h for building host tools.
25989
783a3a05
RB		 259902013-01-15 Richard Biener <rguenther@suse.de>
25991
25992 PR middle-end/55882
25993 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25994 account for bitpos when computing alignment.
25995
3a579e09
VY		 259962013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25997
25998 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25999 (ix86_target_macros_internal): Likewise.
26000
26001 * config/i386/i386.c (m_CORE2I7): Removed.
26002 (m_CORE_HASWELL): New macro.
26003 (m_CORE_ALL): Likewise.
26004 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26005 (initial_ix86_arch_features): Likewise.
26006 (processor_target_table): Initializations for Core avx2.
26007 (cpu_names): New names "core-avx2".
26008 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26009 PROCESSOR_CORE_HASWELL.
26010 (ix86_issue_rate): New case.
26011 (ia32_multipass_dfa_lookahead): Likewise.
26012 (ix86_sched_init_global): Likewise.
26013
26014 * config/i386/i386.h (TARGET_HASWELL): New macro.
26015 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26016 (processor_type): New PROCESSOR_HASWELL.
26017
955f5a07
JJ		 260182013-01-15 Jakub Jelinek <jakub@redhat.com>
26019
ff784829
JJ		 26020 PR tree-optimization/55955
26021 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26022 *SHIFT_EXPR and *ROTATE_EXPR codes.
26023
955f5a07
JJ		 26024 PR tree-optimization/48766
26025 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26026 -ftrapv disable -fwrapv.
26027
aeb8b4e9
GJL		 260282013-01-14 Georg-Johann Lay <avr@gjlay.de>
26029
26030 PR target/55974
26031 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26032 etc. to 1 and not to __flash.
26033 Use LL suffix for __INT24_MAX__ with -mint8.
26034 Use ULL suffix for __UINT24_MAX__ with -mint8.
26035
1c494c6a
GJL		 260362013-01-14 Georg-Johann Lay <avr@gjlay.de>
26037
26038 * config/avr/avr-arch.h
26039 (struct base_arch_s): Use typedef avr_arch_t instead.
26040 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26041 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26042 * config/avr/avr.c: Same.
26043 * config/avr/avr-devices.c: Same.
26044 * config/avr/driver-avr.c: Same.
26045 * config/avr/gen-avr-mmcu-texi.c: Same.
26046 * config/avr/avr-mcus.def: Adjust comment.
26047
a50344cb
TB		 260482013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26049
88e784e6
UB		 26050 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26051 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 26052
47876a2a 260532013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 26054 Andi Kleen <ak@linux.intel.com>
26055
26056 PR target/55948
26057 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26058 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26059 memmodel flag.
26060
00892272
GJL		 260612013-01-14 Georg-Johann Lay <avr@gjlay.de>
26062
26063 * config/avr/avr-stdint.h: Remove trailing blanks.
26064 * config/avr/avr-log.h: Same.
26065 * config/avr/avr-arch.h: Same.
26066 * config/avr/avr-devices.c: Same.
26067 * config/avr/avr-dimode.md: Same.
26068 * config/avr/predicates.md: Same.
26069 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 26070
00892272
GJL		 26071 * config/avr/avr-protos.h: Same. And:
26072 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26073 (init_cumulative_args): Rename to avr_init_cumulative_args.
26074 (expand_prologue): Rename to avr_expand_prologue.
26075 (expand_epilogue): Rename to avr_expand_epilogue.
26076 (adjust_insn_length): Rename to avr_adjust_insn_length.
26077 (notice_update_cc): Rename to avr_notice_update_cc.
26078 (final_prescan_insn): Rename to avr_final_prescan_insn.
26079 * config/avr/avr.c: Same.
26080 * config/avr/avr.h: Same.
26081 * config/avr/avr.md: Remove trailing blanks.
26082 (prologue): Use avr_expand_prologue.
26083 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26084
7bb01996
RB		 260852013-01-14 Richard Biener <rguenther@suse.de>
26086
26087 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26088 verify_location, collect_subblocks): New functions.
26089 (verify_gimple_in_cfg): Verify that locations only reference
26090 BLOCKs in the functions BLOCK tree.
26091
2724573f
RB		 260922013-01-14 Richard Biener <rguenther@suse.de>
26093
26094 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26095 PHI argument.
26096 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26097 unshare reference.
26098 (insert_out_of_ssa_copy_on_edge): Likewise.
26099 (rewrite_close_phi_out_of_ssa): Likewise.
26100 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26101 debug expressions.
26102 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26103 propagated constants.
26104 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26105 can not be shared.
26106
9a0bbab6
GJL		 261072013-01-14 Georg-Johann Lay <avr@gjlay.de>
26108
26109 * config/avr/avr-modes.def: Add GPL copyright notice.
26110
45805f17
UB		 261112013-01-13 Uros Bizjak <ubizjak@gmail.com>
26112
26113 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26114 MEMMODEL_MASK to determine memory model.
26115 (atomic_store<mode>): Ditto from operands[2].
26116 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26117
9d60be38
JJ		 261182013-01-13 Jakub Jelinek <jakub@redhat.com>
26119
26120 PR fortran/55935
45805f17 26121 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 26122 (fold_gimple_assign): Don't call unshare_expr here.
26123 (fold_ctor_reference): Call unshare_expr.
26124
e7f49d92
TG		 261252013-01-13 Terry Guo <terry.guo@arm.com>
26126
aaf1e810
EB		 26127 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26128 * doc/fragments.texi: Document MULTILIB_REUSE.
26129 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 26130 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 26131 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 26132 (tmpmultilib4): Ditto.
26133 (multilib_reuse): New multilib argument.
26134
fbd03a27
RS		 261352013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26136
26137 * Makefile.in: Update copyright.
26138
b3681f13
TV		 261392013-01-12 Tom de Vries <tom@codesourcery.com>
26140
26141 PR middle-end/55890
26142 * calls.c (expand_call): Check if arg_nr is valid.
26143
3f287e4b
MM		 261442013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26145
26146 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26147 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 26148 documentation. Add missing '__' in front of
26149 __builtin_ia32_packssdw256.
3f287e4b 26150
1abcd5eb
AK		 261512013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26152
26153 PR target/55719
26154 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26155 NO_REGS for larl operands.
26156 (s390_reload_larl_operand): Use s390_load_address instead of
26157 emit_move_insn.
26158
980d0812
RB		 261592013-01-11 Richard Biener <rguenther@suse.de>
26160
26161 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26162 (verify_node_sharing): ... here.
26163 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26164
7cb9fd07
EB		 261652013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26166
26167 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26168 Substitute TREECHECKING.
26169 * configure: Regenerate.
26170 * Makefile.in (TREECHECKING): New.
26171
47876a2a 261722013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 26173
26174 PR tree-optimization/44061
26175 * tree-vrp.c (extract_range_basic): Compute zero as
26176 value-range for __builtin_constant_p of function parameters.
26177
d1e082c2
RS		 261782013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26179
45805f17 26180 Update copyright years.
d1e082c2 26181
f85021b0
VM		 261822013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26183
9d60be38 26184 PR rtl-optimization/55672
0160442c
VM		 26185 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26186 const to be eliminable.
f85021b0 26187
7a8b1ec4
DE		 261882013-01-10 David Edelsohn <dje.gcc@gmail.com>
26189
26190 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26191 * configure: Regenerate.
26192
ad2e5b71
RB		 261932013-01-10 Richard Biener <rguenther@suse.de>
26194
ddf9322d 26195 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 26196
ffc5b2bb
RB		 261972013-01-10 Richard Biener <rguenther@suse.de>
26198
26199 PR bootstrap/55792
26200 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26201 locations for virtual PHI arguments.
26202 (rewrite_update_phi_arguments): Likewise.
26203
e1f674e4
JS		 262042013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26205
26206 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26207 on to assembler.
26208
a57fc743
JJ		 262092013-01-10 Jakub Jelinek <jakub@redhat.com>
26210
26211 PR tree-optimization/55921
26212 * tree-complex.c (expand_complex_asm): New function.
26213 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26214
0ff4390d
AK		 262152013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26216
26217 PR target/55718
26218 * config/s390/s390.c (s390_symref_operand_p)
26219 (s390_loadrelative_operand_p): Merge the two functions.
26220 (s390_check_qrst_address, print_operand_address): Add parameters
26221 to s390_loadrelative_operand_p invokation.
26222 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26223 (s390_reload_larl_operand, s390_secondary_reload): Use
26224 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26225 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26226
6f557e0e
MS		 262272013-01-09 Mike Stump <mikestump@comcast.net>
26228
26229 * dse.c (record_store): Remove unnecessary assert.
26230
7770c9e9
JH		 262312013-01-09 Jan Hubicka <jh@suse.cz>
26232
26233 PR tree-optimization/55569
26234 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26235 * cfgloop.h (scale_loop_profile): Likewise.
26236
a19b1432
JH		 262372013-01-09 Jan Hubicka <jh@suse.cz>
26238
26239 PR lto/45375
ddf9322d
UB		 26240 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26241 functions.
a19b1432
JH		 26242 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26243
9a002da8
RS		 262442013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26245
26246 PR middle-end/55114
26247 * expr.h (maybe_emit_group_store): Declare.
26248 * expr.c (maybe_emit_group_store): New function.
26249 * builtins.c (expand_builtin_int_roundingfn): Call it.
26250 (expand_builtin_int_roundingfn_2): Likewise.
26251
511dcace
VM		 262522013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26253
e1f2b729 26254 PR rtl-optimization/55829
511dcace
VM		 26255 * lra-constraints.c (match_reload): Add code for absent output.
26256 (curr_insn_transform): Add code for reloads of matched inputs
26257 without output.
26258
7b0fe4f4
UB		 262592013-01-09 Uros Bizjak <ubizjak@gmail.com>
26260
26261 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26262 attribute of movddup insn to DF.
26263 (*vec_interleave_lowv2df): Ditto.
26264 (vec_dupv2df): Ditto.
26265
870ca331
JH		 262662013-01-09 Jan Hubicka <jh@suse.cz>
26267
26268 PR tree-optimiation/55875
26269 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26270 EVERY_ITERATION parameter.
7b0fe4f4 26271 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 26272 (idx_infer_loop_bounds): Similarly here.
26273 (n_of_executions_at_most): Simplify
26274 to only test for cases where statement is dominated by the
7b0fe4f4 26275 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 26276 (scev_probably_wraps_p): Use max loop iterations info
26277 as a global bound first.
26278
6f575fe4 262792013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 26280 Nick Clifton <nickc@redhat.com>
26281
26282 * config/v850/v850.md (cbranchsf4): New pattern.
26283 (cstoresf4): New pattern.
26284 (cbranchdf4): New pattern.
26285 (cstoredf4): New pattern.
26286 (movsicc): Disallow floating point comparisons.
26287 (cmpsf_le_insn): Fix order of operators.
26288 (cmpsf_lt_insn): Likewise.
26289 (cmpsf_eq_insn): Likewise.
26290 (cmpdf_le_insn): Likewise.
26291 (cmpdf_lt_insn): Likewise.
26292 (cmpdf_eq_insn): Likewise.
26293 (cmpsf_ge_insn): Use LE comparison.
26294 (cmpdf_ge_insn): Likewise.
26295 (cmpsf_gt_insn): Use LT comparison.
26296 (cmpdf_gt_insn): Likewise.
26297 (cmpsf_ne_insn): Delete pattern.
26298 (cmpdf_ne_insn): Delete pattern.
26299 * config/v850/v850.c (v850_gen_float_compare): Use
26300 gen_cmpdf_eq_insn for NE comparison.
26301 (v850_float_z_comparison_operator)
26302 (v850_float_nz_comparison_operator): Move from here ...
26303 * config/v850/predicates.md: ... to here. Move GT and GE
26304 comparisons into v850_float_z_comparison_operator.
26305 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26306 Delete prototype.
26307 (v850_float_nz_comparison_operator): Likewise.
26308
f0d54148
JDA		 263092013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26310
26311 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26312 with calls to gen_insvsi/gen_insvdi.
26313
8f01beca
VK		 263142013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26315
26316 * config/i386/i386.c (initial_ix86_tune_features): Set up
26317 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26318
2aa579ad
JJ		 263192013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26320 Jakub Jelinek <jakub@redhat.com>
26321
26322 PR tree-optimization/48189
26323 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26324 If nitercst is 0, don't predict the exit edge.
26325
6edc3e32 263262013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 26327
26328 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26329 in asm_fprintf with reg_names.
26330 (aarch64_print_operand_address): Likewise.
26331 (aarch64_return_addr): Likewise.
26332 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26333
f6f94d94
JDA		 263342013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26335
26336 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26337 (INT_U6_BITS): Likewise.
26338 * config/pa/predicates.md (uint6_operand): New predicate.
26339 (shift5_operand, shift6_operand): Likewise.
26340 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26341 arith32_operand.
26342 (lshrdi3): Use shift6_operand.
26343 (shrpsi4, shrpdi4): New insn patterns.
26344 (extzv): Delete expander.
26345 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26346 predicates in unamed zero extract patterns. Tighten common constraint.
26347 (extv): Delete expander.
26348 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26349 predicates in unamed sign extract patterns. Tighten common constraint.
26350 (insv): Delete expander.
26351 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26352 predicates in unamed insert patterns. Tighten common constraint.
26353 Change uint32_operand predicate to uint6_operand predicate in unamed
26354 DImode pattern to insert constant values of type 1...1xxxx.
26355
36b72910
JH		 263562013-01-04 Jan Hubicka <jh@suse.cz>
26357
26358 PR tree-optimization/55823
7b0fe4f4
UB		 26359 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26360 issue.
36b72910 26361
47876a2a 263622013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 26363 Uros Bizjak <ubizjak@gmail.com>
26364
26365 PR rtl-optimization/55845
26366 * df-problems.c (can_move_insns_across): Stop scanning at
26367 volatile_insn_p source instruction or give up if
26368 across_from .. across_to range contains any volatile_insn_p
26369 instructions.
26370
4369c11e
TB		 263712013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26372
7b0fe4f4
UB		 26373 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26374 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26375 Declare.
4369c11e 26376 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26377 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26378
aeb7e7c1
JJ		 263792013-01-08 Jakub Jelinek <jakub@redhat.com>
26380
26381 PR fortran/55341
26382 * asan.c (asan_clear_shadow): New function.
26383 (asan_emit_stack_protection): Use it.
26384
a02ad1aa
TB		 263852013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26386
26387 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26388 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26389 with tab instead of space.
26390
f90d87f5
NC		 263912013-01-08 Nick Clifton <nickc@redhat.com>
26392
26393 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26394 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 26395 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26396 MDBH registers.
f90d87f5 26397
385eb93d
JG		 263982013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26399
26400 * config/aarch64/aarch64-simd.md
26401 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26402 (aarch64_simd_bsl): Likewise.
26403 (aarch64_vcond_internal<mode>): Likewise.
26404 (vcond<mode><mode>): Likewise.
26405 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26406 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26407
4dcd1054
JG		 264082013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26409
26410 * config/aarch64/aarch64-builtins.c
26411 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26412
4267a4a6
MJ		 264132013-01-08 Martin Jambor <mjambor@suse.cz>
26414
26415 PR debug/55579
26416 * tree-sra.c (analyze_access_subtree): Return true also after
26417 potentially creating a debug-only replacement.
26418
5f4e6de3
JJ		 264192013-01-08 Jakub Jelinek <jakub@redhat.com>
26420
3138f224
JJ		 26421 PR middle-end/55890
26422 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26423
5f4e6de3
JJ		 26424 PR tree-optimization/54120
26425 * tree-vrp.c (range_fits_type_p): Don't allow
26426 src_precision < precision from signed vr to unsigned_p
26427 if vr->min or vr->max is negative.
26428 (simplify_float_conversion_using_ranges): Test can_float_p
26429 against CODE_FOR_nothing.
26430
f3ef18ff
JJ		 264312013-01-08 Jakub Jelinek <jakub@redhat.com>
26432 Richard Biener <rguenther@suse.de>
26433
26434 PR middle-end/55851
26435 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26436 types instead of just INTEGER_TYPE types.
26437
25c210f9
MK		 264382013-01-07 Mark Kettenis <kettenis@openbsd.org>
26439
26440 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26441 TF_SIZE): Define.
f90d87f5 26442
81826a7b
SE		 264432013-01-07 Steve Ellcey <sellcey@mips.com>
26444
26445 PR target/42661
26446 * config/mips/mips.opt: Change mad to mmad to match documentation.
26447
35678514
GJL		 264482013-01-07 Georg-Johann Lay <avr@gjlay.de>
26449
26450 PR target/55897
26451 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26452 .progmemx.data now.
26453
e5669488
GJL		 264542013-01-07 Georg-Johann Lay <avr@gjlay.de>
26455
26456 PR target/55897
26457 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26458 (avr_addrspace_t): Add .section_name field.
26459 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26460 array size.
26461 (avr_addrspace): Same. Initialize .section_name. Remove last
26462 NULL entry. Put __memx into .progmemx.data.
26463 (progmem_section_prefix): Remove.
26464 (avr_asm_init_sections): No need to initialize progmem_section.
26465 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26466 section name prefix.
26467 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26468 retrieve the progmem section.
26469 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26470 boundary to run over avr_addrspace[].
26471 (avr_register_target_pragmas): Ditto.
26472
656e6f37
JJ		 264732013-01-06 Jakub Jelinek <jakub@redhat.com>
26474
26475 * varasm.c (output_constant_def_contents): For asan_protect_global
26476 protected strings, adjust DECL_ALIGN if needed, before testing for
26477 anchored symbols.
26478 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26479 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26480 normal decls.
26481 (output_object_block): For asan protected decls, emit asan padding
26482 after their contents.
26483 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26484 (asan_finish_file): Test it here instead.
26485
6edc3e32
UB		 264862013-01-07 Nick Clifton <nickc@redhat.com>
26487 Matthias Klose <doko@debian.org>
26488 Doug Kwan <dougkwan@google.com>
26489 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 26490
26491 PR driver/55470
26492 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26493
26494 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26495
26496 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26497
7b0fe4f4 26498 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 26499
26500 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26501
2a095093
GJL		 265022013-01-07 Georg-Johann Lay <avr@gjlay.de>
26503
f3b503f4 26504 PR target/54461
2a095093
GJL		 26505 * doc/install.texi (Cross-Compiler-Specific Options): Document
26506 --with-avrlibc.
26507
383f9b34
TB		 265082013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26509
26510 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26511 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26512 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26513 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26514 vqmovun_high_s64): Fix source operand number and update copyright.
26515
3626621a
RB		 265162013-01-07 Richard Biener <rguenther@suse.de>
26517
26518 PR middle-end/55890
26519 * gimple.h (gimple_call_builtin_p): New overload.
26520 * gimple.c (validate_call): New function.
26521 (gimple_call_builtin_p): Likewise.
26522 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26523 Use gimple_call_builtin_p.
26524 (find_func_clobbers): Likewise.
26525 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26526 (strlen_optimize_stmt): Likewise.
26527
8b2ea410
JG		 265282013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26529
26530 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26531 (vld1q_dup_*): Likewise.
26532 (vld1_*): Likewise.
26533 (vld1q_*): Likewise.
26534 (vld1_lane_*): Likewise.
26535 (vld1q_lane_*): Likewise.
26536
9713d329
RB		 265372013-01-07 Richard Biener <rguenther@suse.de>
26538
26539 * lto-streamer.h (LTO_minor_version): Bump to 2.
26540
3520f7cc
JG		 265412013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26542
26543 * config/aarch64/aarch64-protos.h
26544 (aarch64_const_double_zero_rtx_p): Rename to...
26545 (aarch64_float_const_zero_rtx_p): ...this.
26546 (aarch64_float_const_representable_p): New.
26547 (aarch64_output_simd_mov_immediate): Likewise.
26548 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26549 move immediate case.
26550 * config/aarch64/aarch64.c
26551 (aarch64_const_double_zero_rtx_p): Rename to...
26552 (aarch64_float_const_zero_rtx_p): ...this.
26553 (aarch64_print_operand): Allow printing of new constants.
26554 (aarch64_valid_floating_const): New.
26555 (aarch64_legitimate_constant_p): Check for valid floating-point
26556 constants.
26557 (aarch64_simd_valid_immediate): Likewise.
26558 (aarch64_vect_float_const_representable_p): New.
26559 (aarch64_float_const_representable_p): Likewise.
26560 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26561 (aarch64_output_simd_mov_immediate): New.
26562 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26563 (*movdf_aarch64): Likewise.
26564 * config/aarch64/constraints.md (Ufc): New.
26565 (Y): call aarch64_float_const_zero_rtx.
26566 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26567
e167c04d
RB		 265682013-01-07 Richard Biener <rguenther@suse.de>
26569
26570 PR tree-optimization/55888
26571 PR tree-optimization/55862
26572 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26573 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26574 not if it is contained therein.
26575
0139aaab
GJL		 265762013-01-07 Georg-Johann Lay <avr@gjlay.de>
26577
26578 * config/avr/t-avr: Typo.
26579
4a176b23
GJL		 265802013-01-07 Georg-Johann Lay <avr@gjlay.de>
26581
26582 PR55243
26583 * config/avr/t-avr: Don't automatically rebuild
26584 $(srcdir)/config/avr/t-multilib
26585 $(srcdir)/config/avr/avr-tables.opt
26586 $(srcdir)/doc/avr-mmcu.texi
26587 (avr-mcus): New phony target to build them on request.
26588 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26589 * avr/avr-mcus.def: Adjust comments.
26590
c7afdc98
UB		 265912013-01-07 Uros Bizjak <ubizjak@gmail.com>
26592
26593 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26594
1ab05c31
RS		 265952013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26596
26597 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26598
488e3acc
RS		 265992013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26600
26601 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26602
a89599d2
DE		 266032013-01-05 David Edelsohn <dje.gcc@gmail.com>
26604
26605 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26606 to generate profiling.
26607 * config/rs6000/aix64.h (LIB_SPEC): Same.
26608
70f09188
AP		 266092013-01-04 Andrew Pinski <apinski@cavium.com>
26610
26611 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26612 New function.
26613 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26614
918d445f
UB		 266152013-01-04 Uros Bizjak <ubizjak@gmail.com>
26616
26617 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26618 unconditionally.
26619 (ix86_expand_move): Ditto.
26620 (ix86_zero_extend_to_Pmode): Ditto.
26621 (ix86_expand_call): Ditto.
26622 (ix86_expand_special_args_builtin): Ditto.
26623 (ix86_expand_builtin): Ditto.
26624
361618ec
RB		 266252013-01-04 Richard Biener <rguenther@suse.de>
26626
26627 PR tree-optimization/55862
26628 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26629 translating them through PHI nodes.
26630
87eab554
MJ		 266312013-01-04 Martin Jambor <mjambor@suse.cz>
26632
26633 PR tree-optimization/55755
26634 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26635 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26636
33879b9f
MP		 266372013-01-04 Marek Polacek <polacek@redhat.com>
26638
26639 PR middle-end/55859
26640 * opts.c (default_options_optimization): Clarify error message.
26641
3068819a
RB		 266422013-01-04 Richard Biener <rguenther@suse.de>
26643
26644 PR middle-end/55863
26645 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26646 reassociation.
26647
0e4ae794
JDA		 266482013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26649
26650 PR target/53789
26651 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 26652 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26653 references.
0e4ae794 26654
65c88cef
RH		 266552013-01-03 Richard Henderson <rth@redhat.com>
26656
26657 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26658 after eliminating TLS symbols.
26659
8ac16127
MG		 266602013-01-03 Marc Glisse <marc.glisse@inria.fr>
26661
26662 PR bootstrap/50167
26663 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26664 * graphite-poly.c (debug_gmp_value): Likewise.
26665
bb664f09
UB		 266662013-01-03 Uros Bizjak <ubizjak@gmail.com>
26667
26668 PR target/55712
26669 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26670 selected code model, define __code_mode_small__, __code_model_medium__,
26671 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26672 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26673 xchg temporary register with %k. Declare temporary register as
26674 early clobbered.
26675 [__x86_64__]: For medium and large code models, preserve %rbx register.
26676
bcf1ef00
RB		 266772013-01-03 Richard Biener <rguenther@suse.de>
26678
0506634a 26679 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 26680 (dump_subscript): Adjust.
26681 (finalize_ddr_dependent): Do not dump redundant info.
26682 (analyze_siv_subscript): Adjust.
26683 (subscript_dependence_tester): Likewise.
26684 (compute_affine_dependence): Likewise.
26685
59fd17e3
RB		 266862013-01-03 Richard Biener <rguenther@suse.de>
26687
26688 Revert
26689 2013-01-03 Richard Biener <rguenther@suse.de>
26690
26691 PR tree-optimization/55857
26692 * tree-vect-stmts.c (vectorizable_load): Do not setup
26693 re-alignment for invariant loads.
26694
26695 2013-01-02 Richard Biener <rguenther@suse.de>
26696
26697 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26698 invariant load do not generate a vector load from the scalar location.
59fd17e3 26699
595c2679
RB		 267002013-01-03 Richard Biener <rguenther@suse.de>
26701
26702 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26703 for not vectorizing.
26704 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26705 not build INDIRECT_REFs, call get_name once only.
26706 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26707 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26708
90bb7d7a
RB		 267092013-01-03 Richard Biener <rguenther@suse.de>
26710
26711 PR tree-optimization/55857
26712 * tree-vect-stmts.c (vectorizable_load): Do not setup
26713 re-alignment for invariant loads.
26714
f09b77ca
RB		 267152013-01-03 Richard Biener <rguenther@suse.de>
26716
26717 PR lto/55848
26718 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26719 prefer a built-in decl.
26720
3461a16e
JJ		 267212013-01-03 Jakub Jelinek <jakub@redhat.com>
26722
df8e2b4f
JJ		 26723 * gcc.c (process_command): Update copyright notice dates.
26724 * gcov.c (print_version): Likewise.
26725 * gcov-dump.c (print_version): Likewise.
26726
3461a16e
JJ		 26727 PR rtl-optimization/55838
26728 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26729 iv0.step, iv1.step and step.
26730
8b5546d6
JJ		 267312013-01-03 Jakub Jelinek <jakub@redhat.com>
26732 Marc Glisse <marc.glisse@inria.fr>
26733
26734 PR tree-optimization/55832
26735 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26736 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26737 integer_{one,zero}_node.
26738
8ab1d2e9
JJ		 267392013-01-03 Jakub Jelinek <jakub@redhat.com>
26740
26741 PR debug/54402
26742 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26743 * var-tracking.c (reverse_op): Don't add reverse ops to
26744 VALUEs that have already
0506634a 26745 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 26746
5b9ad1d4
GP		 267472013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26748
26749 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26750
e25a6711
TJ		 267512013-01-02 Teresa Johnson <tejohnson@google.com>
26752
26753 * dumpfile.c (dump_loc): Print filename with location.
26754 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26755 new location_t parameter to emit complete unroll message with
26756 new dump framework.
26757 (canonicalize_loop_induction_variables): Compute loops location
26758 and pass to try_unroll_loop_completely.
26759 * loop-unroll.c (report_unroll_peel): New function.
26760 (peel_loops_completely): Use new dump format with location
26761 for main dumpfile message, and invoke report_unroll_peel on success.
26762 (decide_unrolling_and_peeling): Ditto.
26763 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26764 by report_unroll_peel.
26765 (decide_peel_completely): Ditto.
26766 (decide_unroll_constant_iterations): Ditto.
26767 (decide_unroll_runtime_iterations): Ditto.
26768 (decide_peel_simple): Ditto.
26769 (decide_unroll_stupid): Ditto.
26770 * cfgloop.c (get_loop_location): New function.
26771 * cfgloop.h (get_loop_location): Declare.
26772
77878621
ST		 267732013-01-02 Sriraman Tallam <tmsriram@google.com>
26774
26775 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26776 NULL.
26777
9e65d03e
JDA		 267782013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26779
26780 PR middle-end/55198
26781 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26782 BLKmode objects when EXPAND_MEMORY is specified.
26783
6a7da30f
ST		 267842013-01-02 Sriraman Tallam <tmsriram@google.com>
26785
26786 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26787 in loop predicate.
26788 (fold_builtin_cpu): Do not share cpu model decls across statements.
26789
e78167f2
JM		 267902013-01-02 Jason Merrill <jason@redhat.com>
26791
26792 PR c++/55804
26793 * tree.c (build_array_type_1): Revert earlier change.
26794
8c075fb4
YZ		 267952013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26796
26797 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26798 "cortex-a57".
26799 * config/aarch64/aarch64-tune.md: Re-generate.
26800
0682ed3e
RB		 268012013-01-02 Richard Biener <rguenther@suse.de>
26802
26803 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26804 invariant load do not generate a vector load from the scalar location.
0682ed3e 26805
b8f6a302
RB		 268062013-01-02 Richard Biener <rguenther@suse.de>
26807
26808 PR bootstrap/55784
26809 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26810 * configure: Regenerate.
26811
04b80dbb
RS		 268122013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26813
26814 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26815 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26816 (expand_builtin_int_roundingfn_2): Keep the original target around
26817 for the fallback case.
26818
635b0b0c
RS		 268192013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26820
26821 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26822 to be clear for sign changes.
26823
42d57399
JH		 268242013-01-01 Jan Hubicka <jh@suse.cz>
26825
26826 * ipa-inline-analysis.c: Fix formatting.
26827
5bb6669d
JJ		 268282013-01-01 Jakub Jelinek <jakub@redhat.com>
26829
26830 PR tree-optimization/55831
26831 * tree-vect-loop.c (get_initial_def_for_induction): Use
26832 gsi_after_labels instead of gsi_start_bb.
ad41bd84 26833\f
86a2db33 26834Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 26835
26836Copying and distribution of this file, with or without modification,
26837are permitted in any medium without royalty provided the copyright
26838notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git