1 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3 PR tree-optimization/61136
4 * wide-int.h (multiple_of_p): Define a version that doesn't return
6 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
7 integer_zerop/const_binop pair.
8 (multiple_of_p): Likewise, converting both operands to widest_int
11 2014-05-09 Teresa Johnson <tejohnson@google.com>
13 * cgraphunit.c (analyze_functions): Use correct dump file.
15 2014-05-09 Florian Weimer <fweimer@redhat.com>
17 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
19 (stack_protect_return_slot_p): New function.
20 (expand_used_vars): Call stack_protect_decl_p and
21 stack_protect_return_slot_p for -fstack-protector-strong.
23 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
26 * fold-const.c (fold_binary_loc): Changed width of mask.
28 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
30 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
31 unsigned int initializers for regno_in, regno_out.
33 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
36 * config/avr/avr.md (cc): Add new attribute set_vzn.
37 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
38 Set cc insn attribute to set_vzn instead of set_zn for alternatives
40 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
41 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
42 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
44 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
47 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
49 * wide-int.cc (UTItype): Define.
50 (UDWtype): Define for appropriate W_TYPE_SIZE.
52 2014-05-09 Richard Biener <rguenther@suse.de>
54 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
55 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
56 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
57 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
58 ssa_propagate): Adjust.
60 2014-05-08 Jeff Law <law@redhat.com>
62 PR tree-optimization/61009
63 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
64 tri-state rather than a boolean. When a block is too big to
65 thread through, inform caller via negative return value.
66 (thread_across_edge): If a block was too big for normal threading,
67 then it's too big for a joiner too, so remove temporary equivalences
68 and return immediately.
70 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
71 Matthias Klose <doko@ubuntu.com>
74 * optc-gen.awk: Fix option handling for -Wunused-parameter.
76 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
79 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
81 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
84 * config/alpha/alpha.c: Include gimple-iterator.h.
85 (alpha_gimple_fold_builtin): New function. Move
86 ALPHA_BUILTIN_UMULH folding from ...
87 (alpha_fold_builtin): ... here.
88 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
90 2014-05-08 Wei Mi <wmi@google.com>
93 * config/i386/i386.c (ix86_compute_frame_layout): Update
94 preferred_stack_boundary for call, expanded from tls descriptor.
95 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
96 to depend on SP register.
97 (*tls_local_dynamic_base_32_gnu): Ditto.
98 (*tls_local_dynamic_32_once): Ditto.
99 (tls_global_dynamic_64_<mode>): Set
100 ix86_tls_descriptor_calls_expanded_in_cfun.
101 (tls_local_dynamic_base_64_<mode>): Ditto.
102 (tls_global_dynamic_32): Set
103 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
104 to depend on SP register.
105 (tls_local_dynamic_base_32): Ditto.
107 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
109 * config/arm/arm_neon.h: Update comment.
110 * config/arm/neon-docgen.ml: Delete.
111 * config/arm/neon-gen.ml: Delete.
112 * doc/arm-neon-intrinsics.texi: Update comment.
114 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
116 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
118 (vand, vorr, veor, vorn, vbic): Remove.
119 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
121 (neon_vsub_unspec): Likewise.
122 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
124 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
126 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
127 (vadd_s16): Likewise.
128 (vadd_s32): Likewise.
129 (vadd_f32): Likewise.
131 (vadd_u16): Likewise.
132 (vadd_u32): Likewise.
133 (vadd_s64): Likewise.
134 (vadd_u64): Likewise.
135 (vaddq_s8): Likewise.
136 (vaddq_s16): Likewise.
137 (vaddq_s32): Likewise.
138 (vaddq_s64): Likewise.
139 (vaddq_f32): Likewise.
140 (vaddq_u8): Likewise.
141 (vaddq_u16): Likewise.
142 (vaddq_u32): Likewise.
143 (vaddq_u64): Likewise.
145 (vmul_s16): Likewise.
146 (vmul_s32): Likewise.
147 (vmul_f32): Likewise.
149 (vmul_u16): Likewise.
150 (vmul_u32): Likewise.
152 (vmulq_s8): Likewise.
153 (vmulq_s16): Likewise.
154 (vmulq_s32): Likewise.
155 (vmulq_f32): Likewise.
156 (vmulq_u8): Likewise.
157 (vmulq_u16): Likewise.
158 (vmulq_u32): Likewise.
160 (vsub_s16): Likewise.
161 (vsub_s32): Likewise.
162 (vsub_f32): Likewise.
164 (vsub_u16): Likewise.
165 (vsub_u32): Likewise.
166 (vsub_s64): Likewise.
167 (vsub_u64): Likewise.
168 (vsubq_s8): Likewise.
169 (vsubq_s16): Likewise.
170 (vsubq_s32): Likewise.
171 (vsubq_s64): Likewise.
172 (vsubq_f32): Likewise.
173 (vsubq_u8): Likewise.
174 (vsubq_u16): Likewise.
175 (vsubq_u32): Likewise.
176 (vsubq_u64): Likewise.
178 (vand_s16): Likewise.
179 (vand_s32): Likewise.
181 (vand_u16): Likewise.
182 (vand_u32): Likewise.
183 (vand_s64): Likewise.
184 (vand_u64): Likewise.
185 (vandq_s8): Likewise.
186 (vandq_s16): Likewise.
187 (vandq_s32): Likewise.
188 (vandq_s64): Likewise.
189 (vandq_u8): Likewise.
190 (vandq_u16): Likewise.
191 (vandq_u32): Likewise.
192 (vandq_u64): Likewise.
194 (vorr_s16): Likewise.
195 (vorr_s32): Likewise.
197 (vorr_u16): Likewise.
198 (vorr_u32): Likewise.
199 (vorr_s64): Likewise.
200 (vorr_u64): Likewise.
201 (vorrq_s8): Likewise.
202 (vorrq_s16): Likewise.
203 (vorrq_s32): Likewise.
204 (vorrq_s64): Likewise.
205 (vorrq_u8): Likewise.
206 (vorrq_u16): Likewise.
207 (vorrq_u32): Likewise.
208 (vorrq_u64): Likewise.
210 (veor_s16): Likewise.
211 (veor_s32): Likewise.
213 (veor_u16): Likewise.
214 (veor_u32): Likewise.
215 (veor_s64): Likewise.
216 (veor_u64): Likewise.
217 (veorq_s8): Likewise.
218 (veorq_s16): Likewise.
219 (veorq_s32): Likewise.
220 (veorq_s64): Likewise.
221 (veorq_u8): Likewise.
222 (veorq_u16): Likewise.
223 (veorq_u32): Likewise.
224 (veorq_u64): Likewise.
226 (vbic_s16): Likewise.
227 (vbic_s32): Likewise.
229 (vbic_u16): Likewise.
230 (vbic_u32): Likewise.
231 (vbic_s64): Likewise.
232 (vbic_u64): Likewise.
233 (vbicq_s8): Likewise.
234 (vbicq_s16): Likewise.
235 (vbicq_s32): Likewise.
236 (vbicq_s64): Likewise.
237 (vbicq_u8): Likewise.
238 (vbicq_u16): Likewise.
239 (vbicq_u32): Likewise.
240 (vbicq_u64): Likewise.
242 (vorn_s16): Likewise.
243 (vorn_s32): Likewise.
245 (vorn_u16): Likewise.
246 (vorn_u32): Likewise.
247 (vorn_s64): Likewise.
248 (vorn_u64): Likewise.
249 (vornq_s8): Likewise.
250 (vornq_s16): Likewise.
251 (vornq_s32): Likewise.
252 (vornq_s64): Likewise.
253 (vornq_u8): Likewise.
254 (vornq_u16): Likewise.
255 (vornq_u32): Likewise.
256 (vornq_u64): Likewise.
258 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
260 * wide-int.cc (UTItype): Define.
261 (UDWtype): Define for appropriate W_TYPE_SIZE.
263 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
265 PR tree-optimization/59100
266 * tree-ssa-phiopt.c: Include tree-inline.h.
267 (neutral_element_p, absorbing_element_p): New functions.
268 (value_replacement): Handle conditional binary operations with a
269 neutral or absorbing element.
271 2014-05-08 Richard Biener <rguenther@suse.de>
273 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
274 folding the expression.
275 (valueize_expr): Remove.
276 (visit_reference_op_load): Do not valueize the result of
278 (simplify_binary_expression): Likewise.
279 (simplify_unary_expression): Likewise.
281 2014-05-08 Richard Biener <rguenther@suse.de>
283 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
284 looking at TYPE_ARG_TYPES.
286 2014-05-08 Richard Biener <rguenther@suse.de>
288 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
289 pointer propagation special-case.
291 2014-05-08 Bin Cheng <bin.cheng@arm.com>
293 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
294 core part of address expressions.
296 2014-05-08 Alan Modra <amodra@gmail.com>
299 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
300 loads and stores when -mno-strict-align at any alignment.
301 (expand_block_clear): Similarly. Also correct calculation of
304 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
307 * tree-complex.c (expand_complex_move): Keep line info when expanding
309 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
310 of complex expression. Use new argument to display correct location
311 for values coming from phi statement.
312 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
313 (warn_uninitialized_phi): Pass location of phi argument to
315 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
316 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
318 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
320 * config/rs6000/predicates.md (indexed_address_mem): New.
321 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
322 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
323 fpstore_ux, fpstore_u.
324 (sign_extend, indexed, update): New.
325 (cell_micro): Adjust.
326 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
327 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
328 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
329 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
330 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
331 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
332 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
333 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
334 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
335 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
336 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
337 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
338 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
339 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
340 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
342 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
343 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
344 *vsx_extract_<mode>_store): Adjust.
345 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
346 is_cracked_insn, insn_must_be_first_in_group,
347 insn_must_be_last_in_group): Adjust.
349 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
351 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
352 ppc440-fpstore): Adjust.
353 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
354 ppc476-fpstore): Adjust.
355 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
356 ppc601-fpstore): Adjust.
357 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
359 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
361 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
362 ppc7450-fpstore): Adjust.
363 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
364 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
365 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
367 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
368 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
369 cell-fpstore, cell-fpstore-update): Adjust.
370 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
371 ppce300c3_store, ppce300c3_fpstore): Adjust.
372 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
373 e500mc_fpstore): Adjust.
374 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
375 e500mc64_store, e500mc64_fpstore): Adjust.
376 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
377 e5500_fpstore): Adjust.
378 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
379 e6500_fpstore): Adjust.
380 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
382 * config/rs6000/power4.md (power4-load, power4-load-ext,
383 power4-load-ext-update, power4-load-ext-update-indexed,
384 power4-load-update-indexed, power4-load-update, power4-fpload,
385 power4-fpload-update, power4-store, power4-store-update,
386 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
388 * config/rs6000/power5.md (power5-load, power5-load-ext,
389 power5-load-ext-update, power5-load-ext-update-indexed,
390 power5-load-update-indexed, power5-load-update, power5-fpload,
391 power5-fpload-update, power5-store, power5-store-update,
392 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
394 * config/rs6000/power6.md (power6-load, power6-load-ext,
395 power6-load-update, power6-load-update-indexed,
396 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
397 power6-fpload-update, power6-store, power6-store-update,
398 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
400 * config/rs6000/power7.md (power7-load, power7-load-ext,
401 power7-load-update, power7-load-update-indexed,
402 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
403 power7-fpload-update, power7-store, power7-store-update,
404 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
406 * config/rs6000/power8.md (power8-load, power8-load-update,
407 power8-load-ext, power8-load-ext-update, power8-fpload,
408 power8-fpload-update, power8-store, power8-store-update-indexed,
409 power8-fpstore, power8-fpstore-update): Adjust.
410 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
412 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
413 titan_lsu_store, titan_lsu_fpstore): Adjust.
414 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
416 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
419 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
420 unrolled byte insns. Emit address increments after move insns.
422 2014-05-07 David Malcolm <dmalcolm@redhat.com>
424 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
425 const_gimple, rather than a gimple.
426 (gimple_call_builtin_p): Likewise, for the three variants.
428 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
429 (gimple_call_builtin_p): Likewise, for the three variants.
431 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
433 PR tree-optimization/61095
434 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
436 2014-05-07 Richard Biener <rguenther@suse.de>
438 PR tree-optimization/61034
439 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
440 (maybe_skip_until): Use translate to take into account
441 lattices when trying to do disambiguations.
442 (get_continuation_for_phi_1): Likewise.
443 (get_continuation_for_phi): Adjust for added translate arguments.
444 (walk_non_aliased_vuses): Likewise.
445 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
446 (walk_non_aliased_vuses): Likewise.
447 (call_may_clobber_ref_p_1): Declare.
448 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
449 calls. Stop early if we are only supposed to disambiguate.
450 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
452 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
454 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
455 Emit an error when the function has arguments.
457 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
459 * cfgloop.h (unswitch_loops): Remove.
460 * doc/passes.texi: Remove references to loop-unswitch.c
461 * timevar.def (TV_LOOP_UNSWITCH): Remove.
463 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
465 * tree-vect-data-refs.c (vect_grouped_load_supported): New
466 check for loads group of length 3.
467 (vect_permute_load_chain): New permutations for loads group of
469 * tree-vect-stmts.c (vect_model_load_cost): Change cost
470 of vec_perm_shuffle for the new permutations.
472 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
474 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
475 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
476 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
477 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
478 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
479 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
480 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
481 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
483 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
485 * loop-unswitch.c: Delete.
487 2014-05-07 Richard Biener <rguenther@suse.de>
489 * config.gcc: Always set need_64bit_hwint to yes.
491 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
493 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
494 of using optimize_size.
496 2014-05-06 Mike Stump <mikestump@comcast.net>
498 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
500 2014-05-06 Joseph Myers <joseph@codesourcery.com>
502 * config/i386/sse.md (*mov<mode>_internal)
503 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
504 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
505 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
506 (*<code><mode>3, *andnot<mode>3<mask_name>)
507 (<mask_codefor><code><mode>3<mask_name>): Only consider
508 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
510 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
513 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
515 * lra-constraints.c (valid_address_p): Move earlier in file.
516 Add a constraint argument to the address_info version.
517 (satisfies_memory_constraint_p): New function.
518 (satisfies_address_constraint_p): Likewise.
519 (process_alt_operands, curr_insn_transform): Use them.
520 (process_address): Pass the constraint to valid_address_p when
521 checking address operands.
523 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
525 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
526 to their respective blocks. Fix inadvertent use of "node".
528 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
530 * emit-rtl.c (init_derived_machine_modes): New functionm, split
532 (init_emit_once): ...here.
533 * rtl.h (init_derived_machine_modes): Declare.
534 * toplev.c (do_compile): Call it even if no_backend.
536 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
537 Mike Stump <mikestump@comcast.net>
538 Richard Sandiford <rdsandiford@googlemail.com>
539 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
541 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
542 (rtx_equal_for_memref_p): Update comment.
543 (adjust_offset_for_component_ref): Use wide-int interfaces.
544 * builtins.c (get_object_alignment_2): Likewise.
545 (c_readstr): Likewise.
546 (target_char_cast): Add comment.
547 (determine_block_size): Use wide-int interfaces.
548 (expand_builtin_signbit): Likewise.
549 (fold_builtin_int_roundingfn): Likewise.
550 (fold_builtin_bitop): Likewise.
551 (fold_builtin_bswap): Likewise.
552 (fold_builtin_logarithm): Use signop.
553 (fold_builtin_pow): Likewise.
554 (fold_builtin_memory_op): Use wide-int interfaces.
555 (fold_builtin_object_size): Likewise.
556 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
557 nb_iterations_estimate.
558 (record_niter_bound): Use wide-int interfaces.
559 (get_estimated_loop_iterations_int): Likewise.
560 (get_estimated_loop_iterations): Likewise.
561 (get_max_loop_iterations): Likewise.
562 * cfgloop.h: Include wide-int.h.
563 (struct nb_iter_bound): Change bound to widest_int.
564 (struct loop): Change nb_iterations_upper_bound and
565 nb_iterations_estimate to widest_int.
566 (record_niter_bound): Switch to use widest_int.
567 (get_estimated_loop_iterations): Likewise.
568 (get_max_loop_iterations): Likewise.
569 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
571 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
572 * combine.c (try_combine): Likewise.
573 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
574 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
576 (aarch64_float_const_representable_p): Likewise.
577 * config/arc/arc.c: Include wide-int.h.
578 (arc_can_use_doloop_p): Use wide-int interfaces.
579 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
580 (vfp3_const_double_index): Likewise.
581 * config/avr/avr.c (avr_out_round): Likewise.
582 (avr_fold_builtin): Likewise.
583 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
584 (bfin_can_use_doloop_p): Likewise.
585 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
586 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
587 * config/i386/i386.c: Include wide-int.h.
588 (ix86_data_alignment): Use wide-int interfaces.
589 (ix86_local_alignment): Likewise.
590 (ix86_emit_swsqrtsf): Update real_from_integer.
591 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
592 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
593 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
594 (zero_constant): Likewise.
595 (input_operand): Likewise.
596 (splat_input_operand): Likewise.
597 (non_logical_cint_operand): Change const_double to const_wide_int.
598 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
599 (easy_altivec_constant): Remove comment.
600 (paired_expand_vector_init): Use CONSTANT_P.
601 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
602 (rs6000_emit_move): Update checks.
603 (rs6000_aggregate_candidate): Use wide-int interfaces.
604 (rs6000_expand_ternop_builtin): Likewise.
605 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
606 (rs6000_assemble_integer): Likewise.
607 (rs6000_hash_constant): Likewise.
608 (output_toc): Likewise.
609 (rs6000_rtx_costs): Likewise.
610 (rs6000_emit_swrsqrt); Update call to real_from_integer.
611 * config/rs6000/rs6000-c.c: Include wide-int.h.
612 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
613 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
614 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
615 Handle CONST_WIDE_INT.
616 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
617 Use tree_fits_uhwi_p.
618 * config/sparc/sparc.c: Include wide-int.h.
619 (sparc_fold_builtin): Use wide-int interfaces.
620 * config/vax/vax.c: Include wide-int.h.
621 (vax_float_literal): Use real_from_integer.
622 * coretypes.h (struct hwivec_def): New.
625 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
626 (equiv_constant): Handle CONST_WIDE_INT.
627 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
628 (cselib_hash_rtx): Handle CONST_WIDE_INT.
629 * dbxout.c (stabstr_U): Use wide-int interfaces.
630 (dbxout_type): Update to use cst_fits_shwi_p.
631 * defaults.h (LOG2_BITS_PER_UNIT): Define.
632 (TARGET_SUPPORTS_WIDE_INT): Add default.
633 * dfp.c: Include wide-int.h.
634 (decimal_real_to_integer2): Use wide-int interfaces and rename to
635 decimal_real_to_integer.
636 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
637 decimal_real_to_integer.
638 * doc/generic.texi (Constant expressions): Update for wide_int.
639 * doc/rtl.texi (const_double): Likewise.
640 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
641 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
642 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
643 (REAL_VALUE_FROM_INT): Remove.
644 (TARGET_SUPPORTS_WIDE_INT): New.
645 * doc/tm.texi: Regenerate.
646 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
647 * double-int.h: Include wide-int.h.
648 (struct wi::int_traits): New.
649 * dwarf2out.c (get_full_len): New.
650 (dw_val_equal_p): Add case dw_val_class_wide_int.
651 (size_of_loc_descr): Likewise.
652 (output_loc_operands): Likewise.
653 (insert_double): Remove.
654 (insert_wide_int): New.
656 (print_die): Add case dw_val_class_wide_int.
657 (attr_checksum): Likewise.
658 (attr_checksum_ordered): Likewise.
659 (same_dw_val_p): Likewise.
660 (size_of_die): Likewise.
661 (value_format): Likewise.
662 (output_die): Likewise.
663 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
665 (clz_loc_descriptor): Use wide-int interfaces.
666 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
667 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
668 (round_up_to_align): Use wide-int interfaces.
669 (field_byte_offset): Likewise.
670 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
671 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
672 CONST_DOUBLE handling. Use wide-int interfaces.
673 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
674 (gen_enumeration_type_die): Use add_AT_wide.
675 (hash_loc_operands): Add case dw_val_class_wide_int.
676 (compare_loc_operands): Likewise.
677 * dwarf2out.h: Include wide-int.h.
679 (enum dw_val_class): Add dw_val_class_wide_int.
680 (struct dw_val_struct): Add val_wide.
681 * emit-rtl.c (const_wide_int_htab): New.
682 (const_wide_int_htab_hash): New.
683 (const_wide_int_htab_eq): New.
684 (lookup_const_wide_int): New.
685 (const_double_htab_hash): Use wide-int interfaces.
686 (const_double_htab_eq): Likewise.
687 (rtx_to_double_int): Conditionally compile for wide-int.
688 (immed_double_int_const): Rename to immed_wide_int_const and
690 (immed_double_const): Conditionally compile for wide-int.
691 (init_emit_once): Use wide-int interfaces.
692 * explow.c (plus_constant): Likewise.
693 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
694 (lshift_value): Use wide-int interfaces.
695 (expand_mult): Likewise.
696 (choose_multiplier): Likewise.
697 (expand_smod_pow2): Likewise.
698 (make_tree): Likewise.
699 * expr.c (convert_modes): Consolidate handling of constants.
700 Use wide-int interfaces.
701 (emit_group_load_1): Add note.
702 (store_expr): Update comment.
703 (get_inner_reference): Use wide-int interfaces.
704 (expand_constructor): Update comment.
705 (expand_expr_real_2): Use wide-int interfaces.
706 (expand_expr_real_1): Likewise.
707 (reduce_to_bit_field_precision): Likewise.
708 (const_vector_from_tree): Likewise.
709 * final.c: Include wide-int-print.h.
710 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
711 * fixed-value.c: Include wide-int.h.
712 (fixed_from_string): Use wide-int interfaces.
713 (fixed_to_decimal): Likewise.
714 (fixed_convert_from_real): Likewise.
715 (real_convert_from_fixed): Likewise.
716 * fold-const.h (mem_ref_offset): Return an offset_int.
717 (div_if_zero_remainder): Remove code parameter.
718 * fold-const.c (div_if_zero_remainder): Remove code parameter.
719 Use wide-int interfaces.
720 (may_negate_without_overflow_p): Use wide-int interfaces.
721 (negate_expr_p): Likewise.
722 (fold_negate_expr): Likewise.
723 (int_const_binop_1): Likewise.
724 (const_binop): Likewise.
725 (fold_convert_const_int_from_int): Likewise.
726 (fold_convert_const_int_from_real): Likewise.
727 (fold_convert_const_int_from_fixed): Likewise.
728 (fold_convert_const_fixed_from_int): Likewise.
729 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
730 (sign_bit_p): Use wide-int interfaces.
731 (make_range_step): Likewise.
732 (build_range_check): Likewise. Pass an integer of the correct type
733 instead of using integer_one_node.
734 (range_predecessor): Pass an integer of the correct type instead
735 of using integer_one_node.
736 (range_successor): Likewise.
737 (merge_ranges): Likewise.
738 (unextend): Use wide-int interfaces.
739 (extract_muldiv_1): Likewise.
740 (fold_div_compare): Likewise.
741 (fold_single_bit_test): Likewise.
742 (fold_sign_changed_comparison): Likewise.
743 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
744 (fold_plusminus_mult_expr): Use wide-int interfaces.
745 (native_encode_int): Likewise.
746 (native_interpret_int): Likewise.
747 (fold_unary_loc): Likewise.
748 (pointer_may_wrap_p): Likewise.
749 (size_low_cst): Likewise.
750 (mask_with_tz): Likewise.
751 (fold_binary_loc): Likewise.
752 (fold_ternary_loc): Likewise.
753 (multiple_of_p): Likewise.
754 (tree_call_nonnegative_warnv_p): Update calls to
755 tree_int_cst_min_precision and real_from_integer.
756 (fold_negate_const): Use wide-int interfaces.
757 (fold_abs_const): Likewise.
758 (fold_relational_const): Use tree_int_cst_lt.
759 (round_up_loc): Use wide-int interfaces.
760 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
761 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
762 * gengtype.c: Remove include of double-int.h.
763 (do_typedef): Use wide-int interfaces.
764 (open_base_files): Add wide-int.h.
765 (main): Add offset_int and widest_int typedefs.
766 * gengtype-lex.l: Handle "^".
767 (CXX_KEYWORD): Add "static".
768 * gengtype-parse.c (require3): New.
769 (require_template_declaration): Handle constant template arguments
770 and nested templates.
771 * gengtype-state.c: Don't include "double-int.h".
772 * genpreds.c (write_one_predicate_function): Update comment.
773 (write_tm_constrs_h): Add check for hval and lval use in
775 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
776 (add_to_sequence): Likewise.
777 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
778 and const_double_operand.
779 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
781 * gimple-fold.c (get_base_constructor): Likewise.
782 (fold_array_ctor_reference): Likewise.
783 (fold_nonarray_ctor_reference): Likewise.
784 (fold_const_aggregate_ref_1): Likewise.
785 (gimple_val_nonnegative_real_p): Likewise.
786 (gimple_fold_indirect_ref): Likewise.
787 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
788 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
789 (struct slsr_cand_d): Change index to be widest_int.
790 (struct incr_info_d): Change incr to be widest_int.
791 (alloc_cand_and_find_basis): Use wide-int interfaces.
792 (slsr_process_phi): Likewise.
793 (backtrace_base_for_ref): Likewise. Return a widest_int.
794 (restructure_reference): Take a widest_int instead of a double_int.
795 (slsr_process_ref): Use wide-int interfaces.
796 (create_mul_ssa_cand): Likewise.
797 (create_mul_imm_cand): Likewise.
798 (create_add_ssa_cand): Likewise.
799 (create_add_imm_cand): Take a widest_int instead of a double_int.
800 (slsr_process_add): Use wide-int interfaces.
801 (slsr_process_cast): Likewise.
802 (slsr_process_copy): Likewise.
803 (dump_candidate): Likewise.
804 (dump_incr_vec): Likewise.
805 (replace_ref): Likewise.
806 (cand_increment): Likewise. Return a widest_int.
807 (cand_abs_increment): Likewise.
808 (replace_mult_candidate): Take a widest_int instead of a double_int.
809 (replace_unconditional_candidate): Use wide-int interfaces.
810 (incr_vec_index): Take a widest_int instead of a double_int.
811 (create_add_on_incoming_edge): Likewise.
812 (create_phi_basis): Use wide-int interfaces.
813 (replace_conditional_candidate): Likewise.
814 (record_increment): Take a widest_int instead of a double_int.
815 (record_phi_increments): Use wide-int interfaces.
816 (phi_incr_cost): Take a widest_int instead of a double_int.
817 (lowest_cost_path): Likewise.
818 (total_savings): Likewise.
819 (analyze_increments): Use wide-int interfaces.
820 (ncd_with_phi): Take a widest_int instead of a double_int.
821 (ncd_of_cand_and_phis): Likewise.
822 (nearest_common_dominator_for_cands): Likewise.
823 (insert_initializers): Use wide-int interfaces.
824 (all_phi_incrs_profitable): Likewise.
825 (replace_one_candidate): Likewise.
826 (replace_profitable_candidates): Likewise.
827 * godump.c: Include wide-int-print.h.
828 (go_output_typedef): Use wide-int interfaces.
829 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
830 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
831 (build_loop_iteration_domains): Likewise.
832 * hooks.h: Include wide-int.h rather than double-int.h.
833 (hook_bool_dint_dint_uint_bool_true): Delete.
834 (hook_bool_wint_wint_uint_bool_true): Declare.
835 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
836 (hook_bool_wint_wint_uint_bool_true): New.
837 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
839 (ubsan_expand_si_overflow_mul_check): Likewise.
840 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
841 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
842 (get_ancestor_addr_info): Likewise.
843 (ipa_modify_call_arguments): Likewise.
844 * loop-doloop.c (doloop_modify): Likewise.
845 (doloop_optimize): Likewise.
846 * loop-iv.c (iv_number_of_iterations): Likewise.
847 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
848 (unroll_loop_constant_iterations): Likewise.
849 (decide_unroll_runtime_iterations): Likewise.
850 (unroll_loop_runtime_iterations): Likewise.
851 (decide_peel_simple): Likewise.
852 (decide_unroll_stupid): Likewise.
853 * lto-streamer-in.c (streamer_read_wi): Add.
854 (input_cfg): Use wide-int interfaces.
855 (lto_input_tree_1): Likewise.
856 * lto-streamer-out.c (streamer_write_wi): Add.
857 (hash_tree): Use wide-int interfaces.
858 (output_cfg): Likewise.
859 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
860 (GTFILES): Add wide-int.h and signop.h.
861 (TAGS): Look for .cc files too.
862 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
863 * optabs.c (expand_subword_shift): Likewise.
864 (expand_doubleword_shift): Likewise.
865 (expand_absneg_bit): Likewise.
866 (expand_copysign_absneg): Likewise.
867 (expand_copysign_bit): Likewise.
868 * postreload.c (reload_cse_simplify_set): Likewise.
869 * predict.c (predict_iv_comparison): Likewise.
870 * pretty-print.h: Include wide-int-print.h.
872 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
873 * print-tree.c: Include wide-int-print.h.
874 (print_node_brief): Use wide-int interfaces.
875 (print_node): Likewise.
876 * read-rtl.c (validate_const_wide_int): New.
877 (read_rtx_code): Add CONST_WIDE_INT case.
878 * real.c: Include wide-int.h.
879 (real_to_integer2): Delete.
880 (real_to_integer): New function, returning a wide_int.
881 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
882 (ten_to_ptwo): Update call to real_from_integer.
883 (real_digit): Likewise.
884 * real.h: Include signop.h, wide-int.h and insn-modes.h.
885 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
886 (REAL_VALUE_TO_INT): Delete.
887 (real_to_integer): Declare a wide-int form.
888 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
889 * recog.c (const_int_operand): Improve comment.
890 (const_scalar_int_operand): New.
891 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
892 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
893 (split_double): Likewise.
894 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
895 (rtx_size): Likewise.
896 (rtx_alloc_stat_v): New.
897 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
898 (cwi_output_hex): New.
899 (iterative_hash_rtx): Handle CONST_WIDE_INT.
900 (cwi_check_failed_bounds): New.
901 * rtl.def (CONST_WIDE_INT): New.
902 * rtl.h: Include <utility> and wide-int.h.
903 (struct hwivec_def): New.
904 (CWI_GET_NUM_ELEM): New.
905 (CWI_PUT_NUM_ELEM): New.
906 (struct rtx_def): Add num_elem and hwiv.
907 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
908 (CASE_CONST_UNIQUE): Likewise.
909 (CASE_CONST_ANY): Likewise.
910 (CONST_SCALAR_INT_P): Likewise.
911 (CONST_WIDE_INT_P): New.
914 (cwi_check_failed_bounds): New.
917 (CONST_WIDE_INT_VEC) New.
918 (CONST_WIDE_INT_NUNITS) New.
919 (CONST_WIDE_INT_ELT) New.
920 (rtx_mode_t): New type.
921 (wi::int_traits <rtx_mode_t>): New.
923 (wi::min_value): New.
924 (wi::max_value): New.
926 (const_wide_int_alloc): New.
927 (immed_wide_int_const): New.
928 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
929 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
930 * signop.h: New file.
931 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
932 (simplify_const_unary_operation): Use wide-int interfaces.
933 (simplify_binary_operation_1): Likewise.
934 (simplify_const_binary_operation): Likewise.
935 (simplify_const_relational_operation): Likewise.
936 (simplify_immed_subreg): Likewise.
937 * stmt.c (expand_case): Likewise.
938 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
939 signop rather than a bool.
940 * stor-layout.c (layout_type): Use wide-int interfaces.
941 (initialize_sizetypes): Update calls to
942 set_min_and_max_values_for_integral_type.
943 (set_min_and_max_values_for_integral_type): Take a signop rather
944 than a bool. Use wide-int interfaces.
945 (fixup_signed_type): Update accordingly. Remove
946 HOST_BITS_PER_DOUBLE_INT limit.
947 (fixup_unsigned_type): Likewise.
948 * system.h (STATIC_CONSTANT_P): New.
949 (STATIC_ASSERT): New.
950 * target.def (can_use_doloop_p): Take widest_ints rather than
952 * target.h: Include wide-int.h rather than double-int.h.
953 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
955 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
956 rather than INT_CST_LT_UNSIGNED.
957 (can_use_doloop_if_innermost): Take widest_ints rather than
959 * tree-affine.c: Include wide-int-print.h.
960 (double_int_ext_for_comb): Delete.
961 (wide_int_ext_for_comb): New.
962 (aff_combination_zero): Use wide-int interfaces.
963 (aff_combination_const): Take a widest_int instead of a double_int.
964 (aff_combination_elt): Use wide-int interfaces.
965 (aff_combination_scale): Take a widest_int instead of a double_int.
966 (aff_combination_add_elt): Likewise.
967 (aff_combination_add_cst): Likewise.
968 (aff_combination_add): Use wide-int interfaces.
969 (aff_combination_convert): Likewise.
970 (tree_to_aff_combination): Likewise.
971 (add_elt_to_tree): Take a widest_int instead of a double_int.
972 (aff_combination_to_tree): Use wide-int interfaces.
973 (aff_combination_remove_elt): Likewise.
974 (aff_combination_add_product): Take a widest_int instead of
976 (aff_combination_mult): Use wide-int interfaces.
977 (aff_combination_expand): Likewise.
978 (double_int_constant_multiple_p): Delete.
979 (wide_int_constant_multiple_p): New.
980 (aff_combination_constant_multiple_p): Take a widest_int pointer
981 instead of a double_int pointer.
982 (print_aff): Use wide-int interfaces.
983 (get_inner_reference_aff): Take a widest_int pointer
984 instead of a double_int pointer.
985 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
986 * tree-affine.h: Include wide-int.h.
987 (struct aff_comb_elt): Change type of coef to widest_int.
988 (struct affine_tree_combination): Change type of offset to widest_int.
989 (double_int_ext_for_comb): Delete.
990 (wide_int_ext_for_comb): New.
991 (aff_combination_const): Use widest_int instead of double_int.
992 (aff_combination_scale): Likewise.
993 (aff_combination_add_elt): Likewise.
994 (aff_combination_constant_multiple_p): Likewise.
995 (get_inner_reference_aff): Likewise.
996 (aff_comb_cannot_overlap_p): Likewise.
997 (aff_combination_zero_p): Use wide-int interfaces.
998 * tree.c: Include tree.h.
999 (init_ttree): Use make_int_cst.
1000 (tree_code_size): Removed code for INTEGER_CST case.
1001 (tree_size): Add INTEGER_CST case.
1002 (make_node_stat): Update comment.
1003 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1004 (build_int_cst_type): Use wide-int interfaces.
1005 (double_int_to_tree): Likewise.
1006 (double_int_fits_to_tree_p): Delete.
1007 (force_fit_type_double): Delete.
1008 (force_fit_type): New.
1009 (int_cst_hash_hash): Use wide-int interfaces.
1010 (int_cst_hash_eq): Likewise.
1011 (build_int_cst_wide): Delete.
1012 (wide_int_to_tree): New.
1013 (cache_integer_cst): Use wide-int interfaces.
1014 (build_low_bits_mask): Likewise.
1015 (cst_and_fits_in_hwi): Likewise.
1016 (real_value_from_int_cst): Likewise.
1017 (make_int_cst_stat): New.
1018 (integer_zerop): Use wide_int interfaces.
1019 (integer_onep): Likewise.
1020 (integer_all_onesp): Likewise.
1021 (integer_pow2p): Likewise.
1022 (integer_nonzerop): Likewise.
1023 (tree_log2): Likewise.
1024 (tree_floor_log2): Likewise.
1025 (tree_ctz): Likewise.
1026 (int_size_in_bytes): Likewise.
1027 (mem_ref_offset): Return an offset_int rather than a double_int.
1028 (build_type_attribute_qual_variant): Use wide_int interfaces.
1029 (type_hash_eq): Likewise
1030 (tree_int_cst_equal): Likewise.
1031 (tree_int_cst_lt): Delete.
1032 (tree_int_cst_compare): Likewise.
1033 (tree_fits_shwi_p): Use wide_int interfaces.
1034 (tree_fits_uhwi_p): Likewise.
1035 (tree_int_cst_sign_bit): Likewise.
1036 (tree_int_cst_sgn): Likewise.
1037 (tree_int_cst_min_precision): Take a signop rather than a bool.
1038 (simple_cst_equal): Use wide_int interfaces.
1039 (compare_tree_int): Likewise.
1040 (iterative_hash_expr): Likewise.
1041 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1043 (get_type_static_bounds): Use wide_int interfaces.
1044 (tree_int_cst_elt_check_failed): New.
1045 (build_common_tree_nodes): Reordered to set prec before filling in
1047 (int_cst_value): Check cst_and_fits_in_hwi.
1048 (widest_int_cst_value): Use wide_int interfaces.
1049 (upper_bound_in_type): Likewise.
1050 (lower_bound_in_type): Likewise.
1051 (num_ending_zeros): Likewise.
1052 (drop_tree_overflow): Likewise.
1053 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1054 (gen_conditions_for_pow_cst_base): Likewise.
1055 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1056 (group_case_labels_stmt): Use wide-int interfaces.
1057 (verify_gimple_assign_binary): Likewise.
1058 (print_loop): Likewise.
1059 * tree-chrec.c (tree_fold_binomial): Likewise.
1060 * tree-core.h (struct tree_base): Add int_length.
1061 (struct tree_int_cst): Change rep of value.
1062 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1063 (dr_may_alias_p): Likewise.
1064 (max_stmt_executions_tree): Likewise.
1065 * tree.def (INTEGER_CST): Update comment.
1066 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1067 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1068 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1069 (dequeue_and_dump): Use wide-int interfaces.
1070 * tree.h: Include wide-int.h.
1071 (NULL_TREE): Moved to earlier loc in file.
1072 (TREE_INT_CST_ELT_CHECK): New.
1073 (tree_int_cst_elt_check_failed): New.
1075 (TREE_INT_CST): Delete.
1076 (TREE_INT_CST_LOW): Use wide-int interfaces.
1077 (TREE_INT_CST_HIGH): Delete.
1078 (TREE_INT_CST_NUNITS): New.
1079 (TREE_INT_CST_EXT_NUNITS): Likewise.
1080 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1081 (TREE_INT_CST_ELT): Likewise.
1082 (INT_CST_LT): Delete.
1083 (tree_int_cst_elt_check): New (two forms).
1084 (type_code_size): Update comment.
1085 (make_int_cst_stat, make_int_cst): New.
1086 (tree_to_double_int): Delete.
1087 (double_int_fits_to_tree_p): Delete.
1088 (force_fit_type_double): Delete.
1089 (build_int_cstu): Replace with out-of-line function.
1090 (build_int_cst_wide): Delete.
1091 (tree_int_cst_lt): Define inline.
1092 (tree_int_cst_le): New.
1093 (tree_int_cst_compare): Define inline.
1094 (tree_int_cst_min_precision): Take a signop rather than a bool.
1095 (wi::int_traits <const_tree>): New.
1096 (wi::int_traits <tree>): New.
1097 (wi::extended_tree): New.
1098 (wi::int_traits <wi::extended_tree>): New.
1099 (wi::to_widest): New.
1100 (wi::to_offset): New.
1101 (wi::fits_to_tree_p): New.
1102 (wi::min_value): New.
1103 (wi::max_value): New.
1104 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1105 (copy_tree_body_r): Likewise.
1106 * tree-object-size.c (compute_object_offset): Likewise.
1107 (addr_object_size): Likewise.
1108 * tree-predcom.c: Include wide-int-print.h.
1109 (struct dref_d): Change type of offset to widest_int.
1110 (dump_dref): Call wide-int printer.
1111 (aff_combination_dr_offset): Use wide-int interfaces.
1112 (determine_offset): Take a widest_int pointer rather than a
1114 (split_data_refs_to_components): Use wide-int interfaces.
1115 (suitable_component_p): Likewise.
1116 (order_drefs): Likewise.
1117 (add_ref_to_chain): Likewise.
1118 (valid_initializer_p): Likewise.
1119 (determine_roots_comp): Likewise.
1120 * tree-pretty-print.c: Include wide-int-print.h.
1121 (dump_generic_node): Use wide-int interfaces.
1122 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1123 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1124 (move_fixed_address_to_symbol): Likewise.
1125 (move_hint_to_base): Likewise.
1126 (move_pointer_to_base): Likewise.
1127 (move_variant_to_index): Likewise.
1128 (most_expensive_mult_to_index): Likewise.
1129 (addr_to_parts): Likewise.
1130 (copy_ref_info): Likewise.
1131 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1132 (indirect_refs_may_alias_p): Likewise.
1133 (stmt_kills_ref_p_1): Likewise.
1134 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1135 * tree-ssa-ccp.c: Update comment at top of file. Include
1137 (struct prop_value_d): Change type of mask to widest_int.
1138 (extend_mask): New function.
1139 (dump_lattice_value): Use wide-int interfaces.
1140 (get_default_value): Likewise.
1141 (set_constant_value): Likewise.
1142 (set_value_varying): Likewise.
1143 (valid_lattice_transition): Likewise.
1144 (set_lattice_value): Likewise.
1145 (value_to_double_int): Delete.
1146 (value_to_wide_int): New.
1147 (get_value_from_alignment): Use wide-int interfaces.
1148 (get_value_for_expr): Likewise.
1149 (do_dbg_cnt): Likewise.
1150 (ccp_finalize): Likewise.
1151 (ccp_lattice_meet): Likewise.
1152 (bit_value_unop_1): Use widest_ints rather than double_ints.
1153 (bit_value_binop_1): Likewise.
1154 (bit_value_unop): Use wide-int interfaces.
1155 (bit_value_binop): Likewise.
1156 (bit_value_assume_aligned): Likewise.
1157 (evaluate_stmt): Likewise.
1158 (ccp_fold_stmt): Likewise.
1159 (visit_cond_stmt): Likewise.
1160 (ccp_visit_stmt): Likewise.
1161 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1162 (constant_pointer_difference): Likewise.
1163 (associate_pointerplus): Likewise.
1164 (combine_conversions): Likewise.
1165 * tree-ssa-loop.h: Include wide-int.h.
1166 (struct tree_niter_desc): Change type of max to widest_int.
1167 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1168 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1169 (remove_redundant_iv_tests): Likewise.
1170 (canonicalize_loop_induction_variables): Likewise.
1171 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1172 (constant_multiple_of): Take a widest_int pointer instead of
1173 a double_int pointer.
1174 (get_computation_aff): Use wide-int interfaces.
1175 (ptr_difference_cost): Likewise.
1176 (difference_cost): Likewise.
1177 (get_loop_invariant_expr_id): Likewise.
1178 (get_computation_cost_at): Likewise.
1179 (iv_elimination_compare_lt): Likewise.
1180 (may_eliminate_iv): Likewise.
1181 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1182 instead of double_int.
1183 (max_loop_iterations): Likewise.
1184 (max_stmt_executions): Likewise.
1185 (estimated_stmt_executions): Likewise.
1186 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1187 (split_to_var_and_offset): Use wide-int interfaces.
1188 (determine_value_range): Likewise.
1189 (bound_difference_of_offsetted_base): Likewise.
1190 (bounds_add): Take a widest_int instead of a double_int.
1191 (number_of_iterations_ne_max): Use wide-int interfaces.
1192 (number_of_iterations_ne): Likewise.
1193 (number_of_iterations_lt_to_ne): Likewise.
1194 (assert_loop_rolls_lt): Likewise.
1195 (number_of_iterations_lt): Likewise.
1196 (number_of_iterations_le): Likewise.
1197 (number_of_iterations_cond): Likewise.
1198 (number_of_iterations_exit): Likewise.
1199 (finite_loop_p): Likewise.
1200 (derive_constant_upper_bound_assign): Likewise.
1201 (derive_constant_upper_bound): Return a widest_int.
1202 (derive_constant_upper_bound_ops): Likewise.
1203 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1204 (record_estimate): Take a widest_int rather than a double_int.
1205 (record_nonwrapping_iv): Use wide-int interfaces.
1206 (double_int_cmp): Delete.
1207 (wide_int_cmp): New.
1208 (bound_index): Take a widest_int rather than a double_int.
1209 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1210 (maybe_lower_iteration_bound): Likewise.
1211 (estimate_numbers_of_iterations_loop): Likewise.
1212 (estimated_loop_iterations): Take a widest_int pointer than than
1213 a double_int pointer.
1214 (estimated_loop_iterations_int): Use wide-int interfaces.
1215 (max_loop_iterations): Take a widest_int pointer than than
1216 a double_int pointer.
1217 (max_loop_iterations_int): Use wide-int interfaces.
1218 (max_stmt_executions): Take a widest_int pointer than than
1219 a double_int pointer.
1220 (estimated_stmt_executions): Likewise.
1221 (n_of_executions_at_most): Use wide-int interfaces.
1222 (scev_probably_wraps_p): Likewise.
1223 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1225 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1227 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1228 double_ints. Adjust for trailing_wide_ints <3> representation.
1229 (set_nonzero_bits): Likewise.
1230 (get_range_info): Return wide_ints rather than double_ints.
1231 Adjust for trailing_wide_ints <3> representation.
1232 (get_nonzero_bits): Likewise.
1233 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1235 * tree-ssanames.h (struct range_info_def): Replace min, max and
1236 nonzero_bits with a trailing_wide_ints <3>.
1237 (set_range_info): Use wide_int_refs rather than double_ints.
1238 (set_nonzero_bits): Likewise.
1239 (get_range_info): Return wide_ints rather than double_ints.
1240 (get_nonzero_bits): Likewise.
1241 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1242 * tree-ssa-pre.c (phi_translate_1): Likewise.
1243 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1244 (acceptable_pow_call): Likewise.
1245 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1247 (vn_reference_fold_indirect): Likewise.
1248 (vn_reference_maybe_forwprop_address): Likewise.
1249 (valueize_refs_1): Likewise.
1250 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1251 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1252 tree_int_cst_lt and tree_int_cst_le.
1253 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1255 (streamer_alloc_tree): Likewise.
1256 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1257 (streamer_write_tree_header): Likewise.
1258 (streamer_write_integer_cst): Likewise.
1259 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1260 (build_constructors): Likewise.
1261 (array_value_type): Likewise.
1262 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1263 (vect_check_gather): Likewise.
1264 * tree-vect-generic.c (build_replicated_const): Likewise.
1265 (expand_vector_divmod): Likewise.
1266 * tree-vect-loop.c (vect_transform_loop): Likewise.
1267 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1268 (vect_do_peeling_for_alignment): Likewise.
1269 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1270 * tree-vrp.c: Include wide-int.h.
1271 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1272 (extract_range_from_assert): Use wide-int interfaces.
1273 (vrp_int_const_binop): Likewise.
1274 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1275 double_int pointers.
1276 (ranges_from_anti_range): Use wide-int interfaces.
1277 (quad_int_cmp): Delete.
1278 (quad_int_pair_sort): Likewise.
1279 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1280 (extract_range_from_unary_expr_1): Likewise.
1281 (adjust_range_with_scev): Likewise.
1282 (masked_increment): Take and return wide_ints rather than double_ints.
1283 (register_edge_assert_for_2): Use wide-int interfaces.
1284 (check_array_ref): Likewise.
1285 (search_for_addr_array): Likewise.
1286 (maybe_set_nonzero_bits): Likewise.
1287 (union_ranges): Pass an integer of the correct type instead of
1288 using integer_one_node.
1289 (intersect_ranges): Likewise.
1290 (simplify_truth_ops_using_ranges): Likewise.
1291 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1292 (range_fits_type_p): Likewise.
1293 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1295 (simplify_conversion_using_ranges): Use wide-int interfaces.
1296 (simplify_float_conversion_using_ranges): Likewise.
1297 (vrp_finalize): Likewise.
1298 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1299 (gimple_stringops_transform): Likewise.
1300 * varasm.c (decode_addr_const): Likewise.
1301 (const_hash_1): Likewise.
1302 (const_rtx_hash_1): Likewise
1303 (output_constant): Likewise.
1304 (array_size_for_constructor): Likewise.
1305 (output_constructor_regular_field): Likewise.
1306 (output_constructor_bitfield): Likewise.
1307 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1308 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1310 * gencheck.c: Define BITS_PER_UNIT.
1313 * wide-int-print.cc: New.
1314 * wide-int-print.h: New.
1316 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1318 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1320 2014-05-06 Richard Biener <rguenther@suse.de>
1322 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1323 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1324 (TODO_verify_all): Adjust.
1325 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1326 TODO_verify_stmts and TODO_verify_rtl_sharing.
1327 * bb-reorder.c: Likewise.
1328 * cfgexpand.c: Likewise.
1329 * cprop.c: Likewise.
1331 * function.c: Likewise.
1332 * fwprop.c: Likewise.
1334 * gimple-ssa-isolate-paths.c: Likewise.
1335 * gimple-ssa-strength-reduction.c: Likewise.
1336 * ipa-split.c: Likewise.
1337 * loop-init.c: Likewise.
1338 * loop-unroll.c: Likewise.
1339 * lower-subreg.c: Likewise.
1340 * modulo-sched.c: Likewise.
1341 * postreload-gcse.c: Likewise.
1342 * predict.c: Likewise.
1343 * recog.c: Likewise.
1344 * sched-rgn.c: Likewise.
1345 * store-motion.c: Likewise.
1346 * tracer.c: Likewise.
1347 * trans-mem.c: Likewise.
1348 * tree-call-cdce.c: Likewise.
1349 * tree-cfg.c: Likewise.
1350 * tree-cfgcleanup.c: Likewise.
1351 * tree-complex.c: Likewise.
1352 * tree-eh.c: Likewise.
1353 * tree-emutls.c: Likewise.
1354 * tree-if-conv.c: Likewise.
1355 * tree-into-ssa.c: Likewise.
1356 * tree-loop-distribution.c: Likewise.
1357 * tree-object-size.c: Likewise.
1358 * tree-parloops.c: Likewise.
1359 * tree-pass.h: Likewise.
1360 * tree-sra.c: Likewise.
1361 * tree-ssa-ccp.c: Likewise.
1362 * tree-ssa-copy.c: Likewise.
1363 * tree-ssa-copyrename.c: Likewise.
1364 * tree-ssa-dce.c: Likewise.
1365 * tree-ssa-dom.c: Likewise.
1366 * tree-ssa-dse.c: Likewise.
1367 * tree-ssa-forwprop.c: Likewise.
1368 * tree-ssa-ifcombine.c: Likewise.
1369 * tree-ssa-loop-ch.c: Likewise.
1370 * tree-ssa-loop-ivcanon.c: Likewise.
1371 * tree-ssa-loop.c: Likewise.
1372 * tree-ssa-math-opts.c: Likewise.
1373 * tree-ssa-phiopt.c: Likewise.
1374 * tree-ssa-phiprop.c: Likewise.
1375 * tree-ssa-pre.c: Likewise.
1376 * tree-ssa-reassoc.c: Likewise.
1377 * tree-ssa-sink.c: Likewise.
1378 * tree-ssa-strlen.c: Likewise.
1379 * tree-ssa-tail-merge.c: Likewise.
1380 * tree-ssa-uncprop.c: Likewise.
1381 * tree-switch-conversion.c: Likewise.
1382 * tree-tailcall.c: Likewise.
1383 * tree-vect-generic.c: Likewise.
1384 * tree-vectorizer.c: Likewise.
1385 * tree-vrp.c: Likewise.
1387 * var-tracking.c: Likewise.
1388 * bt-load.c: Likewise.
1389 * cfgcleanup.c: Likewise.
1390 * combine-stack-adj.c: Likewise.
1391 * combine.c: Likewise.
1392 * compare-elim.c: Likewise.
1393 * config/epiphany/resolve-sw-modes.c: Likewise.
1394 * config/i386/i386.c: Likewise.
1395 * config/mips/mips.c: Likewise.
1396 * config/s390/s390.c: Likewise.
1397 * config/sh/sh_treg_combine.cc: Likewise.
1398 * config/sparc/sparc.c: Likewise.
1401 * final.c: Likewise.
1402 * ifcvt.c: Likewise.
1403 * mode-switching.c: Likewise.
1404 * passes.c: Likewise.
1405 * postreload.c: Likewise.
1407 * reg-stack.c: Likewise.
1408 * regcprop.c: Likewise.
1409 * regrename.c: Likewise.
1412 2014-05-06 Richard Biener <rguenther@suse.de>
1415 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1416 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1418 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1421 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1423 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1424 Tom de Vries <tom@codesourcery.com>
1426 * target.def (call_fusage_contains_non_callee_clobbers): New
1428 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1430 (@node Miscellaneous Register Hooks): New node.
1431 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1432 * doc/tm.texi: Regenerate.
1434 2014-05-05 Marek Polacek <polacek@redhat.com>
1437 * opts.c (common_handle_option): Call error_at instead of warning_at.
1439 2014-05-05 Richard Biener <rguenther@suse.de>
1441 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1442 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1443 under the TODO_verify_il umbrella.
1445 2014-05-05 Richard Biener <rguenther@suse.de>
1447 * passes.c (execute_function_todo): Move TODO_verify_flow under
1448 the TODO_verify_ul umbrella.
1450 2014-05-05 Richard Biener <rguenther@suse.de>
1453 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1454 X & CST away from a CST that is the mask of a mode.
1456 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1458 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1459 int argument to enum machine_mode.
1460 (picochip_class_max_nregs): Ditto.
1461 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1462 (picochip_class_max_nregs): Ditto.
1464 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1466 * target.def: Add new target hook.
1467 * doc/tm.texi: Regenerate.
1468 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1469 * targhooks.c (default_keep_leaf_when_profiled): New function.
1471 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1472 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1474 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1476 PR tree-optimization/60363
1477 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1478 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1479 (update_destination_phis): New parameter.
1480 (create_edge_and_update_destination_phis): Ditto.
1481 (ssa_fix_duplicate_block_edges): Pass new arguments.
1482 (thread_single_edge): Ditto.
1484 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1486 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1487 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1488 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1489 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1490 Use RS6000_BTM_HARD_FLOAT.
1491 (BU_MISC_2): Likewise.
1492 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1493 RS6000_BTM_HARD_FLOAT.
1494 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1496 (rs6000_invalid_builtin): Add hard floating builtin support.
1497 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1498 hard float builtins.
1499 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1501 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1503 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1504 Add missing function* argument.
1506 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1508 * lra-constraints.c (valid_address_p): Move earlier in file.
1509 Add a constraint argument to the address_info version.
1510 (satisfies_memory_constraint_p): New function.
1511 (satisfies_address_constraint_p): Likewise.
1512 (process_alt_operands, curr_insn_transform): Use them.
1513 (process_address): Pass the constraint to valid_address_p when
1514 checking address operands.
1516 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1518 * config/mips/mips.c (mips_isa_rev): New variable.
1519 (mips_set_architecture): Set it.
1520 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1522 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1523 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1524 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1525 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1526 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1527 conditions in terms of mips_isa_rev.
1528 (mips_isa_rev): Declare.
1530 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1532 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1533 (prob_unlikely, prob_likely): Make variables const.
1535 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1537 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1539 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1541 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1543 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1545 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1546 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1547 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1549 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1550 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1552 Replace usage of ROUND_REG with sh_round_reg.
1553 Use CEIL instead of ROUND_ADVANCE.
1555 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1558 * config/sh/sh.c: Include stdlib headers before everything else.
1560 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1562 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1563 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1564 (gimplify_adjust_omp_clauses): Simd region is never
1565 directly nested in combined parallel. Instead, for linear
1566 with copyin/copyout, if in combined for simd loop, make decl
1567 firstprivate/lastprivate on OMP_FOR.
1568 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1569 expand_omp_for_static_chunk): When setting endvar, also set
1570 fd->loop.v to the same value.
1572 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1574 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1576 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1578 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1581 2014-05-02 Marek Polacek <polacek@redhat.com>
1583 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1585 2014-05-02 Kito Cheng <kito@0xlab.org>
1587 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1588 to a C expression marco.
1589 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1590 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1591 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1592 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1593 HONOR_REG_ALLOC_ORDER.
1594 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1596 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1598 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1600 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1602 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1604 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1606 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1607 (convert_scalar_cond_reduction): Likewise.
1608 (predicate_scalar_phi): Add recognition and transformation
1609 of simple conditioanl reduction to be vectorizable.
1611 2014-05-01 Marek Polacek <polacek@redhat.com>
1614 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1616 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1618 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1619 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1620 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1621 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1622 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1623 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1624 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1625 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1627 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1629 * config/arc/arc.opt (mlra): Move comment above option name
1630 to avoid mis-parsing as language options.
1632 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1634 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1635 * config/sol2.h: ... here.
1636 * config/sol2-10.h: Remove.
1638 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1639 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1640 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1641 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1642 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1643 * config/sol2.h: ... here.
1644 (SECTION_NAME_FORMAT): Don't redefine.
1645 (STARTFILE_ARCH32_SPEC): Rename to ...
1646 (STARTFILE_ARCH_SPEC): ... this.
1647 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1648 * config/sparc/sol2.h: ... here.
1649 (SECTION_NAME_FORMAT): Don't undef.
1650 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1651 (SUBTARGET_EXTRA_SPECS): Remove.
1652 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1654 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1655 (MD_STARTFILE_PREFIX): Remove.
1656 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1657 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1658 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1659 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1660 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1661 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1662 * config/i386/sol2.h: ... here.
1663 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1664 * config/i386/sol2-bi.h: Remove.
1665 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1666 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1668 * config/i386/t-sol2-64: Rename to ...
1669 * config/i386/t-sol2: ... this.
1670 * config/sparc/t-sol2-64: Rename to ...
1671 * config/sparc/t-sol2: ... this.
1673 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1674 sol2_tm_file_head, sol2_tm_file_tail.
1675 Include ${cpu_type}/sol2.h before sol2.h.
1677 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1678 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1679 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1680 Reflect i386/t-sol2-64 renaming.
1681 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1682 Reflect sparc/t-sol2-64 renaming.
1684 2014-04-30 Richard Biener <rguenther@suse.de>
1686 * passes.c (execute_function_todo): Move TODO_verify_stmts
1687 and TODO_verify_ssa under the TODO_verify_il umbrella.
1688 * tree-ssa.h (verify_ssa): Adjust prototype.
1689 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1690 we should verify SSA operands.
1691 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1692 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1693 whether we should verify whether not throwing stmts have EH info.
1694 * graphite-scop-detection.c (create_sese_edges): Adjust.
1695 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1696 * tree-eh.c (lower_try_finally_switch): Do not add the
1697 default case label twice.
1699 2014-04-30 Marek Polacek <polacek@redhat.com>
1701 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1702 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1703 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1704 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1706 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1708 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1709 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1710 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1711 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1712 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1713 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1714 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1715 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1717 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1719 * tree-cfg.c (dump_function_to_file): Dump the return type of
1720 functions, in a line to itself before the function body, mimicking
1721 the layout of a C function.
1723 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1725 PR tree-optimization/60971
1726 * tree-tailcall.c (process_assignment): Reject conversions which
1729 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1731 * calls.c (initialize_argument_information): Always treat
1732 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1733 (expand_call): Likewise.
1734 (emit_library_call_calue_1): Likewise.
1735 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1736 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1739 2014-04-29 Nick Clifton <nickc@redhat.com>
1741 * config/msp430/msp430.md (umulsidi): Fix typo.
1742 (mulhisi3): Enable even inside interrupt handlers.
1743 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1744 bigger return address pushed in large mode.
1746 2014-04-29 Nick Clifton <nickc@redhat.com>
1748 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1749 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1751 * config/m32r/m32r.c (init_reg_tables): Likewise.
1752 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1753 enum to hold the modes.
1755 2014-04-29 Richard Biener <rguenther@suse.de>
1757 * dominance.c (free_dominance_info): Add overload with
1759 (dom_info_state): Likewise.
1760 (dom_info_available_p): Likewise.
1761 * basic-block.h (free_dominance_info, dom_info_state,
1762 dom_info_available_p): Declare overloads.
1763 * passes.c (execute_function_todo): Verify that verifiers
1764 don't change dominator info state. Drop dominator info
1765 for IPA pass invocations.
1766 * cgraph.c (release_function_body): Restore asserts that
1767 dominator information is released.
1769 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1771 * doc/invoke.texi: Fix typo.
1772 * tree-vrp.c: Fix typos.
1773 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1775 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1777 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1779 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1781 * config/aarch64/aarch64-builtins.c
1782 (aarch64_types_storestruct_lane_qualifiers): New.
1783 (TYPES_STORESTRUCT_LANE): Likewise.
1784 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1785 (st3_lane): Likewise.
1786 (st4_lane): Likewise.
1787 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1788 (vec_store_lanesci_lane<mode>): Likewise.
1789 (vec_store_lanesxi_lane<mode>): Likewise.
1790 (aarch64_st2_lane<VQ:mode>): Likewise.
1791 (aarch64_st3_lane<VQ:mode>): Likewise.
1792 (aarch64_st4_lane<VQ:mode>): Likewise.
1793 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1794 * config/aarch64/arm_neon.h
1795 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1796 use new macro arguments.
1797 (__ST3_LANE_FUNC): Likewise.
1798 (__ST4_LANE_FUNC): Likewise.
1799 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1800 (V_THREE_ELEM): Likewise.
1801 (V_FOUR_ELEM): Likewise.
1803 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1805 * doc/gimple.texi: Replace the description of the now-defunct
1806 union gimple_statement_d with a diagram showing the
1807 gimple_statement_base class hierarchy and its relationships to
1808 the GSS_ and GIMPLE_ enums.
1810 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1812 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1813 * config/aarch64/aarch64.c
1814 (aarch64_cannot_change_mode_class): Weaken conditions.
1815 (aarch64_modes_tieable_p): New.
1816 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1818 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1820 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1821 (loadsync_<mode>): Change mode.
1822 (load_quadpti, store_quadpti): New.
1823 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1824 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1826 2014-04-28 Martin Jambor <mjambor@suse.cz>
1828 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1829 same alias type as the original statement.
1830 (subreplacement_assignment_data): New type.
1831 (handle_unscalarized_data_in_subtree): New type of parameter,
1832 generate new memory accesses with same alias type as the original
1834 (load_assign_lhs_subreplacements): Likewise.
1835 (sra_modify_constructor_assign): Generate new memory accesses with
1836 same alias type as the original statement.
1838 2014-04-28 Richard Biener <rguenther@suse.de>
1840 * tree-pass.h (TODO_verify_il): Define.
1841 (TODO_verify_all): Complete properly.
1842 * passes.c (execute_function_todo): Move existing loop-closed
1843 SSA verification under TODO_verify_il.
1844 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1845 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1846 Fix tree sharing issue.
1848 2014-04-28 Richard Biener <rguenther@suse.de>
1851 * builtins.def (DEF_C11_BUILTIN): Add.
1852 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1853 * coretypes.h (enum function_class): Add function_c11_misc.
1854 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1855 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1856 (call_may_clobber_ref_p_1): Likewise.
1857 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1858 (mark_all_reaching_defs_necessary_1): Likewise.
1859 (propagate_necessity): Likewise.
1860 (eliminate_unnecessary_stmts): Likewise.
1861 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1863 2014-04-28 Richard Biener <rguenther@suse.de>
1865 * tree-vrp.c (vrp_var_may_overflow): Remove.
1866 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1867 with overflow immediately bump to one before that value and
1868 let iteration figure out overflow status.
1870 2014-04-28 Richard Biener <rguenther@suse.de>
1872 * configure.ac: Do valgrind header checks unconditionally.
1873 Add --enable-valgrind-annotations.
1874 * system.h: Guard valgrind header inclusion with
1875 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1876 * alloc-pool.c (pool_alloc, pool_free): Use
1877 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1878 to guard possibly dead code.
1879 * config.in: Regenerated.
1880 * configure: Likewise.
1882 2014-04-28 Jeff Law <law@redhat.com>
1884 PR tree-optimization/60902
1885 * tree-ssa-threadedge.c
1886 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1887 over real defs when invalidating outputs from statements that do not
1888 produce useful outputs for threading.
1890 2014-04-28 Richard Biener <rguenther@suse.de>
1892 PR tree-optimization/60979
1893 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1894 SCOPs that end in a block with a successor with abnormal
1897 2014-04-28 Richard Biener <rguenther@suse.de>
1899 * tree-pass.h (execute_pass_list): Adjust prototype.
1900 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1901 (do_per_function): Change callback signature, push all actual
1902 work to the callbals.
1903 (do_per_function_toporder): Likewise.
1904 (execute_function_dump): Adjust.
1905 (execute_function_todo): Likewise.
1906 (clear_last_verified): Likewise.
1907 (verify_curr_properties): Likewise.
1908 (update_properties_after_pass): Likewise.
1909 (execute_pass_list_1): Split out from ...
1910 (execute_pass_list): ... here. Adjust.
1911 (execute_ipa_pass_list): Likewise.
1912 * cgraphunit.c (cgraph_add_new_function): Adjust.
1913 (analyze_function): Likewise.
1914 (expand_function): Likewise.
1915 * cgraph.c (release_function_body): Free dominance info
1916 here instead of asserting it was magically freed elsewhere.
1918 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1920 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1921 * configure: Regenerate.
1922 * config/sparc/sparc.opt (muser-mode): New option.
1923 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1925 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1926 * doc/invoke.texi (SPARC options): Document -muser-mode.
1928 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1930 * cselib.c (find_slot_memmode): Delete.
1931 (cselib_hasher): Change compare_type to a struct.
1932 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1934 (preserve_constants_and_equivs): Adjust for new compare_type.
1935 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1936 (wrap_constant): Delete.
1937 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1939 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1941 * doc/install.texi (Building with profile feedback): Remove
1944 2014-04-26 Tom de Vries <tom@codesourcery.com>
1946 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1949 2014-04-25 Cary Coutant <ccoutant@google.com>
1952 * dwarf2out.c (should_move_die_to_comdat): A type definition
1953 can contain a subprogram definition, but don't move it to a
1955 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1956 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1958 (clone_tree_hash): Rename to...
1959 (clone_tree_partial): ...this; change callers. Copy
1960 DW_TAG_subprogram DIEs as declarations.
1961 (copy_decls_walk): Don't copy children of a declaration into a
1964 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1967 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1970 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1972 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1973 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1975 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1978 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1980 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1982 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1984 PR tree-optimization/60930
1985 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1986 creating a multiply candidate by folding two constant
1987 multiplicands when the result overflows.
1989 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1991 PR tree-optimization/60960
1992 * tree-vect-generic.c (expand_vector_operation): Only call
1993 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1995 2014-04-25 Tom de Vries <tom@codesourcery.com>
1997 * expr.c (clobber_reg_mode): New function.
1998 * expr.h (clobber_reg): New function.
2000 2014-04-25 Tom de Vries <tom@codesourcery.com>
2002 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2005 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2006 Tom de Vries <tom@codesourcery.com>
2008 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2010 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2011 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2012 new argument to find_all_hard_reg_sets call.
2014 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2016 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2017 Use HOST_WIDE_INT_C for mask literal.
2018 (aarch_rev16_shleft_mask_imm_p): Likewise.
2020 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2023 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2025 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2027 PR preprocessor/56540
2028 * config/i386/i386-c.c (ix86_target_macros): Define
2029 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2031 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2033 * configure.ac (tga_func): Remove.
2034 (LIB_TLS_SPEC): Remove.
2035 * configure: Regenerate.
2036 * config.in: Regenerate.
2037 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2039 2014-04-25 Richard Biener <rguenther@suse.de>
2042 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2043 call stmt use/clobber sets during stmt walk instead of
2044 walking the possibly incomplete set of caller edges.
2046 2014-04-25 Richard Biener <rguenther@suse.de>
2049 * passes.c (apply_ipa_transforms): Inline into only caller ...
2050 (execute_one_pass): ... here. Properly bring in function
2051 bodies for nodes we want to apply IPA transforms to.
2053 2014-04-24 Cong Hou <congh@google.com>
2055 PR tree-optimization/60896
2056 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2057 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2058 (vect_mark_pattern_stmts): Set the def type of all statements in
2059 PATTERN_DEF_SEQ as vect_internal_def.
2061 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2063 * doc/extend.texi (PowerPC Built-in Functions): Document new
2064 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2065 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2067 * config/rs6000/predicates.md (const_0_to_3_operand): New
2068 predicate to match 0..3 integer constants.
2070 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2071 to support adding miscellaneous builtin functions.
2072 (BU_DFP_MISC_2): Likewise.
2073 (BU_P7_MISC_1): Likewise.
2074 (BU_P7_MISC_2): Likewise.
2075 (BU_P8V_MISC_3): Likewise.
2076 (BU_MISC_1): Likewise.
2077 (BU_MISC_2): Likewise.
2078 (DIVWE): Add extended divide builtin functions.
2081 (DIVWEUO): Likewise.
2085 (DIVDEUO): Likewise.
2086 (DXEX): Add decimal floating-point builtin functions.
2089 (DDEDPDQ): Likewise.
2091 (DENBCDQ): Likewise.
2098 (CDTBCD): Add new BCD builtin functions.
2102 (BCDADD_LT): Likewise.
2103 (BCDADD_EQ): Likewise.
2104 (BCDADD_GT): Likewise.
2105 (BCDADD_OV): Likewise.
2107 (BCDSUB_LT): Likewise.
2108 (BCDSUB_EQ): Likewise.
2109 (BCDSUB_GT): Likewise.
2110 (BCDSUB_OV): Likewise.
2111 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2112 (UNPACK_TD): Likewise.
2113 (PACK_TF): Likewise.
2114 (UNPACK_TF): Likewise.
2115 (UNPACK_TF_0): Likewise.
2116 (UNPACK_TF_1): Likewise.
2117 (PACK_V1TI): Likewise.
2118 (UNPACK_V1TI): Likewise.
2120 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2121 support for decimal floating point builtin functions.
2122 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2123 functions that take constant arguments.
2124 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2125 (rs6000_init_builtins): Setup long double, _Decimal64, and
2126 _Decimal128 types for new builtin functions.
2127 (builtin_function_type): Set the unsigned flags appropriately for
2128 the new builtin functions.
2129 (rs6000_opt_masks): Add support for decimal floating point builtin
2132 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2133 floating point builtin functions.
2134 (RS6000_BTM_COMMON): Likewise.
2135 (RS6000_BTI_long_double): Likewise.
2136 (RS6000_BTI_dfloat64): Likewise.
2137 (RS6000_BTI_dfloat128): Likewise.
2138 (long_double_type_internal_node): Likewise.
2139 (dfloat64_type_internal_node): Likewise.
2140 (dfloat128_type_internal_node): Likewise.
2142 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2143 2.07 bcd arithmetic instructions.
2144 (UNSPEC_BCDSUB): Likewise.
2145 (UNSPEC_BCD_OVERFLOW): Likewise.
2146 (UNSPEC_BCD_ADD_SUB): Likewise.
2147 (bcd_add_sub): Likewise.
2148 (BCD_TEST): Likewise.
2149 (bcd<bcd_add_sub>): Likewise.
2150 (bcd<bcd_add_sub>_test): Likewise.
2151 (bcd<bcd_add_sub>_test2): Likewise.
2152 (bcd<bcd_add_sub>_<code>): Likewise.
2153 (peephole2 for combined bcd ops): Likewise.
2155 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2156 decimal floating point builtin functions.
2157 (UNSPEC_DENBCD): Likewise.
2158 (UNSPEC_DXEX): Likewise.
2159 (UNSPEC_DIEX): Likewise.
2160 (UNSPEC_DSCLI): Likewise.
2161 (UNSPEC_DSCRI): Likewise.
2162 (D64_D128): Likewise.
2163 (dfp_suffix): Likewise.
2164 (dfp_ddedpd_<mode>): Likewise.
2165 (dfp_denbcd_<mode>): Likewise.
2166 (dfp_dxex_<mode>): Likewise.
2167 (dfp_diex_<mode>): Likewise.
2168 (dfp_dscli_<mode>): Likewise.
2169 (dfp_dscri_<mode>): Likewise.
2171 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2173 (UNSPEC_CDTBCD): Likewise.
2174 (UNSPEC_CBCDTD): Likewise.
2175 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2176 (UNSPEC_DIVEO): Likewise.
2177 (UNSPEC_DIVEU): Likewise.
2178 (UNSPEC_DIVEUO): Likewise.
2179 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2180 pack/unpack 128-bit types.
2181 (UNSPEC_PACK_128BIT): Likewise.
2182 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2183 (udiv<mode>3): Use idiv_ldiv mode attribute.
2184 (div<mode>3): Likewise.
2185 (addg6s): Add new BCD builtin functions.
2188 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2189 (div_extend): Likewise.
2190 (div<div_extend>_<mode>"): Likewise.
2191 (FP128_64): Add support for new builtin functions to pack/unpack
2193 (unpack<mode>): Likewise.
2194 (unpacktf_0): Likewise.
2195 (unpacktf_1): Likewise.
2196 (unpack<mode>_dm): Likewise.
2197 (unpack<mode>_nodm): Likewise.
2198 (pack<mode>): Likewise.
2199 (unpackv1ti): Likewise.
2200 (packv1ti): Likewise.
2202 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2204 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2207 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2209 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2210 * gimplify.c (omp_is_private): Change last argument's type to int.
2211 Only diagnose lastprivate if the simd argument is 1, only diagnose
2212 linear if the simd argument is 2.
2213 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2214 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2215 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2216 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2217 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2218 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2219 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2220 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2221 * tree-nested.c (convert_nonlocal_omp_clauses,
2222 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2224 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2227 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2230 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2232 * flag-types.h (enum ivar_visibility): Add.
2234 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2236 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2237 function * argument.
2239 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2241 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2243 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2244 Tom de Vries <tom@codesourcery.com>
2246 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2247 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2249 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2250 * emit-rtl.c (try_split): Same.
2252 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2253 Tom de Vries <tom@codesourcery.com>
2255 * common.opt (fuse-caller-save): New option.
2257 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2259 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2260 elements for big-endian.
2262 2014-04-24 Richard Biener <rguenther@suse.de>
2264 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2265 during TER and instead use the sepops interface for expanding
2266 non-GIMPLE_SINGLE_RHS.
2268 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2270 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2271 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2273 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2275 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2276 assembler 64-bit option.
2277 * configure: Regenerate.
2279 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2281 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2282 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2283 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2284 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2285 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2287 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2289 * config/aarch64/aarch64-builtins.c
2290 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2291 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2292 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2293 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2295 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2296 (Vrevsuff): New mode attribute.
2298 2014-04-24 Terry Guo <terry.guo@arm.com>
2300 * config/arm/arm.h (machine_function): Define variable
2301 after_arm_reorg here.
2302 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2303 (arm_split_constant): Update the way to access variable
2306 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2308 2014-04-23 Tom de Vries <tom@codesourcery.com>
2310 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2312 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2314 * is-a.h: Update comments to reflect the following changes to the
2315 "pointerness" of the API, making the template parameter match the
2316 return type, allowing use of is-a.h with typedefs of pointers.
2317 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2318 that the return type matches the parameter to the is_a_helper.
2320 (dyn_cast): Likewise.
2322 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2323 pointer from the is-a.h API.
2325 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2326 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2328 (is_a_helper <varpool_node>::test): Likewise, convert to...
2329 (is_a_helper <varpool_node *>::test): ...this.
2331 (varpool_first_variable): Update for removal of implicit pointer
2332 from the is-a.h API.
2333 (varpool_next_variable): Likewise.
2334 (varpool_first_static_initializer): Likewise.
2335 (varpool_next_static_initializer): Likewise.
2336 (varpool_first_defined_variable): Likewise.
2337 (varpool_next_defined_variable): Likewise.
2338 (cgraph_first_defined_function): Likewise.
2339 (cgraph_next_defined_function): Likewise.
2340 (cgraph_first_function): Likewise.
2341 (cgraph_next_function): Likewise.
2342 (cgraph_first_function_with_gimple_body): Likewise.
2343 (cgraph_next_function_with_gimple_body): Likewise.
2344 (cgraph_alias_target): Likewise.
2345 (varpool_alias_target): Likewise.
2346 (cgraph_function_or_thunk_node): Likewise.
2347 (varpool_variable_node): Likewise.
2348 (symtab_real_symbol_p): Likewise.
2349 * cgraphunit.c (referred_to_p): Likewise.
2350 (analyze_functions): Likewise.
2351 (handle_alias_pairs): Likewise.
2352 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2353 * gimple-ssa.h (gimple_vuse_op): Likewise.
2354 (gimple_vdef_op): Likewise.
2355 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2356 * gimple.c (gimple_build_asm_1): Likewise.
2357 (gimple_build_try): Likewise.
2358 (gimple_build_resx): Likewise.
2359 (gimple_build_eh_dispatch): Likewise.
2360 (gimple_build_omp_for): Likewise.
2361 (gimple_omp_for_set_clauses): Likewise.
2363 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2364 (is_a_helper <gimple_statement_asm *>::test): ...this.
2365 (is_a_helper <gimple_statement_bind>::test): Convert to...
2366 (is_a_helper <gimple_statement_bind *>::test): ...this.
2367 (is_a_helper <gimple_statement_call>::test): Convert to...
2368 (is_a_helper <gimple_statement_call *>::test): ...this.
2369 (is_a_helper <gimple_statement_catch>::test): Convert to...
2370 (is_a_helper <gimple_statement_catch *>::test): ...this.
2371 (is_a_helper <gimple_statement_resx>::test): Convert to...
2372 (is_a_helper <gimple_statement_resx *>::test): ...this.
2373 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2374 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2375 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2376 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2377 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2378 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2379 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2380 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2381 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2382 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2383 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2384 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2385 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2386 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2387 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2388 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2389 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2390 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2391 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2392 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2393 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2394 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2395 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2396 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2397 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2398 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2399 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2400 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2401 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2402 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2403 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2404 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2405 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2406 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2407 (is_a_helper <gimple_statement_phi>::test): Convert to...
2408 (is_a_helper <gimple_statement_phi *>::test): ...this.
2409 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2410 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2411 (is_a_helper <gimple_statement_try>::test): Convert to...
2412 (is_a_helper <gimple_statement_try *>::test): ...this.
2413 (is_a_helper <gimple_statement_wce>::test): Convert to...
2414 (is_a_helper <gimple_statement_wce *>::test): ...this.
2415 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2416 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2417 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2418 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2419 (is_a_helper <const gimple_statement_call>::test): Convert to...
2420 (is_a_helper <const gimple_statement_call *>::test): ...this.
2421 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2422 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2423 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2424 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2425 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2426 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2427 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2428 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2429 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2431 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2433 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2435 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2437 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2438 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2439 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2441 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2442 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2444 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2445 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2446 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2447 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2448 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2449 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2451 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2452 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2453 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2454 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2456 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2457 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2458 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2459 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2460 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2461 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2462 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2463 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2464 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2465 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2466 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2467 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2468 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2469 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2470 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2471 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2473 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2475 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2476 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2478 (gimple_use_ops): Update for removal of implicit pointer from the
2480 (gimple_set_use_ops): Likewise.
2481 (gimple_vuse): Likewise.
2482 (gimple_vdef): Likewise.
2483 (gimple_vuse_ptr): Likewise.
2484 (gimple_vdef_ptr): Likewise.
2485 (gimple_set_vuse): Likewise.
2486 (gimple_set_vdef): Likewise.
2487 (gimple_omp_return_set_lhs): Likewise.
2488 (gimple_omp_return_lhs): Likewise.
2489 (gimple_omp_return_lhs_ptr): Likewise.
2490 (gimple_call_fntype): Likewise.
2491 (gimple_call_set_fntype): Likewise.
2492 (gimple_call_set_internal_fn): Likewise.
2493 (gimple_call_use_set): Likewise.
2494 (gimple_call_clobber_set): Likewise.
2495 (gimple_bind_vars): Likewise.
2496 (gimple_bind_set_vars): Likewise.
2497 (gimple_bind_body_ptr): Likewise.
2498 (gimple_bind_set_body): Likewise.
2499 (gimple_bind_add_stmt): Likewise.
2500 (gimple_bind_block): Likewise.
2501 (gimple_bind_set_block): Likewise.
2502 (gimple_asm_ninputs): Likewise.
2503 (gimple_asm_noutputs): Likewise.
2504 (gimple_asm_nclobbers): Likewise.
2505 (gimple_asm_nlabels): Likewise.
2506 (gimple_asm_input_op): Likewise.
2507 (gimple_asm_input_op_ptr): Likewise.
2508 (gimple_asm_output_op): Likewise.
2509 (gimple_asm_output_op_ptr): Likewise.
2510 (gimple_asm_set_output_op): Likewise.
2511 (gimple_asm_clobber_op): Likewise.
2512 (gimple_asm_set_clobber_op): Likewise.
2513 (gimple_asm_label_op): Likewise.
2514 (gimple_asm_set_label_op): Likewise.
2515 (gimple_asm_string): Likewise.
2516 (gimple_catch_types): Likewise.
2517 (gimple_catch_types_ptr): Likewise.
2518 (gimple_catch_handler_ptr): Likewise.
2519 (gimple_catch_set_types): Likewise.
2520 (gimple_catch_set_handler): Likewise.
2521 (gimple_eh_filter_types): Likewise.
2522 (gimple_eh_filter_types_ptr): Likewise.
2523 (gimple_eh_filter_failure_ptr): Likewise.
2524 (gimple_eh_filter_set_types): Likewise.
2525 (gimple_eh_filter_set_failure): Likewise.
2526 (gimple_eh_must_not_throw_fndecl): Likewise.
2527 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2528 (gimple_eh_else_n_body_ptr): Likewise.
2529 (gimple_eh_else_e_body_ptr): Likewise.
2530 (gimple_eh_else_set_n_body): Likewise.
2531 (gimple_eh_else_set_e_body): Likewise.
2532 (gimple_try_eval_ptr): Likewise.
2533 (gimple_try_cleanup_ptr): Likewise.
2534 (gimple_try_set_eval): Likewise.
2535 (gimple_try_set_cleanup): Likewise.
2536 (gimple_wce_cleanup_ptr): Likewise.
2537 (gimple_wce_set_cleanup): Likewise.
2538 (gimple_phi_capacity): Likewise.
2539 (gimple_phi_num_args): Likewise.
2540 (gimple_phi_result): Likewise.
2541 (gimple_phi_result_ptr): Likewise.
2542 (gimple_phi_set_result): Likewise.
2543 (gimple_phi_arg): Likewise.
2544 (gimple_phi_set_arg): Likewise.
2545 (gimple_resx_region): Likewise.
2546 (gimple_resx_set_region): Likewise.
2547 (gimple_eh_dispatch_region): Likewise.
2548 (gimple_eh_dispatch_set_region): Likewise.
2549 (gimple_omp_critical_name): Likewise.
2550 (gimple_omp_critical_name_ptr): Likewise.
2551 (gimple_omp_critical_set_name): Likewise.
2552 (gimple_omp_for_clauses): Likewise.
2553 (gimple_omp_for_clauses_ptr): Likewise.
2554 (gimple_omp_for_set_clauses): Likewise.
2555 (gimple_omp_for_collapse): Likewise.
2556 (gimple_omp_for_index): Likewise.
2557 (gimple_omp_for_index_ptr): Likewise.
2558 (gimple_omp_for_set_index): Likewise.
2559 (gimple_omp_for_initial): Likewise.
2560 (gimple_omp_for_initial_ptr): Likewise.
2561 (gimple_omp_for_set_initial): Likewise.
2562 (gimple_omp_for_final): Likewise.
2563 (gimple_omp_for_final_ptr): Likewise.
2564 (gimple_omp_for_set_final): Likewise.
2565 (gimple_omp_for_incr): Likewise.
2566 (gimple_omp_for_incr_ptr): Likewise.
2567 (gimple_omp_for_set_incr): Likewise.
2568 (gimple_omp_for_pre_body_ptr): Likewise.
2569 (gimple_omp_for_set_pre_body): Likewise.
2570 (gimple_omp_parallel_clauses): Likewise.
2571 (gimple_omp_parallel_clauses_ptr): Likewise.
2572 (gimple_omp_parallel_set_clauses): Likewise.
2573 (gimple_omp_parallel_child_fn): Likewise.
2574 (gimple_omp_parallel_child_fn_ptr): Likewise.
2575 (gimple_omp_parallel_set_child_fn): Likewise.
2576 (gimple_omp_parallel_data_arg): Likewise.
2577 (gimple_omp_parallel_data_arg_ptr): Likewise.
2578 (gimple_omp_parallel_set_data_arg): Likewise.
2579 (gimple_omp_task_clauses): Likewise.
2580 (gimple_omp_task_clauses_ptr): Likewise.
2581 (gimple_omp_task_set_clauses): Likewise.
2582 (gimple_omp_task_child_fn): Likewise.
2583 (gimple_omp_task_child_fn_ptr): Likewise.
2584 (gimple_omp_task_set_child_fn): Likewise.
2585 (gimple_omp_task_data_arg): Likewise.
2586 (gimple_omp_task_data_arg_ptr): Likewise.
2587 (gimple_omp_task_set_data_arg): Likewise.
2588 (gimple_omp_taskreg_clauses): Likewise.
2589 (gimple_omp_taskreg_clauses_ptr): Likewise.
2590 (gimple_omp_taskreg_set_clauses): Likewise.
2591 (gimple_omp_taskreg_child_fn): Likewise.
2592 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2593 (gimple_omp_taskreg_set_child_fn): Likewise.
2594 (gimple_omp_taskreg_data_arg): Likewise.
2595 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2596 (gimple_omp_taskreg_set_data_arg): Likewise.
2597 (gimple_omp_task_copy_fn): Likewise.
2598 (gimple_omp_task_copy_fn_ptr): Likewise.
2599 (gimple_omp_task_set_copy_fn): Likewise.
2600 (gimple_omp_task_arg_size): Likewise.
2601 (gimple_omp_task_arg_size_ptr): Likewise.
2602 (gimple_omp_task_set_arg_size): Likewise.
2603 (gimple_omp_task_arg_align): Likewise.
2604 (gimple_omp_task_arg_align_ptr): Likewise.
2605 (gimple_omp_task_set_arg_align): Likewise.
2606 (gimple_omp_single_clauses): Likewise.
2607 (gimple_omp_single_clauses_ptr): Likewise.
2608 (gimple_omp_single_set_clauses): Likewise.
2609 (gimple_omp_target_clauses): Likewise.
2610 (gimple_omp_target_clauses_ptr): Likewise.
2611 (gimple_omp_target_set_clauses): Likewise.
2612 (gimple_omp_target_child_fn): Likewise.
2613 (gimple_omp_target_child_fn_ptr): Likewise.
2614 (gimple_omp_target_set_child_fn): Likewise.
2615 (gimple_omp_target_data_arg): Likewise.
2616 (gimple_omp_target_data_arg_ptr): Likewise.
2617 (gimple_omp_target_set_data_arg): Likewise.
2618 (gimple_omp_teams_clauses): Likewise.
2619 (gimple_omp_teams_clauses_ptr): Likewise.
2620 (gimple_omp_teams_set_clauses): Likewise.
2621 (gimple_omp_sections_clauses): Likewise.
2622 (gimple_omp_sections_clauses_ptr): Likewise.
2623 (gimple_omp_sections_set_clauses): Likewise.
2624 (gimple_omp_sections_control): Likewise.
2625 (gimple_omp_sections_control_ptr): Likewise.
2626 (gimple_omp_sections_set_control): Likewise.
2627 (gimple_omp_for_set_cond): Likewise.
2628 (gimple_omp_for_cond): Likewise.
2629 (gimple_omp_atomic_store_set_val): Likewise.
2630 (gimple_omp_atomic_store_val): Likewise.
2631 (gimple_omp_atomic_store_val_ptr): Likewise.
2632 (gimple_omp_atomic_load_set_lhs): Likewise.
2633 (gimple_omp_atomic_load_lhs): Likewise.
2634 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2635 (gimple_omp_atomic_load_set_rhs): Likewise.
2636 (gimple_omp_atomic_load_rhs): Likewise.
2637 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2638 (gimple_omp_continue_control_def): Likewise.
2639 (gimple_omp_continue_control_def_ptr): Likewise.
2640 (gimple_omp_continue_set_control_def): Likewise.
2641 (gimple_omp_continue_control_use): Likewise.
2642 (gimple_omp_continue_control_use_ptr): Likewise.
2643 (gimple_omp_continue_set_control_use): Likewise.
2644 (gimple_transaction_body_ptr): Likewise.
2645 (gimple_transaction_label): Likewise.
2646 (gimple_transaction_label_ptr): Likewise.
2647 (gimple_transaction_set_body): Likewise.
2648 (gimple_transaction_set_label): Likewise.
2650 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2651 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2652 * ipa-ref.c (ipa_record_reference): Likewise.
2653 * ipa-reference.c (analyze_function): Likewise.
2654 (ipa_reference_write_optimization_summary): Likewise.
2655 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2656 (address_taken_from_non_vtable_p): Likewise.
2657 (comdat_can_be_unshared_p_1): Likewise.
2658 * lto-cgraph.c (lto_output_ref): Likewise.
2659 (add_references): Likewise.
2660 (compute_ltrans_boundary): Likewise.
2661 (output_symtab): Likewise.
2662 (input_ref): Likewise.
2663 (input_cgraph_1): Likewise.
2664 (output_cgraph_opt_summary): Likewise.
2665 * lto-streamer-out.c (lto_output): Likewise.
2666 (output_symbol_p): Likewise.
2667 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2668 (lsei_start_function_in_partition): Likewise.
2669 (lsei_next_variable_in_partition): Likewise.
2670 (lsei_start_variable_in_partition): Likewise.
2671 * symtab.c (insert_to_assembler_name_hash): Likewise.
2672 (unlink_from_assembler_name_hash): Likewise.
2673 (symtab_unregister_node): Likewise.
2674 (symtab_remove_node): Likewise.
2675 (dump_symtab_node): Likewise.
2676 (verify_symtab_base): Likewise.
2677 (verify_symtab_node): Likewise.
2678 (symtab_make_decl_local): Likewise.
2679 (symtab_alias_ultimate_target): Likewise.
2680 (symtab_resolve_alias): Likewise.
2681 (symtab_get_symbol_partitioning_class): Likewise.
2682 * tree-phinodes.c (allocate_phi_node): Likewise.
2683 (reserve_phi_args_for_new_edge): Likewise.
2684 (remove_phi_args): Likewise.
2685 * varpool.c (varpool_node_for_asm): Likewise.
2686 (varpool_remove_unreferenced_decls): Likewise.
2688 2014-04-23 Jeff Law <law@redhat.com>
2690 PR tree-optimization/60902
2691 * tree-ssa-threadedge.c
2692 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2693 invalidate outputs from statements that do not produce useful
2694 outputs for threading.
2696 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2698 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2699 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2700 machine descriptions for Stack Smashing Protector.
2702 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2704 * aarch64.md (<optab>_rol<mode>3): New pattern.
2705 (<optab>_rolsi3_uxtw): Likewise.
2706 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2708 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2710 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2711 (arm_cortex_a12_tune): Likewise.
2713 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2715 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2717 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2719 * config/arm/arm.md (arm_rev16si2): New pattern.
2720 (arm_rev16si2_alt): Likewise.
2721 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2723 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2725 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2726 (rev16<mode>2_alt): Likewise.
2727 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2728 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2729 (aarch_rev16_shleft_mask_imm_p): Likewise.
2730 (aarch_rev16_p_1): Likewise.
2731 (aarch_rev16_p): Likewise.
2732 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2733 (aarch_rev16_shright_mask_imm_p): Likewise.
2734 (aarch_rev16_shleft_mask_imm_p): Likewise.
2736 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2738 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2739 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2741 (cortex_a53_extra_costs): Likewise.
2742 (cortex_a57_extra_costs): Likewise.
2743 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2744 (cortexa7_extra_costs): Likewise.
2745 (cortexa8_extra_costs): Likewise.
2746 (cortexa12_extra_costs): Likewise.
2747 (cortexa15_extra_costs): Likewise.
2748 (v7m_extra_costs): Likewise.
2749 (arm_new_rtx_costs): Handle BSWAP.
2751 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2753 * config/arm/arm.c (cortexa8_extra_costs): New table.
2754 (arm_cortex_a8_tune): New tuning struct.
2755 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2757 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2759 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2761 2014-04-23 Richard Biener <rguenther@suse.de>
2763 * Makefile.in (OBJS): Remove loop-unswitch.o.
2764 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2765 * passes.def (pass_rtl_unswitch): Likewise.
2766 * loop-init.c (gate_rtl_unswitch): Likewise.
2767 (rtl_unswitch): Likewise.
2768 (pass_data_rtl_unswitch): Likewise.
2769 (pass_rtl_unswitch): Likewise.
2770 (make_pass_rtl_unswitch): Likewise.
2771 * rtl.h (reversed_condition): Likewise.
2772 (compare_and_jump_seq): Likewise.
2773 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2775 * loop-unroll.c (compare_and_jump_seq): Likewise.
2777 2014-04-23 Richard Biener <rguenther@suse.de>
2779 PR tree-optimization/60903
2780 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2781 commented code block.
2782 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2783 loop flags to newly created BBs and edges.
2785 2014-04-23 Nick Clifton <nickc@redhat.com>
2787 * config/msp430/msp430.c (msp430_handle_option): Move function
2789 (msp430_option_override): Simplify mcu and mcpu option handling.
2790 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2791 support for -mhwmult command line option.
2792 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2793 -mhwmult command line option.
2794 (msp430_hwmult_enabled): Delete.
2795 (msp43o_output_labelref): Add support for -mhwmult command line option.
2796 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2797 (umulsidi3): Likewise.
2798 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2799 (mcpu, mlarge, msmall): Likewise.
2800 (mhwmult): New option.
2801 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2803 (msp430_is_f5_mcu): Remove prototype.
2804 (msp430_use_f5_series_hwmult): Add prototype.
2805 * config/msp430/msp430-opts.h: New file.
2806 * common/config/msp430: New directory.
2807 * common/config/msp430/msp430-common.c: New file.
2808 * config.gcc (msp430): Remove target_has_targetm_common.
2809 * doc/invoke.texi: Document -mhwmult command line option.
2811 2014-04-23 Nick Clifton <nickc@redhat.com>
2813 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2814 default-manifest.o if it can be found in the search path.
2815 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2817 2014-04-23 Terry Guo <terry.guo@arm.com>
2819 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2821 2014-04-23 Richard Biener <rguenther@suse.de>
2824 * tree-inline.c (declare_return_variable): Use mark_addressable.
2826 2014-04-23 Richard Biener <rguenther@suse.de>
2829 * loop-init.c (loop_optimizer_init): Make sure to apply
2830 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2832 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2835 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2837 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2838 if flag_sanitize_undefined_trap_on_error.
2839 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2840 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2841 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2842 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2843 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2844 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2845 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2846 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2847 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2848 * ubsan.c (ubsan_instrument_unreachable): Return
2849 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2850 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2851 if flag_sanitize_undefined_trap_on_error and
2852 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2853 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2854 instrument_bool_enum_load): Emit __builtin_trap () if
2855 flag_sanitize_undefined_trap_on_error and
2856 __builtin_handle_*_abort () if !flag_sanitize_recover.
2857 * doc/invoke.texi (-fsanitize-recover,
2858 -fsanitize-undefined-trap-on-error): Document.
2860 2014-04-22 Christian Bruel <christian.bruel@st.com>
2862 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2863 Force immediates to SImode.
2865 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2867 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2869 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2870 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2871 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2872 (nios2_fpu_insn): Add entry for round.
2873 (N2FPU_NO_ERRNO_P): Define.
2874 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2876 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2878 2014-04-22 Richard Henderson <rth@redhat.com>
2880 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2881 (add<GPI>3_compare0): Remove leading * from name.
2882 (add<GPI>3_carryin): Likewise.
2883 (sub<GPI>3_compare0): Likewise.
2884 (sub<GPI>3_carryin): Likewise.
2885 (<su_optab>mulditi3): New expander.
2886 (multi3): New expander.
2887 (madd<GPI>): Remove leading * from name.
2889 2014-04-22 Martin Jambor <mjambor@suse.cz>
2891 * cgraphclones.c (cgraph_function_versioning): Copy
2892 ipa_transforms_to_apply instead of asserting it is empty.
2894 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2897 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2898 on count_exp to get mode.
2900 2014-04-22 Andrew Pinski <apinski@cavium.com>
2902 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2903 Handle TLS for ILP32.
2904 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2905 (tlsie_small_<mode>): this and handle PTR.
2906 (tlsie_small_sidi): New pattern.
2907 (tlsle_small): Change to an expand to handle ILP32.
2908 (tlsle_small_<mode>): New pattern.
2909 (tlsdesc_small): Rename to ...
2910 (tlsdesc_small_<mode>): this and handle PTR.
2912 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2914 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2916 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2918 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2919 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2920 (aarch64_types_signed_poly_qualifiers): Likewise.
2921 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2922 (aarch64_types_poly_signed_qualifiers): Likewise.
2923 (TYPES_REINTERP_SS): Type macro added.
2924 (TYPES_REINTERP_SU): Likewise.
2925 (TYPES_REINTERP_SP): Likewise.
2926 (TYPES_REINTERP_US): Likewise.
2927 (TYPES_REINTERP_PS): Likewise.
2928 (aarch64_fold_builtin): New expression folding added.
2929 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2930 Declarations removed.
2931 (REINTERP_SS): Declarations added.
2932 (REINTERP_US): Likewise.
2933 (REINTERP_PS): Likewise.
2934 (REINTERP_SU): Likewise.
2935 (REINTERP_SP): Likewise.
2936 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2937 (vreinterpretq_p8_f64): Likewise.
2938 (vreinterpret_p16_f64): Likewise.
2939 (vreinterpretq_p16_f64): Likewise.
2940 (vreinterpret_f32_f64): Likewise.
2941 (vreinterpretq_f32_f64): Likewise.
2942 (vreinterpret_f64_f32): Likewise.
2943 (vreinterpret_f64_p8): Likewise.
2944 (vreinterpret_f64_p16): Likewise.
2945 (vreinterpret_f64_s8): Likewise.
2946 (vreinterpret_f64_s16): Likewise.
2947 (vreinterpret_f64_s32): Likewise.
2948 (vreinterpret_f64_s64): Likewise.
2949 (vreinterpret_f64_u8): Likewise.
2950 (vreinterpret_f64_u16): Likewise.
2951 (vreinterpret_f64_u32): Likewise.
2952 (vreinterpret_f64_u64): Likewise.
2953 (vreinterpretq_f64_f32): Likewise.
2954 (vreinterpretq_f64_p8): Likewise.
2955 (vreinterpretq_f64_p16): Likewise.
2956 (vreinterpretq_f64_s8): Likewise.
2957 (vreinterpretq_f64_s16): Likewise.
2958 (vreinterpretq_f64_s32): Likewise.
2959 (vreinterpretq_f64_s64): Likewise.
2960 (vreinterpretq_f64_u8): Likewise.
2961 (vreinterpretq_f64_u16): Likewise.
2962 (vreinterpretq_f64_u32): Likewise.
2963 (vreinterpretq_f64_u64): Likewise.
2964 (vreinterpret_s64_f64): Likewise.
2965 (vreinterpretq_s64_f64): Likewise.
2966 (vreinterpret_u64_f64): Likewise.
2967 (vreinterpretq_u64_f64): Likewise.
2968 (vreinterpret_s8_f64): Likewise.
2969 (vreinterpretq_s8_f64): Likewise.
2970 (vreinterpret_s16_f64): Likewise.
2971 (vreinterpretq_s16_f64): Likewise.
2972 (vreinterpret_s32_f64): Likewise.
2973 (vreinterpretq_s32_f64): Likewise.
2974 (vreinterpret_u8_f64): Likewise.
2975 (vreinterpretq_u8_f64): Likewise.
2976 (vreinterpret_u16_f64): Likewise.
2977 (vreinterpretq_u16_f64): Likewise.
2978 (vreinterpret_u32_f64): Likewise.
2979 (vreinterpretq_u32_f64): Likewise.
2981 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2983 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2984 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2985 (vreinterpret_p8_s8): Likewise.
2986 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2987 (vreinterpret_p8_s16): Likewise.
2988 (vreinterpret_p8_s32): Likewise.
2989 (vreinterpret_p8_s64): Likewise.
2990 (vreinterpret_p8_f32): Likewise.
2991 (vreinterpret_p8_u8): Likewise.
2992 (vreinterpret_p8_u16): Likewise.
2993 (vreinterpret_p8_u32): Likewise.
2994 (vreinterpret_p8_u64): Likewise.
2995 (vreinterpret_p8_p16): Likewise.
2996 (vreinterpretq_p8_s8): Likewise.
2997 (vreinterpretq_p8_s16): Likewise.
2998 (vreinterpretq_p8_s32): Likewise.
2999 (vreinterpretq_p8_s64): Likewise.
3000 (vreinterpretq_p8_f32): Likewise.
3001 (vreinterpretq_p8_u8): Likewise.
3002 (vreinterpretq_p8_u16): Likewise.
3003 (vreinterpretq_p8_u32): Likewise.
3004 (vreinterpretq_p8_u64): Likewise.
3005 (vreinterpretq_p8_p16): Likewise.
3006 (vreinterpret_p16_s8): Likewise.
3007 (vreinterpret_p16_s16): Likewise.
3008 (vreinterpret_p16_s32): Likewise.
3009 (vreinterpret_p16_s64): Likewise.
3010 (vreinterpret_p16_f32): Likewise.
3011 (vreinterpret_p16_u8): Likewise.
3012 (vreinterpret_p16_u16): Likewise.
3013 (vreinterpret_p16_u32): Likewise.
3014 (vreinterpret_p16_u64): Likewise.
3015 (vreinterpret_p16_p8): Likewise.
3016 (vreinterpretq_p16_s8): Likewise.
3017 (vreinterpretq_p16_s16): Likewise.
3018 (vreinterpretq_p16_s32): Likewise.
3019 (vreinterpretq_p16_s64): Likewise.
3020 (vreinterpretq_p16_f32): Likewise.
3021 (vreinterpretq_p16_u8): Likewise.
3022 (vreinterpretq_p16_u16): Likewise.
3023 (vreinterpretq_p16_u32): Likewise.
3024 (vreinterpretq_p16_u64): Likewise.
3025 (vreinterpretq_p16_p8): Likewise.
3026 (vreinterpret_f32_s8): Likewise.
3027 (vreinterpret_f32_s16): Likewise.
3028 (vreinterpret_f32_s32): Likewise.
3029 (vreinterpret_f32_s64): Likewise.
3030 (vreinterpret_f32_u8): Likewise.
3031 (vreinterpret_f32_u16): Likewise.
3032 (vreinterpret_f32_u32): Likewise.
3033 (vreinterpret_f32_u64): Likewise.
3034 (vreinterpret_f32_p8): Likewise.
3035 (vreinterpret_f32_p16): Likewise.
3036 (vreinterpretq_f32_s8): Likewise.
3037 (vreinterpretq_f32_s16): Likewise.
3038 (vreinterpretq_f32_s32): Likewise.
3039 (vreinterpretq_f32_s64): Likewise.
3040 (vreinterpretq_f32_u8): Likewise.
3041 (vreinterpretq_f32_u16): Likewise.
3042 (vreinterpretq_f32_u32): Likewise.
3043 (vreinterpretq_f32_u64): Likewise.
3044 (vreinterpretq_f32_p8): Likewise.
3045 (vreinterpretq_f32_p16): Likewise.
3046 (vreinterpret_s64_s8): Likewise.
3047 (vreinterpret_s64_s16): Likewise.
3048 (vreinterpret_s64_s32): Likewise.
3049 (vreinterpret_s64_f32): Likewise.
3050 (vreinterpret_s64_u8): Likewise.
3051 (vreinterpret_s64_u16): Likewise.
3052 (vreinterpret_s64_u32): Likewise.
3053 (vreinterpret_s64_u64): Likewise.
3054 (vreinterpret_s64_p8): Likewise.
3055 (vreinterpret_s64_p16): Likewise.
3056 (vreinterpretq_s64_s8): Likewise.
3057 (vreinterpretq_s64_s16): Likewise.
3058 (vreinterpretq_s64_s32): Likewise.
3059 (vreinterpretq_s64_f32): Likewise.
3060 (vreinterpretq_s64_u8): Likewise.
3061 (vreinterpretq_s64_u16): Likewise.
3062 (vreinterpretq_s64_u32): Likewise.
3063 (vreinterpretq_s64_u64): Likewise.
3064 (vreinterpretq_s64_p8): Likewise.
3065 (vreinterpretq_s64_p16): Likewise.
3066 (vreinterpret_u64_s8): Likewise.
3067 (vreinterpret_u64_s16): Likewise.
3068 (vreinterpret_u64_s32): Likewise.
3069 (vreinterpret_u64_s64): Likewise.
3070 (vreinterpret_u64_f32): Likewise.
3071 (vreinterpret_u64_u8): Likewise.
3072 (vreinterpret_u64_u16): Likewise.
3073 (vreinterpret_u64_u32): Likewise.
3074 (vreinterpret_u64_p8): Likewise.
3075 (vreinterpret_u64_p16): Likewise.
3076 (vreinterpretq_u64_s8): Likewise.
3077 (vreinterpretq_u64_s16): Likewise.
3078 (vreinterpretq_u64_s32): Likewise.
3079 (vreinterpretq_u64_s64): Likewise.
3080 (vreinterpretq_u64_f32): Likewise.
3081 (vreinterpretq_u64_u8): Likewise.
3082 (vreinterpretq_u64_u16): Likewise.
3083 (vreinterpretq_u64_u32): Likewise.
3084 (vreinterpretq_u64_p8): Likewise.
3085 (vreinterpretq_u64_p16): Likewise.
3086 (vreinterpret_s8_s16): Likewise.
3087 (vreinterpret_s8_s32): Likewise.
3088 (vreinterpret_s8_s64): Likewise.
3089 (vreinterpret_s8_f32): Likewise.
3090 (vreinterpret_s8_u8): Likewise.
3091 (vreinterpret_s8_u16): Likewise.
3092 (vreinterpret_s8_u32): Likewise.
3093 (vreinterpret_s8_u64): Likewise.
3094 (vreinterpret_s8_p8): Likewise.
3095 (vreinterpret_s8_p16): Likewise.
3096 (vreinterpretq_s8_s16): Likewise.
3097 (vreinterpretq_s8_s32): Likewise.
3098 (vreinterpretq_s8_s64): Likewise.
3099 (vreinterpretq_s8_f32): Likewise.
3100 (vreinterpretq_s8_u8): Likewise.
3101 (vreinterpretq_s8_u16): Likewise.
3102 (vreinterpretq_s8_u32): Likewise.
3103 (vreinterpretq_s8_u64): Likewise.
3104 (vreinterpretq_s8_p8): Likewise.
3105 (vreinterpretq_s8_p16): Likewise.
3106 (vreinterpret_s16_s8): Likewise.
3107 (vreinterpret_s16_s32): Likewise.
3108 (vreinterpret_s16_s64): Likewise.
3109 (vreinterpret_s16_f32): Likewise.
3110 (vreinterpret_s16_u8): Likewise.
3111 (vreinterpret_s16_u16): Likewise.
3112 (vreinterpret_s16_u32): Likewise.
3113 (vreinterpret_s16_u64): Likewise.
3114 (vreinterpret_s16_p8): Likewise.
3115 (vreinterpret_s16_p16): Likewise.
3116 (vreinterpretq_s16_s8): Likewise.
3117 (vreinterpretq_s16_s32): Likewise.
3118 (vreinterpretq_s16_s64): Likewise.
3119 (vreinterpretq_s16_f32): Likewise.
3120 (vreinterpretq_s16_u8): Likewise.
3121 (vreinterpretq_s16_u16): Likewise.
3122 (vreinterpretq_s16_u32): Likewise.
3123 (vreinterpretq_s16_u64): Likewise.
3124 (vreinterpretq_s16_p8): Likewise.
3125 (vreinterpretq_s16_p16): Likewise.
3126 (vreinterpret_s32_s8): Likewise.
3127 (vreinterpret_s32_s16): Likewise.
3128 (vreinterpret_s32_s64): Likewise.
3129 (vreinterpret_s32_f32): Likewise.
3130 (vreinterpret_s32_u8): Likewise.
3131 (vreinterpret_s32_u16): Likewise.
3132 (vreinterpret_s32_u32): Likewise.
3133 (vreinterpret_s32_u64): Likewise.
3134 (vreinterpret_s32_p8): Likewise.
3135 (vreinterpret_s32_p16): Likewise.
3136 (vreinterpretq_s32_s8): Likewise.
3137 (vreinterpretq_s32_s16): Likewise.
3138 (vreinterpretq_s32_s64): Likewise.
3139 (vreinterpretq_s32_f32): Likewise.
3140 (vreinterpretq_s32_u8): Likewise.
3141 (vreinterpretq_s32_u16): Likewise.
3142 (vreinterpretq_s32_u32): Likewise.
3143 (vreinterpretq_s32_u64): Likewise.
3144 (vreinterpretq_s32_p8): Likewise.
3145 (vreinterpretq_s32_p16): Likewise.
3146 (vreinterpret_u8_s8): Likewise.
3147 (vreinterpret_u8_s16): Likewise.
3148 (vreinterpret_u8_s32): Likewise.
3149 (vreinterpret_u8_s64): Likewise.
3150 (vreinterpret_u8_f32): Likewise.
3151 (vreinterpret_u8_u16): Likewise.
3152 (vreinterpret_u8_u32): Likewise.
3153 (vreinterpret_u8_u64): Likewise.
3154 (vreinterpret_u8_p8): Likewise.
3155 (vreinterpret_u8_p16): Likewise.
3156 (vreinterpretq_u8_s8): Likewise.
3157 (vreinterpretq_u8_s16): Likewise.
3158 (vreinterpretq_u8_s32): Likewise.
3159 (vreinterpretq_u8_s64): Likewise.
3160 (vreinterpretq_u8_f32): Likewise.
3161 (vreinterpretq_u8_u16): Likewise.
3162 (vreinterpretq_u8_u32): Likewise.
3163 (vreinterpretq_u8_u64): Likewise.
3164 (vreinterpretq_u8_p8): Likewise.
3165 (vreinterpretq_u8_p16): Likewise.
3166 (vreinterpret_u16_s8): Likewise.
3167 (vreinterpret_u16_s16): Likewise.
3168 (vreinterpret_u16_s32): Likewise.
3169 (vreinterpret_u16_s64): Likewise.
3170 (vreinterpret_u16_f32): Likewise.
3171 (vreinterpret_u16_u8): Likewise.
3172 (vreinterpret_u16_u32): Likewise.
3173 (vreinterpret_u16_u64): Likewise.
3174 (vreinterpret_u16_p8): Likewise.
3175 (vreinterpret_u16_p16): Likewise.
3176 (vreinterpretq_u16_s8): Likewise.
3177 (vreinterpretq_u16_s16): Likewise.
3178 (vreinterpretq_u16_s32): Likewise.
3179 (vreinterpretq_u16_s64): Likewise.
3180 (vreinterpretq_u16_f32): Likewise.
3181 (vreinterpretq_u16_u8): Likewise.
3182 (vreinterpretq_u16_u32): Likewise.
3183 (vreinterpretq_u16_u64): Likewise.
3184 (vreinterpretq_u16_p8): Likewise.
3185 (vreinterpretq_u16_p16): Likewise.
3186 (vreinterpret_u32_s8): Likewise.
3187 (vreinterpret_u32_s16): Likewise.
3188 (vreinterpret_u32_s32): Likewise.
3189 (vreinterpret_u32_s64): Likewise.
3190 (vreinterpret_u32_f32): Likewise.
3191 (vreinterpret_u32_u8): Likewise.
3192 (vreinterpret_u32_u16): Likewise.
3193 (vreinterpret_u32_u64): Likewise.
3194 (vreinterpret_u32_p8): Likewise.
3195 (vreinterpret_u32_p16): Likewise.
3196 (vreinterpretq_u32_s8): Likewise.
3197 (vreinterpretq_u32_s16): Likewise.
3198 (vreinterpretq_u32_s32): Likewise.
3199 (vreinterpretq_u32_s64): Likewise.
3200 (vreinterpretq_u32_f32): Likewise.
3201 (vreinterpretq_u32_u8): Likewise.
3202 (vreinterpretq_u32_u16): Likewise.
3203 (vreinterpretq_u32_u64): Likewise.
3204 (vreinterpretq_u32_p8): Likewise.
3205 (vreinterpretq_u32_p16): Likewise.
3207 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3209 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3211 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3213 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3214 (vqnegd_s64): Likewise.
3215 (vqabs_s64): Likewise.
3216 (vqabsd_s64): Likewise.
3218 2014-04-22 Richard Henderson <rth@redhat.com>
3220 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3221 computation to the top of the loop.
3223 2014-04-22 Renlin <renlin.li@arm.com>
3224 Jiong Wang <jiong.wang@arm.com>
3226 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3227 * config/aarch64/aarch64.c (aarch64_layout_frame)
3228 (aarch64_initial_elimination_offset): Likewise.
3230 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3232 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3235 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3237 * machmode.h (bitwise_mode_for_mode): Declare.
3238 * stor-layout.h (bitwise_type_for_mode): Likewise.
3239 * stor-layout.c (bitwise_mode_for_mode): New function.
3240 (bitwise_type_for_mode): Likewise.
3241 * builtins.c (fold_builtin_memory_op): Use it instead of
3242 int_mode_for_mode and build_nonstandard_integer_type.
3244 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3246 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3247 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3248 (*-*-solaris2*): Simplify.
3249 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3250 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3251 *-*-solaris2.9* handling.
3253 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3255 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3256 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3258 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3259 * configure: Regenerate.
3261 * config/i386/sol2-9.h: Remove.
3263 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3264 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3265 Remove Solaris 9 references.
3267 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3269 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3270 (floatuns<GPI:mode><GPF:mode>2): Remove.
3271 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3272 and floatuns conversions.
3273 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3274 and floatuns conversions.
3275 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3276 (w1,w2): New mode attributes for inequal width conversions.
3278 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3280 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3281 the output asm format.
3283 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3285 * config/aarch64/aarch64-simd.md
3286 (aarch64_cm<optab>di): Always split.
3287 (*aarch64_cm<optab>di): New.
3288 (aarch64_cmtstdi): Always split.
3289 (*aarch64_cmtstdi): New.
3291 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3293 PR tree-optimization/60823
3294 * omp-low.c (ipa_simd_modify_function_body): Go through
3295 all SSA_NAMEs and for those refering to vector arguments
3296 which are going to be replaced adjust SSA_NAME_VAR and,
3297 if it is a default definition, change it into a non-default
3298 definition assigned at the beginning of function from new_decl.
3299 (ipa_simd_modify_stmt_ops): Rewritten.
3300 * tree-dfa.c (set_ssa_default_def): When removing default def,
3301 check for NULL loc instead of NULL *loc.
3303 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3305 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3306 restrictions on core registers for DImode values in Thumb2.
3308 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3310 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3311 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3313 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3315 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3316 (*iordi_notzesidi_di): Likewise.
3317 (*iordi_notsesidi_di): Likewise.
3319 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3321 * config/arm/arm-protos.h (tune_params): New struct members.
3322 * config/arm/arm.c: Initialise tune_params per processor.
3323 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3324 for speed, based on new tune_params.
3326 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3328 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3329 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3330 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3331 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3332 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3333 (vrnda_f64): Likewise.
3334 (vrndi_f64): Likewise.
3335 (vrndm_f64): Likewise.
3336 (vrndn_f64): Likewise.
3337 (vrndp_f64): Likewise.
3338 (vrndx_f64): Likewise.
3340 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3342 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3343 GET_MODE_SIZE argument is enum machine_mode.
3345 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3348 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3349 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3351 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3354 * asan.c (asan_emit_stack_protection): Force the base to align to
3355 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3356 appropriate bits if STRICT_ALIGNMENT.
3357 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3359 (expand_used_vars): Leave a space in the stack frame for alignment
3360 if STRICT_ALIGNMENT.
3362 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3364 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3366 (gimple_store_p): Likewise.
3367 (gimple_assign_load_p): Likewise.
3368 (gimple_assign_cast_p): Likewise.
3369 (gimple_clobber_p): Likewise.
3371 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3372 rather than a gimple.
3373 (gimple_assign_cast_p): Likewise.
3375 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3378 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3379 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3381 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3382 more debug information for E500 if -mdebug=reg.
3384 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3387 * config/i386/i386.c (ix86_expand_builtin)
3388 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3389 register for target RTX.
3390 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3392 2014-04-18 Cong Hou <congh@google.com>
3394 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3395 the widen-mult pattern by handling two operands with different sizes,
3396 and operands whose size is smaller than half of the result type.
3398 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3400 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3401 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3402 (do_estimate_edge_time): Compute it.
3403 * ipa-inline.c (want_inline_small_function_p): Bypass
3404 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3406 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3408 * ipa-inline.c (spec_rem): New static variable.
3409 (dump_overall_stats): New function.
3410 (dump_inline_stats): New function.
3412 2014-04-18 Richard Henderson <rth@redhat.com>
3414 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3415 to GET_MODE_SIZE, not a reg_class_t.
3417 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3419 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3420 (vsx_xxmrglw_<mode>): Likewise.
3422 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3425 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3426 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3427 (rs6000_init_hard_regno_mode_ok): Likewise.
3429 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3431 * ipa-inline.c (inline_small_functions): Account only non-cold
3433 * doc/invoke.texi (inline-unit-growth): Update documentation.
3435 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3437 * config/rs6000/rs6000.md (addti3, subti3): New.
3439 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3442 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3443 comment. Check optimize_insn_for_size_p instead of
3444 optimize_insn_for_speed_p.
3446 2014-04-17 Martin Jambor <mjambor@suse.cz>
3448 * gimple-iterator.c (gsi_start_edge): New function.
3449 * gimple-iterator.h (gsi_start_edge): Declare.
3450 * tree-sra.c (single_non_eh_succ): New function.
3451 (disqualify_ops_if_throwing_stmt): Renamed to
3452 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3453 having one non-EH successor BB.
3454 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3455 generate loads into replacements.
3456 (sra_modify_assign): Likewise and and also use the simple path for
3458 (sra_modify_function_body): Commit statements on edges.
3460 2014-04-17 Richard Biener <rguenther@suse.de>
3463 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3464 comparison results and add clarifying comment.
3466 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3468 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3469 (blank_mode): Initialize it.
3470 (emit_mode_size_inline, emit_mode_nunits_inline,
3471 emit_mode_inner_inline): New functions.
3472 (emit_insn_modes_h): Call them and surround their output with
3473 #if GCC_VERSION >= 4001 ... #endif.
3474 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3475 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3476 mode_* arrays if the argument is __builtin_constant_p.
3477 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3478 is enum machine_mode.
3480 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3482 * passes.c (opt_pass::execute): Adjust.
3483 (pass_manager::execute_pass_mode_switching): Likewise.
3484 (early_local_passes::execute): Likewise.
3485 (execute_one_pass): Pass cfun to the pass's execute method.
3486 * tree-pass.h (opt_pass::execute): Add function * argument.
3487 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3488 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3489 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3490 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3491 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3492 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3493 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3494 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3495 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3496 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3497 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3498 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3499 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3500 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3501 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3502 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3503 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3504 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3505 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3506 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3507 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3508 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3509 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3510 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3511 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3512 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3513 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3514 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3515 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3518 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3520 * passes.c (opt_pass::gate): Take function * argument.
3521 (gate_all_early_local_passes): Merge into
3522 (early_local_passes::gate): this.
3523 (gate_all_early_optimizations): Merge into
3524 (all_early_optimizations::gate): this.
3525 (gate_all_optimizations): Mege into
3526 (all_optimizations::gate): this.
3527 (gate_all_optimizations_g): Merge into
3528 (all_optimizations_g::gate): this.
3529 (gate_rest_of_compilation): Mege into
3530 (rest_of_compilation::gate): this.
3531 (gate_postreload): Merge into
3532 (postreload::gate): this.
3533 (dump_one_pass): Pass cfun to the pass's gate method.
3534 (execute_ipa_summary_passes): Likewise.
3535 (execute_one_pass): Likewise.
3536 (ipa_write_summaries_2): Likewise.
3537 (ipa_write_optimization_summaries_1): Likewise.
3538 (ipa_read_summaries_1): Likewise.
3539 (ipa_read_optimization_summaries_1): Likewise.
3540 (execute_ipa_stmt_fixups): Likewise.
3541 * tree-pass.h (opt_pass::gate): Add function * argument.
3542 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3543 combine-stack-adj.c, combine.c, compare-elim.c,
3544 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3545 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3546 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3547 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3548 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3549 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3550 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3551 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3552 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3553 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3554 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3555 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3556 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3557 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3558 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3559 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3560 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3561 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3562 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3563 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3564 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3565 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3566 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3567 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3568 var-tracking.c, vtable-verify.c, web.c: Adjust.
3570 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3572 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3573 * configure: Regenerate.
3575 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3577 * passes.c (dump_one_pass): don't check pass->has_gate.
3578 (execute_ipa_summary_passes): Likewise.
3579 (execute_one_pass): Likewise.
3580 (ipa_write_summaries_2): Likewise.
3581 (ipa_write_optimization_summaries_1): Likewise.
3582 (ipa_read_optimization_summaries_1): Likewise.
3583 (execute_ipa_stmt_fixups): Likewise.
3584 * tree-pass.h (pass_data::has_gate): Remove.
3585 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3586 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3587 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3588 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3589 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3590 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3591 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3592 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3593 gimple-low.c, gimple-ssa-isolate-paths.c,
3594 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3595 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3596 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3597 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3598 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3599 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3600 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3601 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3602 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3603 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3604 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3605 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3606 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3607 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3608 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3609 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3610 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3611 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3612 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3613 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3614 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3615 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3618 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3620 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3622 * passes.c (pass_manager::register_dump_files_1): Merge into
3623 (pass_manager::register_dump_files): this, and remove its handling of
3624 properties since the pass always has the properties anyway.
3625 (pass_manager::pass_manager): Adjust.
3627 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3629 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3630 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3631 dealing with properties.
3632 (pass_manager::register_dump_files): Adjust.
3634 2014-03-20 Mark Wielaard <mjw@redhat.com>
3636 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3637 then represent the bound as normal constant value.
3639 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3642 Forward port from 4.8 branch
3643 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3645 * config/i386/bmiintrin.h (_blsi_u32): New.
3649 (_blsmsk_u32): Ditto.
3650 (_blsmsk_u64): Ditto.
3651 (_tzcnt_u32): Ditto.
3652 (_tzcnt_u64): Ditto.
3654 2014-04-17 Kito Cheng <kito@0xlab.org>
3656 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3658 2014-04-17 Richard Biener <rguenther@suse.de>
3661 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3662 boolean results for comparisons.
3664 2014-04-17 Richard Biener <rguenther@suse.de>
3666 PR tree-optimization/60836
3667 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3668 initial PHI args to be gimple values.
3670 2014-04-17 Richard Biener <rguenther@suse.de>
3672 PR tree-optimization/60841
3673 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3674 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3675 of stmts to SLP build.
3676 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3677 (vect_analyze_slp): Likewise.
3678 (vect_analyze_slp_instance): Likewise.
3679 (vect_build_slp_tree): Limit overall SLP tree growth.
3680 * tree-vectorizer.h (vect_analyze_data_refs,
3681 vect_analyze_slp): Adjust prototypes.
3683 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3685 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3688 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3690 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3691 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3692 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3693 for TARGET_SLOW_PSHUFB
3695 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3697 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3698 * config/i386/i386.c (intel_cost): Ditto.
3700 2014-04-17 Joey Ye <joey.ye@arm.com>
3702 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3704 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3706 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3709 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3711 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3712 (type_all_derivations_known_p): New predicate.
3713 (type_all_ctors_visible_p): New predicate.
3714 (type_possibly_instantiated_p): New predicate.
3715 (get_odr_type): Compute all_derivations_known.
3716 (dump_odr_type): Dump the flag.
3717 (maybe_record_type): Cleanup.
3718 (record_target_from_binfo): Add bases_to_consider array;
3719 record bases for types w/o instances and skip CXX destructor.
3720 (possible_polymorphic_call_targets_1): Add bases_to_consider
3721 and consider_construction parameters; check if type may have instance.
3722 (get_polymorphic_call_info): Set maybe_in_construction to true
3723 when we know nothing.
3724 (record_targets_from_bases): Skip CXX destructors; they are
3725 never called for types in construction.
3726 (possible_polymorphic_call_targets): Do not record target when
3727 type may not have instance.
3729 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3732 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3733 external aliases alive, too.
3735 2014-04-16 Andrew Pinski <apinski@cavium.com>
3737 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3740 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3742 * final.c (compute_alignments): Do not apply loop alignment to a block
3743 falling through to the exit.
3745 2014-04-16 Catherine Moore <clm@codesourcery.com>
3747 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3748 Adjust constraints for microMIPS store patterns.
3750 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3752 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3754 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3756 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3757 (append_use): Run at -O0.
3758 (append_vdef): Likewise.
3759 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3760 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3762 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3764 PR tree-optimization/60844
3765 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3766 (propagate_op_to_single_use, remove_visited_stmt_chain,
3767 linearize_expr, repropagate_negates, reassociate_bb): Use it
3768 instead of gsi_remove.
3770 2014-04-16 Martin Jambor <mjambor@suse.cz>
3772 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3773 ipa_transforms_to_apply.
3774 (cgraph_function_versioning): Assert that old_node has empty
3775 ipa_transforms_to_apply.
3776 * trans-mem.c (ipa_tm_create_version): Likewise.
3777 * tree-inline.c (tree_function_versioning): Do not duplicate
3778 ipa_transforms_to_apply.
3780 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3783 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3785 Pass necessary as flags on 64-bit Solaris/x86.
3786 Use lowercase relocs for x86_64-*-*.
3787 * configure: Regenerate.
3789 2014-04-15 Jan Hubicka <jh@suse.cz>
3791 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3792 (maybe_record_node, likely_target_p): Use it.
3794 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3797 Revert following patch
3799 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3802 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3803 software floating point or no floating point registers, do not
3804 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3805 in GPRs that occurs after we tested for GPRs that would never be
3808 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3809 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3810 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3811 specifically allow DDmode, since that does not use the SPE SIMD
3814 2014-03-21 Mark Wielaard <mjw@redhat.com>
3816 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3817 as unsigned or int depending on type and value used.
3819 2014-04-15 Richard Biener <rguenther@suse.de>
3821 PR rtl-optimization/56965
3822 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3823 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3825 * alias.c (true_dependence_1): Do not call
3826 nonoverlapping_component_refs_p.
3827 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3828 nonoverlapping_component_refs_p.
3829 (indirect_refs_may_alias_p): Likewise.
3831 2014-04-15 Teresa Johnson <tejohnson@google.com>
3833 * cfg.c (dump_bb_info): Fix flags check.
3834 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3836 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3838 PR rtl-optimization/60663
3839 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3842 2014-04-15 Richard Biener <rguenther@suse.de>
3844 * lto-streamer.h (LTO_major_version): Bump to 4.
3846 2014-04-15 Richard Biener <rguenther@suse.de>
3848 * common.opt (lto_partition_model): New enum.
3849 (flto-partition=): Merge separate options with a single with argument,
3850 add -flto-partition=one support.
3851 * flag-types.h (enum lto_partition_model): Declare.
3852 * opts.c (finish_options): Remove duplicate -flto-partition=
3854 * lto-wrapper.c (run_gcc): Adjust.
3856 2014-04-15 Richard Biener <rguenther@suse.de>
3858 * alias.c (ncr_compar): New function.
3859 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3861 2014-04-15 Richard Biener <rguenther@suse.de>
3863 * alias.c (record_component_aliases): Do not walk BINFOs.
3865 2014-04-15 Richard Biener <rguenther@suse.de>
3867 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3868 Add struct function argument and adjust.
3869 (find_func_aliases_for_call): Likewise.
3870 (find_func_aliases): Likewise.
3871 (find_func_clobbers): Likewise.
3872 (intra_create_variable_infos): Likewise.
3873 (compute_points_to_sets): Likewise.
3874 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3876 2014-04-15 Richard Biener <rguenther@suse.de>
3878 * tree.c (iterative_hash_expr): Use enum tree_code_class
3879 to store TREE_CODE_CLASS.
3880 (tree_block): Likewise.
3881 (tree_set_block): Likewise.
3882 * tree.h (fold_build_pointer_plus_loc): Use
3883 convert_to_ptrofftype_loc.
3885 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3888 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3891 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3893 * cfgloop.h (struct loop): Move force_vectorize down.
3894 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3895 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3896 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3897 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3898 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3899 * tree-core.h (enum annot_expr_kind): Add new kind values.
3900 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3901 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3903 * tree.def (ANNOTATE_EXPR): Tweak comment.
3905 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3907 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3910 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3912 * tree.h (TYPE_IDENTIFIER): Declare.
3913 * tree.c (subrange_type_for_debug_p): Use it.
3914 * godump.c (go_format_type): Likewise.
3915 * dwarf2out.c (is_cxx_auto, modified_type_die,
3916 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3917 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3919 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3922 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3924 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3926 * config/i386/i386.c (examine_argument): Return bool. Return true if
3927 parameter should be passed in memory.
3928 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3929 (construct_container): Update calls to examine_argument.
3930 (function_arg_advance_64): Ditto.
3931 (return_in_memory_32): Merge with ix86_return_in_memory.
3932 (return_in_memory_64): Ditto.
3933 (return_in_memory_ms_64): Ditto.
3935 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3937 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3938 * coverage.c (coverage_compute_profile_id): Handle externally visible
3941 2014-04-14 Martin Jambor <mjambor@suse.cz>
3943 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3944 DECL_DISREGARD_INLINE_LIMITS functions.
3946 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3949 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3951 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3954 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3955 optimize_insn_for_speed_p instead of
3956 optimize_function_for_speed_p.
3958 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3960 * doc/invoke.texi (free): Document AArch64.
3962 2014-04-14 Richard Biener <rguenther@suse.de>
3964 PR tree-optimization/60042
3965 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3966 (insert_into_preds_of_block): Do not prevent PHI insertion
3967 for REFERENCE exprs here ...
3968 (eliminate_dom_walker::before_dom_children): ... but prevent
3969 their use here under similar conditions when applied to the
3970 IL after PRE optimizations.
3972 2014-04-14 Richard Biener <rguenther@suse.de>
3974 * passes.def: Move early points-to after early SRA.
3976 2014-04-14 Richard Biener <rguenther@suse.de>
3978 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3979 check for which sign-changes we allow when forwarding
3980 a converted value into a switch.
3982 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3984 * stor-layout.c (place_field): Finalize non-constant offset for the
3987 2014-04-14 Richard Biener <rguenther@suse.de>
3989 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3991 (expand_switch_using_bit_tests_p): Likewise.
3992 (process_switch): Compute and pass on speed_p based on the
3994 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3995 optimize_bb_for_speed_p.
3997 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3999 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4000 * function.h (struct function): Rename has_force_vect_loops into
4001 has_force_vectorize_loops.
4002 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4003 (input_struct_function_base): Likewise.
4004 * lto-streamer-out.c (output_cfg): Likewise.
4005 (output_struct_function_base): Likewise.
4006 * omp-low.c (expand_omp_simd): Likewise.
4007 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4008 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4009 (version_loop_for_if_conversion): Likewise.
4010 (tree_if_conversion): Likewise.
4011 (main_tree_if_conversion): Likewise.
4012 (gate_tree_if_conversion): Likewise.
4013 * tree-inline.c (copy_loops): Likewise.
4014 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4015 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4016 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4017 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4018 * tree-vectorizer.c (vectorize_loops): Likewise.
4019 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4021 2014-04-14 Richard Biener <rguenther@suse.de>
4024 * lto-streamer-out.c (wrap_refs): New function.
4025 (lto_output): Wrap symbol references in global initializes in
4026 type-preserving MEM_REFs.
4028 2014-04-14 Christian Bruel <christian.bruel@st.com>
4030 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4032 2014-04-14 Christian Bruel <christian.bruel@st.com>
4034 * config/sh/sh.md (setmemqi): New expand pattern.
4035 * config/sh/sh.h (CLEAR_RATIO): Define.
4036 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4037 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4039 2014-04-14 Richard Biener <rguenther@suse.de>
4042 * fold-const.c (negate_expr_p): Don't negate directional rounding
4044 (fold_negate_expr): Likewise.
4046 2014-04-14 Richard Biener <rguenther@suse.de>
4048 PR tree-optimization/59817
4049 PR tree-optimization/60453
4050 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4051 recursion to catch all CHRECs in the scalar evolution and restrict
4052 the predicate for the remains appropriately.
4054 2014-04-12 Catherine Moore <clm@codesourcery.com>
4056 * config/mips/constraints.md: Add new register constraint "kb".
4057 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4058 (*movhi_internal): Likewise.
4059 (*movqi_internal): Likewise.
4060 * config/mips/mips.h (M16_STORE_REGS): New register class.
4061 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4062 (REG_CLASS_CONTENTS): Likewise.
4063 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4065 2014-04-11 Tobias Burnus <burnus@net-b.de>
4068 * doc/invoke.texi (-Wformat-signedness): Document it.
4069 (Wformat=2): Mention that this enables -Wformat-signedness.
4071 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4073 * common/config/epiphany/epiphany-common.c
4074 (epiphany_option_optimization_table): Enable section anchors by
4075 default at -O1 or higher.
4076 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4077 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4078 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4079 carries no extra cost.
4080 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4081 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4082 * config/epiphany/predicates.md (memclob_operand): New predicate.
4083 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4084 Use memclob_operand predicate and X constraint for operand 3.
4086 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4088 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4089 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4092 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4094 PR rtl-optimization/60651
4095 * mode-switching.c (optimize_mode_switching): Make sure to emit
4096 sets of a lower numbered entity before sets of a higher numbered
4097 entity to a mode of the same or lower priority.
4098 When creating a seginfo for a basic block that starts with a code
4099 label, move the insertion point past the code label.
4100 (new_seginfo): Document and enforce requirement that
4101 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4102 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4103 * doc/tm.texi: Regenerate.
4105 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4108 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4110 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4112 * BASE-VER: Set to 4.10.0.
4114 2014-04-11 Tobias Burnus <burnus@net-b.de>
4117 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4118 * doc/gcc.texi (Service): Update description in the @menu
4119 * doc/invoke.texi (Option Summary): Remove misplaced and
4122 2014-04-11 Steve Ellcey <sellcey@mips.com>
4123 Jakub Jelinek <jakub@redhat.com>
4126 * expr.c (convert_move): Use emit_store_flag_force instead of
4127 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4130 2014-04-11 Richard Biener <rguenther@suse.de>
4133 * varasm.c (assemble_alias): Avoid endless error reporting
4134 recursion by setting TREE_ASM_WRITTEN.
4136 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4138 * config/s390/s390.md: Add a splitter for NOT rtx.
4140 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4142 PR rtl-optimization/60663
4143 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4145 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4146 Jakub Jelinek <jakub@redhat.com>
4149 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4150 flag from decl_node to node.
4152 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4155 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4156 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4157 ameliorating the cases where it can be.
4159 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4162 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4164 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4165 (loadsync_<mode>): Change mode.
4166 (load_quadpti, store_quadpti): New.
4167 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4168 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4169 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4171 2014-04-09 Cong Hou <congh@google.com>
4174 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4177 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4179 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4180 instead of vnor to exploit possible fusion opportunity in the
4182 (altivec_expand_vec_perm_const_le): Likewise.
4184 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4186 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4187 (loadsync_<mode>): Change mode.
4188 (load_quadpti, store_quadpti): New.
4189 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4190 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4192 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4195 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4196 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4197 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4199 2014-04-08 Richard Biener <rguenther@suse.de>
4202 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4203 a 64bit widest int print double-int similar to on HWI64 hosts.
4205 2014-04-08 Richard Biener <rguenther@suse.de>
4207 PR tree-optimization/60785
4208 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4209 default defs properly.
4211 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4213 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4214 (Weffc++): Likewise.
4216 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4218 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4219 set completep to false rather than true.
4221 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4224 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4225 ARM_TARGET2_DWARF_FORMAT.
4227 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4230 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4231 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4234 2014-04-07 Richard Biener <rguenther@suse.de>
4236 PR tree-optimization/60766
4237 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4238 (may_eliminate_iv): Convert cand_value_at result to desired type.
4240 2014-04-07 Jason Merrill <jason@redhat.com>
4243 * common.opt (-fno-gnu-unique): Add.
4244 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4246 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4248 * haifa-sched.c: Fix outdated function reference and minor
4249 grammar errors in introductory comment.
4251 2014-04-07 Richard Biener <rguenther@suse.de>
4254 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4256 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4258 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4261 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4262 size accounting for thunks.
4263 (pa_asm_output_mi_thunk): Use final_start_function() and
4264 final_end_function() to output function start and end directives.
4266 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4268 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4269 device specific ISA/ feature information. Remove short_sp and
4270 errata_skip ds. Add avr_device_specific_features enum to have device
4272 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4273 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4274 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4275 updated device specific info.
4276 * config/avr/avr-mcus.def: Merge device specific details to
4277 dev_attribute field.
4278 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4280 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4281 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4282 assembler if RMW isa supported by current device.
4283 * config/avr/genmultilib.awk: Update as device info structure changed.
4284 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4286 2014-04-04 Cong Hou <congh@google.com>
4288 PR tree-optimization/60656
4289 * tree-vect-stmts.c (supportable_widening_operation):
4290 Fix a bug that elements in a vector with vect_used_by_reduction
4291 property are incorrectly reordered when the operation on it is not
4292 consistant with the one in reduction operation.
4294 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4296 PR rtl-optimization/60155
4297 * gcse.c (record_set_data): New function.
4298 (single_set_gcse): New function.
4299 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4300 (hoist_code): Likewise.
4301 (get_pressure_class_and_nregs): Likewise.
4303 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4305 * explow.c (probe_stack_range): Emit a final optimization blockage.
4307 2014-04-04 Anthony Green <green@moxielogic.com>
4309 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4312 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4315 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4316 flags are set only during streaming.
4317 * ipa.c (process_references, walk_polymorphic_call_targets,
4318 symtab_remove_unreachable_nodes): Drop bodies of always inline
4319 after early inlining.
4320 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4322 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4323 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4326 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4329 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4332 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4334 (cortex_a53_fdivd): Likewise.
4336 2014-04-04 Martin Jambor <mjambor@suse.cz>
4339 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4341 * cgraph.c (clone_of_p): Also return true if thunks match.
4342 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4343 cgraph_function_or_thunk_node and an obsolete comment.
4344 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4346 (build_function_decl_skip_args): Likewise.
4347 (set_new_clone_decl_and_node_flags): New function.
4348 (duplicate_thunk_for_node): Likewise.
4349 (redirect_edge_duplicating_thunks): Likewise.
4350 (cgraph_clone_node): New parameter args_to_skip, pass it to
4351 redirect_edge_duplicating_thunks which is called instead of
4352 cgraph_redirect_edge_callee.
4353 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4354 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4356 2014-04-04 Jeff Law <law@redhat.com>
4359 * config/arm/predicates.md (const_int_I_operand): New predicate.
4360 (const_int_M_operand): Similarly.
4361 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4363 (insv_t2, extv_reg, extzv_t2): Likewise.
4364 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4365 (pop_multiple_with_writeback_and_return): Likewise.
4366 (vfp_pop_multiple_with_writeback): Likewise
4368 2014-04-04 Richard Biener <rguenther@suse.de>
4371 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4372 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4374 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4375 * gimplify.c (gimple_add_tmp_var_fn): New function.
4376 * gimple-expr.h (create_tmp_reg_fn): Declare.
4377 * gimple-expr.c (create_tmp_reg_fn): New function.
4378 * gimple-low.c (record_vars_into): Don't change cfun.
4379 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4380 code generation without cfun.
4382 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4385 * Makefile.in (install-driver): Fix shell scripting.
4387 2014-04-03 Cong Hou <congh@google.com>
4389 PR tree-optimization/60505
4390 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4391 threshold of number of iterations below which no vectorization
4393 * tree-vect-loop.c (new_loop_vec_info):
4394 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4395 * tree-vect-loop.c (vect_analyze_loop_operations):
4396 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4397 * tree-vect-loop.c (vect_transform_loop):
4398 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4399 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4400 of iterations of the loop and see if we should build the epilogue.
4402 2014-04-03 Richard Biener <rguenther@suse.de>
4404 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4405 (streamer_tree_cache_create): Adjust.
4406 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4407 to allow optional nodes array.
4408 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4409 (streamer_tree_cache_append): Likewise.
4410 (streamer_tree_cache_create): Create nodes array optionally
4411 as specified by parameter.
4412 * lto-streamer-out.c (create_output_block): Avoid maintaining
4413 the node array in the writer cache.
4414 (DFS_write_tree): Remove assertion.
4415 (produce_asm_for_decls): Free the out decl state hash table early.
4416 * lto-streamer-in.c (lto_data_in_create): Adjust for
4417 streamer_tree_cache_create prototype change.
4419 2014-04-03 Richard Biener <rguenther@suse.de>
4421 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4422 set TREE_CHAIN to NULL_TREE.
4424 2014-04-03 Richard Biener <rguenther@suse.de>
4426 PR tree-optimization/60740
4427 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4428 over all GIMPLE_COND operands.
4430 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4432 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4433 (Weffc++): Remove Scott's numbering, merge lists and reference
4436 2014-04-03 Nick Clifton <nickc@redhat.com>
4438 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4441 2014-04-03 Martin Jambor <mjambor@suse.cz>
4443 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4444 mention gcc_unreachable before failing.
4445 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4448 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4451 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4452 inconsistent code and instead mark the context inconsistent.
4453 (possible_polymorphic_call_targets): For inconsistent contexts
4454 return empty complete list.
4456 2014-04-02 Anthony Green <green@moxielogic.com>
4458 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4459 (extendqisi2, extendhisi2): Define.
4460 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4461 (WCHAR_TYPE): Change to unsigned int.
4463 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4465 PR tree-optimization/60733
4466 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4467 insertion point for PHI candidates to be the end of the feeding
4468 block for the PHI argument.
4470 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4472 PR rtl-optimization/60650
4473 * lra-constraints.c (process_alt_operands): Decrease reject for
4474 earlyclobber matching.
4476 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4478 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4480 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4482 * config/spu/spu.c (pad_bb): Do not crash when the last
4483 insn is CODE_FOR_blockage.
4485 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4487 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4488 lies outside the target mode.
4490 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4493 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4494 software floating point or no floating point registers, do not
4495 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4496 in GPRs that occurs after we tested for GPRs that would never be
4499 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4500 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4501 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4502 specifically allow DDmode, since that does not use the SPE SIMD
4505 2014-04-02 Richard Biener <rguenther@suse.de>
4508 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4509 MODE_INTs. Properly use negv_optab.
4510 (expand_abs): Likewise.
4512 2014-04-02 Richard Biener <rguenther@suse.de>
4515 * Makefile.in (install-driver): Guard extra installs with special
4518 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4520 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4523 2014-04-01 Richard Henderson <rth@redhat.com>
4526 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4527 alternative enabled before register allocation.
4529 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4531 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4532 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4534 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4535 (nios2_got_address): Update nios2_large_got_address call site.
4536 (nios2_delegitimize_address): New function.
4537 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4538 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4539 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4541 2014-04-01 Martin Husemann <martin@duskware.de>
4543 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4546 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4548 PR rtl-optimization/60604
4549 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4550 check from register_operand.
4551 (register_operand): Redefine in terms of general_operand.
4552 (nonmemory_operand): Use register_operand for the non-constant cases.
4554 2014-04-01 Richard Biener <rguenther@suse.de>
4556 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4558 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4560 * doc/invoke.texi (mapp-regs): Clarify.
4562 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4564 * config/i386/avx512fintrin.h (__v32hi): Define type.
4565 (__v64qi): Likewise.
4566 (_mm512_set1_epi8): Define.
4567 (_mm512_set1_epi16): Define.
4568 (_mm512_set4_epi32): Define.
4569 (_mm512_set4_epi64): Define.
4570 (_mm512_set4_pd): Define.
4571 (_mm512_set4_ps): Define.
4572 (_mm512_setr4_epi64): Define.
4573 (_mm512_setr4_epi32): Define.
4574 (_mm512_setr4_pd): Define.
4575 (_mm512_setr4_ps): Define.
4576 (_mm512_setzero_epi32): Define.
4578 2014-03-31 Martin Jambor <mjambor@suse.cz>
4581 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4582 callsite_arguments_match_p. Updated all callers. Also check types of
4583 corresponding formal parameters and actual arguments.
4584 (not_all_callers_have_enough_arguments_p) Renamed to
4585 some_callers_have_mismatched_arguments_p.
4587 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4589 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4591 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4594 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4597 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4599 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4600 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4602 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4603 Use FMAMODE_NOVF512 mode iterator.
4604 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4605 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4606 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4607 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4609 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4610 Use VF_128_256 mode iterator.
4611 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4614 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4616 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4617 static chain if needed.
4619 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4622 * lra-constraints.c (index_part_to_reg): New.
4623 (process_address): Use it.
4625 2014-03-27 Jeff Law <law@redhat.com>
4626 Jakub Jelinek <jakub@redhat.com>
4629 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4630 gen_rtx_{PLUS,MULT} to build up the address expression.
4632 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4633 creating non-canonical RTL.
4635 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4638 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4639 functions; reorganize to make cheap checks first.
4640 (inline_small_functions): Do not estimate growth when dumping;
4642 * ipa-inline.h (inline_summary): Add min_size.
4643 (growth_likely_positive): New function.
4644 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4645 (set_cond_stmt_execution_predicate): Cleanup.
4646 (estimate_edge_size_and_time): Compute min_size.
4647 (estimate_calls_size_and_time): Likewise.
4648 (estimate_node_size_and_time): Likewise.
4649 (inline_update_overall_summary): Update min_size.
4650 (do_estimate_edge_time): Likewise.
4651 (do_estimate_edge_size): Update.
4652 (do_estimate_edge_hints): Update.
4653 (growth_likely_positive): New function.
4655 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4658 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4659 also if addr has VOIDmode.
4661 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4663 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4664 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4666 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4667 instructions as well as AdvancedSIMD loads.
4669 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4671 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4672 Use crypto_aese type.
4673 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4674 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4675 crypto_aese, crypto_aesmc. Move to types.md.
4676 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4678 * config/arm/iterators.md (crypto_type): Likewise.
4680 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4682 * cgraph.c: Include expr.h and tree-dfa.h.
4683 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4686 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4689 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4690 regs from checking multi-reg pseudos.
4692 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4694 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4696 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4698 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4699 if it would clobber the stack pointer, even temporarily.
4701 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4703 * mode-switching.c: Make small adjustments to the top comment.
4705 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4707 * config/rs6000/constraints.md (wD constraint): New constraint to
4708 match the constant integer to get the top DImode/DFmode out of a
4709 vector in a VSX register.
4711 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4712 match the constant integer to get the top DImode/DFmode out of a
4713 vector in a VSX register.
4715 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4718 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4721 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4722 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4724 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4725 Optimize vec_extract of 64-bit values, where the value being
4726 extracted is in the top word, where we can use scalar
4727 instructions. Add direct move and store support. Combine the big
4728 endian/little endian vector select load support into a single insn.
4729 (vsx_extract_<mode>_internal1): Likewise.
4730 (vsx_extract_<mode>_internal2): Likewise.
4731 (vsx_extract_<mode>_load): Likewise.
4732 (vsx_extract_<mode>_store): Likewise.
4733 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4734 combined into vsx_extract_<mode>_load.
4735 (vsx_extract_<mode>_one_le): Likewise.
4737 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4738 define the top 64-bit vector element.
4740 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4743 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4744 Document vec_vbpermq builtin.
4747 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4748 enable use of xxsldwi and xxpermdi builtin functions.
4749 (vec_xxpermdi): Likewise.
4751 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4752 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4754 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4756 PR rtl-optimization/60650
4757 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4759 (find_spills_for): New.
4760 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4761 Spill all pseudos on the second iteration.
4763 2014-03-27 Marek Polacek <polacek@redhat.com>
4766 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4769 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4771 * config/s390/s390.c (s390_can_use_return_insn): Check for
4772 call-saved FPRs on 31 bit.
4774 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4777 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4778 if they need regimplification, just drop them instead of
4779 calling gimple_regimplify_operands on them.
4781 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4784 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4785 (aarch64_frame_pointer_required): Adjust logic.
4786 (aarch64_can_eliminate): Adjust logic.
4787 (aarch64_override_options_after_change): Adjust logic.
4789 2014-03-27 Dehao Chen <dehao@google.com>
4791 * ipa-inline.c (early_inliner): Update node's inline info.
4793 2014-03-26 Dehao Chen <dehao@google.com>
4795 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4796 compiler inserted conditional jumps for NAN float check.
4798 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4800 * ubsan.h (ubsan_create_data): Change second argument's type
4801 to const location_t *.
4802 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4804 (ubsan_create_data): Change second argument to const location_t *PLOC.
4805 Create Loc field whenever PLOC is non-NULL.
4806 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4807 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4811 * real.c (real_to_integer2): Change type of low to UHWI.
4813 2014-03-26 Tobias Burnus <burnus@net-b.de>
4815 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4816 (CILK_SELF_SPECS): New define.
4817 (driver_self_specs): Use it.
4819 2014-03-26 Richard Biener <rguenther@suse.de>
4821 * tree-pretty-print.c (percent_K_format): Implement special
4822 case for LTO and its stripped down BLOCK tree.
4824 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4827 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4829 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4830 one range is range_int_cst_p, but not both, at least optimize
4831 addition/subtraction of 0 and multiplication by 0 or 1.
4832 * gimple-fold.c (gimple_fold_call): Fold
4833 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4834 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4835 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4837 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4839 PR rtl-optimization/60452
4840 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4841 <case REG>: Return 1 for invalid offsets from the frame pointer.
4843 2014-03-26 Marek Polacek <polacek@redhat.com>
4846 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4849 2014-03-26 Marek Polacek <polacek@redhat.com>
4852 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4855 2014-03-26 Marek Polacek <polacek@redhat.com>
4858 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4859 multiplication in unsigned type.
4861 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4863 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4865 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4867 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4869 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4872 * cif-code.def (UNREACHABLE) New code.
4873 * ipa-inline.c (inline_small_functions): Skip edges to
4874 __builtlin_unreachable.
4875 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4876 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4877 predicate to __bulitin_unreachable.
4878 (set_cond_stmt_execution_predicate): Fix issue when
4879 invert_tree_comparison returns ERROR_MARK.
4880 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4881 propagate to inline clones.
4882 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4884 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4885 * cgraphclones.c (cgraph_clone_node): If call destination is already
4886 ureachable, do not redirect it back.
4887 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4890 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4892 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4893 Do not modify inline clones.
4895 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4897 * config/i386/i386.md (general_sext_operand): New mode attr.
4898 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4899 don't generate (sign_extend (const_int)).
4900 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4901 operands[2]. Use We constraint instead of <i> and
4902 <general_sext_operand> predicate instead of <general_operand>.
4903 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4904 * config/i386/constraints.md (We): New constraint.
4905 * config/i386/predicates.md (x86_64_sext_operand,
4906 sext_operand): New predicates.
4908 2014-03-25 Martin Jambor <mjambor@suse.cz>
4911 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4912 inconsistent devirtualizations to __builtin_unreachable.
4914 2014-03-25 Marek Polacek <polacek@redhat.com>
4917 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4919 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4921 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4922 order of elements for big-endian.
4924 2014-03-25 Richard Biener <rguenther@suse.de>
4927 * gimplify-me.c (gimple_regimplify_operands): Update the
4930 2014-03-25 Martin Jambor <mjambor@suse.cz>
4933 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4934 (lto_output_varpool_node): Likewise.
4935 (input_overwrite_node): Likewise.
4936 (input_varpool_node): Likewise.
4938 2014-03-25 Richard Biener <rguenther@suse.de>
4940 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4941 (run_gcc): Likewise.
4943 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4945 * combine.c (simplify_compare_const): Add MODE argument.
4946 Handle mode_width 0 as very large mode_width.
4947 (try_combine, simplify_comparison): Adjust callers.
4949 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4950 type to avoid signed integer overflow.
4951 * explow.c (plus_constant): Likewise.
4953 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4955 * doc/generic.texi: Correct typos.
4957 2014-03-24 Tobias Burnus <burnus@net-b.de>
4959 * doc/invoke.texi (-flto): Expand section about
4960 using static libraries with LTO.
4962 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4964 PR rtl-optimization/60501
4965 * optabs.def (addptr3_optab): New optab.
4966 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4967 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4968 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4970 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4972 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4974 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4976 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4979 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4980 (_mm256_undefined_ps): Define.
4981 (_mm256_undefined_pd): Define.
4982 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4983 (_mm_undefined_pd): Define.
4984 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4985 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4986 (_mm512_undefined_ps): Define.
4987 (_mm512_undefined_pd): Define.
4988 Use _mm*_undefined_*.
4989 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4991 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4993 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4994 (lshr_simd): DI mode added.
4995 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4996 (aarch64_ushr_simddi): Likewise.
4997 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4998 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4999 (vshrd_n_u64): Likewise.
5001 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5003 * Makefile.in (s-macro_list): Depend on cc1.
5005 2014-03-23 Teresa Johnson <tejohnson@google.com>
5007 * ipa-utils.c (ipa_print_order): Use specified dump file.
5009 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5011 PR rtl-optimization/60601
5012 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5014 * gcc.c (eval_spec_function): Initialize save_growing_value.
5016 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5019 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5020 code == MINUS_EXPR, never swap op0 with op1.
5022 * toplev.c (init_local_tick): Avoid signed integer multiplication
5024 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5025 shift by first operand's bitsize.
5027 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5030 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5032 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5033 TARGET_ISA_64BIT_P(x).
5035 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5037 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5038 pattern for vector nor instead of subtract from splat(-1).
5039 (altivec_expand_vec_perm_const_le): Likewise.
5041 2014-03-21 Richard Henderson <rth@twiddle.net>
5044 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5045 related insns after epilogue_completed.
5047 2014-03-21 Martin Jambor <mjambor@suse.cz>
5050 * cgraph.h (symtab_node): New flag body_removed.
5051 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5052 when removing bodies.
5053 * symtab.c (dump_symtab_base): Dump body_removed flag.
5054 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5055 had their bodies removed.
5057 2014-03-21 Martin Jambor <mjambor@suse.cz>
5060 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5063 2014-03-21 Richard Biener <rguenther@suse.de>
5065 PR tree-optimization/60577
5066 * tree-core.h (struct tree_base): Document nothrow_flag use
5067 in VAR_DECL_NONALIASED.
5068 * tree.h (VAR_DECL_NONALIASED): New.
5069 (may_be_aliased): Adjust.
5070 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5072 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5074 * expr.c (expand_expr_real_1): Remove outdated comment.
5076 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5079 * ira.c (adjust_cleared_regs): Call copy_rtx on
5080 *reg_equiv[REGNO (loc)].src_p before passing it to
5081 simplify_replace_fn_rtx.
5084 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5085 into CONST, put pic register as first operand of PLUS. Use
5086 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5088 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5090 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5092 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5094 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5095 around for store forwarding issue in the FPU on the UT699.
5096 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5097 loads and operations if -mfix-ut699 is specified.
5098 (divtf3_hq): Tweak attribute.
5099 (sqrttf2_hq): Likewise.
5101 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5103 * calls.c (store_one_arg): Remove incorrect const qualification on the
5104 type of the temporary.
5105 * cfgexpand.c (expand_return): Likewise.
5106 * expr.c (expand_constructor): Likewise.
5107 (expand_expr_real_1): Likewise.
5109 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5111 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5114 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5117 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5119 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5121 * config/arm/aarch-common-protos.h
5122 (alu_cost_table): Fix spelling of "extend".
5123 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5125 2014-03-19 Richard Biener <rguenther@suse.de>
5128 * tree-core.h (tree_type_common): Re-order pointer members
5129 to reduce recursion depth during GC walks.
5131 2014-03-19 Marek Polacek <polacek@redhat.com>
5134 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5135 before accessing it.
5137 2014-03-19 Richard Biener <rguenther@suse.de>
5140 * lto-streamer-in.c (input_function): In WPA stage do not drop
5143 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5145 PR tree-optimization/60559
5146 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5147 with build_zero_cst assignment.
5149 2014-03-18 Kai Tietz <ktietz@redhat.com>
5151 PR rtl-optimization/56356
5152 * sdbout.c (sdbout_parms): Verify that parms'
5153 incoming argument is valid.
5154 (sdbout_reg_parms): Likewise.
5156 2014-03-18 Richard Henderson <rth@redhat.com>
5159 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5160 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5161 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5163 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5165 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5166 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5167 Italicize plugin event names in description. Explain that
5168 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5169 Remind that no GCC functions should be called after PLUGIN_FINISH.
5170 Explain what pragmas with expansion are.
5172 2014-03-18 Martin Liska <mliska@suse.cz>
5174 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5175 gimple call statement is update.
5176 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5177 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5179 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5182 * ubsan.c (ubsan_instrument_unreachable): Call
5183 initialize_sanitizer_builtins.
5184 (ubsan_pass): Likewise.
5187 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5188 varpool_finalize_decl instead of rest_of_decl_compilation.
5190 2014-03-18 Richard Biener <rguenther@suse.de>
5192 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5193 by using bitmap_and_compl instead of bitmap_and_compl_into.
5194 (df_rd_transfer_function): Likewise.
5196 2014-03-18 Richard Biener <rguenther@suse.de>
5198 * doc/lto.texi (fresolution): Fix typo.
5200 2014-03-18 Richard Biener <rguenther@suse.de>
5202 * doc/invoke.texi (flto): Update for changes in 4.9.
5204 2014-03-18 Richard Biener <rguenther@suse.de>
5206 * doc/loop.texi: Remove section on the removed lambda framework.
5207 Update loop docs with recent changes in preserving loop structure.
5209 2014-03-18 Richard Biener <rguenther@suse.de>
5211 * doc/lto.texi (-fresolution): Document.
5213 2014-03-18 Richard Biener <rguenther@suse.de>
5215 * doc/contrib.texi: Adjust my name.
5217 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5220 * internal-fn.c: Include diagnostic-core.h.
5221 (expand_BUILTIN_EXPECT): New function.
5222 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5223 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5224 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5225 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5227 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5228 Revert 3 argument __builtin_expect code.
5229 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5230 * gimple-fold.c (gimple_fold_call): Likewise.
5231 * tree.h (fold_builtin_expect): New prototype.
5232 * builtins.c (build_builtin_expect_predicate): Add predictor
5233 argument, if non-NULL, create 3 argument __builtin_expect.
5234 (fold_builtin_expect): No longer static. Add ARG2 argument,
5235 pass it through to build_builtin_expect_predicate.
5236 (fold_builtin_2): Adjust caller.
5237 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5238 * internal-fn.def (BUILTIN_EXPECT): New.
5240 2014-03-18 Tobias Burnus <burnus@net-b.de>
5243 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5244 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5245 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5247 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5250 * predict.c (combine_predictions_for_bb): Fix up formatting.
5251 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5252 fill what it points to if non-NULL.
5253 (tree_predict_by_opcode): Adjust caller, use the predictor.
5254 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5256 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5258 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5259 proper constant for the store mode.
5261 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5263 * symtab.c (change_decl_assembler_name): Fix transparent alias
5266 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5268 * config/aarch64/aarch64.c: Correct the comments about the
5269 aarch64 stack layout.
5271 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5273 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5274 check for GF_OMP_FOR_KIND_FOR.
5276 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5278 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5279 ymm and zmm register names.
5281 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5284 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5285 note creation for the 2010-08-31 changes.
5287 2014-03-17 Marek Polacek <polacek@redhat.com>
5290 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5291 as -fno-tree-loop-vectorize.
5292 (expand_omp_simd): Likewise.
5294 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5296 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5297 (eligible_for_call_delay): New prototype.
5298 * config/sparc/sparc.c (tls_call_delay): Rename into...
5299 (eligible_for_call_delay): ...this. Return false if the instruction
5300 cannot be put in the delay slot of a branch.
5301 (eligible_for_restore_insn): Simplify.
5302 (eligible_for_return_delay): Return false if the instruction cannot be
5303 put in the delay slot of a branch and simplify.
5304 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5305 put in the delay slot of a branch.
5306 * config/sparc/sparc.md (fix_ut699): New attribute.
5307 (tls_call_delay): Delete.
5308 (in_call_delay): Reimplement.
5309 (eligible_for_sibcall_delay): Rename into...
5310 (in_sibcall_delay): ...this.
5311 (eligible_for_return_delay): Rename into...
5312 (in_return_delay): ...this.
5313 (in_branch_delay): Reimplement.
5314 (in_uncond_branch_delay): Delete.
5315 (in_annul_branch_delay): Delete.
5317 2014-03-14 Richard Henderson <rth@redhat.com>
5320 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5321 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5322 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5323 (floathi splitters): Remove.
5324 (float<SWI48x>xf2): New pattern.
5325 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5326 code that tried to handle DImode for 32-bit, but which was excluded
5327 by the pattern's condition. Drop allocation of stack temporary.
5328 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5329 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5330 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5331 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5332 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5333 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5334 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5335 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5336 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5337 (*float<SWI48x><X87MODEF>2_i387): Remove.
5338 (all float _with_temp splitters): Remove.
5339 (*float<SWI48x><MODEF>2_i387): New pattern.
5340 (*float<SWI48><MODEF>2_sse): New pattern.
5341 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5342 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5344 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5345 Marek Polacek <polacek@redhat.com>
5348 * common.opt (dump_base_name_prefixed): New Variable.
5349 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5350 if x_dump_base_name_prefixed is already set, set it at the end.
5352 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5354 PR rtl-optimization/60508
5355 * lra-constraints.c (get_reload_reg): Add new parameter
5357 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5358 Pass the new parameter values.
5360 2014-03-14 Richard Biener <rguenther@suse.de>
5362 * common.opt: Revert unintented changes from r205065.
5365 2014-03-14 Richard Biener <rguenther@suse.de>
5368 * cfghooks.c (split_block): Properly adjust all loops the
5369 block was a latch of.
5371 2014-03-14 Martin Jambor <mjambor@suse.cz>
5374 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5377 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5380 * config/avr/avr.c (avr_set_current_function): Pass function name
5381 through default_strip_name_encoding before sanity checking instead
5382 of skipping the first char of the assembler name.
5384 2014-03-13 Richard Henderson <rth@redhat.com>
5387 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5388 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5389 * config/i386/i386-protos.h: Likewise.
5390 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5391 in the expander instead of a splitter.
5392 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5393 any possibility of requiring a memory.
5394 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5395 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5396 (fp branch splitters): Update for ix86_split_fp_branch.
5397 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5398 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5399 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5400 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5401 (*fop_<MODEF>_3_i387): Likewise.
5402 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5403 (splitters for the fop_* register patterns): Remove.
5404 (fscalexf4_i387): Rename from *fscalexf4_i387.
5405 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5407 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5409 PR tree-optimization/59779
5410 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5411 type for bitsize and maxsize instead of HOST_WIDE_INT.
5413 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5415 PR rtl-optimization/57320
5416 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5417 the CFG after thread_prologue_and_epilogue_insns.
5419 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5421 PR rtl-optimization/57189
5422 * lra-constraints.c (process_alt_operands): Disfavor spilling
5425 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5427 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5429 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5431 PR tree-optimization/59025
5433 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5434 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5436 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5439 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5440 calls of avr_out_plus_1.
5442 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5444 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5445 BB's single pred and update the father loop's latch info later.
5447 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5449 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5453 (VEC_base): Likewise.
5454 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5455 registers, we need to swap double words in little endian mode.
5457 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5458 to be a container mode for 128-bit integer operations added in ISA
5459 2.07. Unlike TImode and PTImode, the preferred register set is
5460 the Altivec/VMX registers for the 128-bit operations.
5462 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5464 (rs6000_split_128bit_ok_p): Likewise.
5466 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5467 macros for creating ISA 2.07 normal and overloaded builtin
5468 functions with 3 arguments.
5469 (BU_P8V_OVERLOAD_3): Likewise.
5470 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5471 for use as overloaded functions.
5472 (VPERM_1TI_UNS): Likewise.
5473 (VSEL_1TI): Likewise.
5474 (VSEL_1TI_UNS): Likewise.
5475 (ST_INTERNAL_1ti): Likewise.
5476 (LD_INTERNAL_1ti): Likewise.
5477 (XXSEL_1TI): Likewise.
5478 (XXSEL_1TI_UNS): Likewise.
5479 (VPERM_1TI): Likewise.
5480 (VPERM_1TI_UNS): Likewise.
5481 (XXPERMDI_1TI): Likewise.
5482 (SET_1TI): Likewise.
5483 (LXVD2X_V1TI): Likewise.
5484 (STXVD2X_V1TI): Likewise.
5485 (VEC_INIT_V1TI): Likewise.
5486 (VEC_SET_V1TI): Likewise.
5487 (VEC_EXT_V1TI): Likewise.
5488 (EQV_V1TI): Likewise.
5489 (NAND_V1TI): Likewise.
5490 (ORC_V1TI): Likewise.
5491 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5492 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5494 (VADDUQM): Likewise.
5495 (VSUBCUQ): Likewise.
5496 (VADDEUQM): Likewise.
5497 (VADDECUQ): Likewise.
5498 (VSUBEUQM): Likewise.
5499 (VSUBECUQ): Likewise.
5501 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5502 __int128_t and __uint128_t types.
5503 (__uint128_type): Likewise.
5504 (altivec_categorize_keyword): Add support for vector __int128_t,
5505 vector __uint128_t, vector __int128, and vector unsigned __int128
5506 as a container type for TImode operations that need to be done in
5507 VSX/Altivec registers.
5508 (rs6000_macro_to_expand): Likewise.
5509 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5510 to support 128-bit integer instructions vaddcuq, vadduqm,
5511 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5512 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5514 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5515 for V1TImode, and set up preferences to use VSX/Altivec registers.
5516 Setup VSX reload handlers.
5517 (rs6000_debug_reg_global): Likewise.
5518 (rs6000_init_hard_regno_mode_ok): Likewise.
5519 (rs6000_preferred_simd_mode): Likewise.
5520 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5521 (easy_altivec_constant): Likewise.
5522 (output_vec_const_move): Likewise.
5523 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5525 (rs6000_expand_vector_extract): Likewise.
5526 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5528 (rs6000_const_vec): Add support for V1TImode.
5529 (rs6000_emit_le_vsx_load): Swap double words when loading or
5530 storing TImode/V1TImode.
5531 (rs6000_emit_le_vsx_store): Likewise.
5532 (rs6000_emit_le_vsx_move): Likewise.
5533 (rs6000_emit_move): Add support for V1TImode.
5534 (altivec_expand_ld_builtin): Likewise.
5535 (altivec_expand_st_builtin): Likewise.
5536 (altivec_expand_vec_init_builtin): Likewise.
5537 (altivec_expand_builtin): Likewise.
5538 (rs6000_init_builtins): Add support for V1TImode type. Add
5539 support for ISA 2.07 128-bit integer builtins. Define type names
5540 for the VSX/Altivec vector types.
5541 (altivec_init_builtins): Add support for overloaded vector
5542 functions with V1TImode type.
5543 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5544 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5546 (rs6000_split_128bit_ok_p): Likewise.
5547 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5548 __int128_t and vector __uint128_t.
5550 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5551 and mode attributes.
5558 (VS_scalar): Likewise.
5559 (VS_double): Likewise.
5560 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5562 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5563 we support the ISA 2.07 128-bit integer arithmetic instructions.
5564 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5565 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5566 and TImode types for use with the builtin functions.
5567 (V1TI_type_node): Likewise.
5568 (unsigned_V1TI_type_node): Likewise.
5569 (intTI_type_internal_node): Likewise.
5570 (uintTI_type_internal_node): Likewise.
5572 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5573 128-bit builtin functions.
5574 (UNSPEC_VADDEUQM): Likewise.
5575 (UNSPEC_VADDECUQ): Likewise.
5576 (UNSPEC_VSUBCUQ): Likewise.
5577 (UNSPEC_VSUBEUQM): Likewise.
5578 (UNSPEC_VSUBECUQ): Likewise.
5579 (VM): Add V1TImode to vector mode iterators.
5581 (VI_unit): Likewise.
5582 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5583 (altivec_vaddcuq): Likewise.
5584 (altivec_vsubuqm): Likewise.
5585 (altivec_vsubcuq): Likewise.
5586 (altivec_vaddeuqm): Likewise.
5587 (altivec_vaddecuq): Likewise.
5588 (altivec_vsubeuqm): Likewise.
5589 (altivec_vsubecuq): Likewise.
5591 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5593 (BOOL_128): Likewise.
5594 (BOOL_REGS_OUTPUT): Likewise.
5595 (BOOL_REGS_OP1): Likewise.
5596 (BOOL_REGS_OP2): Likewise.
5597 (BOOL_REGS_UNARY): Likewise.
5598 (BOOL_REGS_AND_CR0): Likewise.
5600 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5601 128-bit integer builtin support.
5602 (vec_vadduqm): Likewise.
5603 (vec_vaddecuq): Likewise.
5604 (vec_vaddeuqm): Likewise.
5605 (vec_vsubecuq): Likewise.
5606 (vec_vsubeuqm): Likewise.
5607 (vec_vsubcuq): Likewise.
5608 (vec_vsubuqm): Likewise.
5610 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5611 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5612 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5613 128-bit integer add/subtract to ISA 2.07.
5615 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5617 * config/arc/arc.c (arc_predicate_delay_insns):
5618 Fix third argument passed to conditionalize_nonjump.
5620 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5622 * config/aarch64/aarch64-builtins.c
5623 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5624 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5625 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5626 instead of __builtin_lfloor.
5627 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5629 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5631 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5632 (tree_ssa_ifcombine_bb_1): New function.
5633 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5634 is an empty forwarder block to then_bb or vice versa and then_bb
5635 and else_bb are effectively swapped.
5637 2014-03-12 Christian Bruel <christian.bruel@st.com>
5640 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5641 REG_CFA_DEF_CFA note.
5642 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5643 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5645 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5647 PR tree-optimization/60454
5648 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5650 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5652 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5653 Do not define target_cpu_default2 to generic.
5654 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5655 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5656 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5658 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5659 Marc Glisse <marc.glisse@inria.fr>
5661 PR tree-optimization/60502
5662 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5663 instead of build_low_bits_mask.
5665 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5668 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5669 if there are multiple uses, but op doesn't live on E edge.
5670 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5671 clobber stmts before __builtin_unreachable.
5673 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5675 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5676 hard_frame_pointer_rtx.
5677 * cse.c (cse_insn): Remove volatile check.
5678 * cselib.c (cselib_process_insn): Likewise.
5679 * dse.c (scan_insn): Likewise.
5681 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5683 * config/arc/arc.c (conditionalize_nonjump): New function,
5685 (arc_ifcvt): ... this.
5686 (arc_predicate_delay_insns): Use it.
5688 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5690 * config/arc/predicates.md (extend_operand): During/after reload,
5691 allow const_int_operand.
5692 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5693 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5694 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5696 (umulsi3_highpart_i): Likewise.
5698 2014-03-11 Richard Biener <rguenther@suse.de>
5700 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5701 Add asserts to guard possible wrong-code bugs.
5703 2014-03-11 Richard Biener <rguenther@suse.de>
5705 PR tree-optimization/60429
5706 PR tree-optimization/60485
5707 * tree-ssa-structalias.c (set_union_with_increment): Properly
5708 take into account all fields that overlap the shifted vars.
5709 (do_sd_constraint): Likewise.
5710 (do_ds_constraint): Likewise.
5711 (get_constraint_for_ptr_offset): Likewise.
5713 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5715 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5716 (nios2_compute_frame_layout):
5717 Add calculation of cfun->machine->fp_save_offset.
5718 (nios2_expand_prologue): Correct setting of frame pointer register
5720 (nios2_expand_epilogue): Update recovery of stack pointer from
5721 frame pointer accordingly.
5722 (nios2_initial_elimination_offset): Update calculation of offset
5723 for eliminating to HARD_FRAME_POINTER_REGNUM.
5725 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5728 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5729 cgraph_get_create_node on VAR_DECLs.
5731 2014-03-10 Richard Biener <rguenther@suse.de>
5734 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5736 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5738 * config/vms/vms.opt (vms_float_format): New variable.
5740 2014-03-08 Tobias Burnus <burnus@net-b.de>
5742 * doc/invoke.texi (-fcilkplus): Update implementation status.
5744 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5745 Richard Biener <rguenther@suse.de>
5747 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5748 consistently accross all TUs.
5749 (run_gcc): Enable -fshort-double automatically at link at link-time
5750 and disallow override.
5752 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5755 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5756 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5757 if they can't be used.
5759 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5761 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5762 for Solaris 11/x86 ld.
5763 * configure: Regenerate.
5765 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5767 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5768 (LIB_TLS_SPEC): Save as ld_tls_libs.
5769 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5770 (HAVE_AS_IX86_TLSLDM): New test.
5771 * configure, config.in: Regenerate.
5772 * config/i386/i386.c (legitimize_tls_address): Fall back to
5773 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5774 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5775 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5776 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5778 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5780 * common.opt (fira-loop-pressure): Mark as optimization.
5782 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5784 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5785 an OpenMP mappable type.
5787 2014-03-06 Matthias Klose <doko@ubuntu.com>
5789 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5790 MULTILIB_OSDIRNAMES is not defined.
5792 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5793 Meador Inge <meadori@codesourcery.com>
5796 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5797 (arm_legitimize_address): Call legitimize_tls_address for any
5798 arm_tls_referenced_p expression, handle constant addend. Call it
5799 before testing for !TARGET_ARM.
5800 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5802 2014-03-06 Richard Biener <rguenther@suse.de>
5808 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5810 * tree-streamer.c (record_common_node): Assert we don't record
5811 nodes with type double.
5812 (preload_common_node): Skip type double, complex double and double
5813 pointer since it is now frontend dependent due to fshort-double option.
5815 2014-03-06 Richard Biener <rguenther@suse.de>
5817 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5818 or -fno-lto is specified and the linker has full plugin support.
5819 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5820 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5821 * lto-wrapper.c (merge_and_complain): Merge compile-time
5822 optimization levels.
5823 (run_gcc): And pass it through to the link options.
5825 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5829 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5831 * cselib.c (remove_useless_values): Skip to avoid quadratic
5832 behavior if the condition moved from...
5833 (cselib_process_insn): ... here holds.
5835 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5838 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5839 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5842 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5843 (TM_H): Add x86-tune.def.
5845 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5847 * config/aarch64/aarch64.c (generic_tunings):
5848 Use cortexa57_extra_costs.
5850 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5853 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5854 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5855 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5858 2014-03-04 Heiher <r@hev.cc>
5860 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5861 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5863 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5865 * config/i386/predicates.md (const2356_operand): Change to ...
5866 (const2367_operand): ... this.
5867 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5869 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5870 (*avx512pf_scatterpf<mode>sf): Ditto.
5871 (avx512pf_scatterpf<mode>df): Ditto.
5872 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5873 (*avx512pf_scatterpf<mode>df): Ditto.
5874 * config/i386/i386.c (ix86_expand_builtin): Update
5875 incorrect hint operand error message.
5877 2014-03-04 Richard Biener <rguenther@suse.de>
5879 * lto-section-in.c (lto_get_section_data): Fix const cast.
5881 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5883 * tree-streamer.c (record_common_node): Assert we don't record
5884 nodes with type double.
5885 (preload_common_node): Skip type double, complex double and double
5886 pointer since it is now frontend dependent due to fshort-double option.
5888 2014-03-04 Richard Biener <rguenther@suse.de>
5891 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5892 (lto_input_toplevel_asms): Likewise.
5893 * lto-section-in.c (lto_get_section_data): Instead do it here
5896 2014-03-04 Richard Biener <rguenther@suse.de>
5898 PR tree-optimization/60382
5899 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5900 dead PHIs a reduction.
5902 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5904 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5906 (_mm_prefetch): Move out of GCC target("sse") pragma.
5907 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5908 GCC target("prfchw") pragma.
5909 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5911 * config/i386/i386.c (ix86_option_override_internal): Enable
5912 -mprfchw with -mprefetchwt1.
5914 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5916 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5919 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5921 * opts.h (CL_PCH_IGNORE): Define.
5922 * targhooks.c (option_affects_pch_p):
5923 Return false for options that have CL_PCH_IGNORE set.
5924 * opt-functions.awk: Process PchIgnore.
5925 * doc/options.texi: Document PchIgnore.
5927 * config/arc/arc.opt (misize): Add PchIgnore property.
5929 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5931 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5932 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5933 constraint on constants to permit them being loaded into
5934 GENERAL_REGS or BASE_REGS.
5936 2014-03-03 Nick Clifton <nickc@redhat.com>
5938 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5939 anti-cacnonical alternatives.
5940 (negandhi3_real): New pattern.
5941 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5943 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5945 * config/avr/avr-mcus.def: Remove atxmega16x1.
5946 * config/avr/avr-tables.opt: Regenerate.
5947 * config/avr/t-multilib: Regenerate.
5948 * doc/avr-mmcu.texi: Regenerate.
5950 2014-03-03 Tobias Grosser <tobias@grosser.es>
5951 Mircea Namolaru <mircea.namolaru@inria.fr>
5953 PR tree-optimization/58028
5954 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5957 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5959 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5960 not handled by recognizers.
5962 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5965 * function.c (expand_function_end): Don't emit
5966 clobber_return_register sequence if clobber_after is a BARRIER.
5967 * cfgexpand.c (construct_exit_block): Append instructions before
5968 return_label to prev_bb.
5970 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5972 * config/rs6000/constraints.md: Document reserved use of "wc".
5974 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5977 * ipa.c (function_and_variable_visibility): When dissolving comdat
5978 group, also set all symbols to local.
5980 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5985 2013-12-14 Jan Hubicka <jh@suse.cz>
5987 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5989 2014-03-02 Jon Beniston <jon@beniston.com>
5995 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5996 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5997 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5998 (simple_return, *simple_return): New patterns
5999 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6000 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6002 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6004 * dwarf2out.c (gen_subprogram_die): Tidy.
6006 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6009 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6010 (*mov_t_msb_neg_negc): ... this new insn.
6012 2014-02-28 Jason Merrill <jason@redhat.com>
6015 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6018 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6021 * dwarf2out.c (decltype_auto_die): New static.
6022 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6023 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6024 (is_cxx_auto): Likewise.
6026 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6028 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6029 we are not using general regs only.
6031 2014-02-28 Richard Biener <rguenther@suse.de>
6034 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6035 previous fix and only allow to remove trivial pre-headers
6036 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6037 (remove_forwarder_block): Properly update the latch of a loop.
6039 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6042 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6043 (cselib_preserved_hash_table): New.
6044 (preserve_constants_and_equivs): Move preserved vals to it.
6045 (cselib_find_slot): Look it up first.
6046 (cselib_init): Initialize it.
6047 (cselib_finish): Release it.
6048 (dump_cselib_table): Dump it.
6050 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6053 * cselib.c (remove_useless_values): Skip to avoid quadratic
6054 behavior if the condition moved from...
6055 (cselib_process_insn): ... here holds.
6057 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6060 * var-tracking.c (vt_initialize): Apply the same condition to
6061 preserve the CFA base value.
6063 2014-02-28 Joey Ye <joey.ye@arm.com>
6066 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6067 if reload in progress or completed.
6069 2014-02-28 Tobias Burnus <burnus@net-b.de>
6072 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6075 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6077 * doc/tm.texi.in (Condition Code Status): Update documention for
6078 relative locations of cc0-setter and cc0-user.
6080 2014-02-27 Jeff Law <law@redhat.com>
6082 PR rtl-optimization/52714
6083 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6084 into two independent simple sets, if I3 is a jump, ensure the
6085 pattern we place into I3 is a (set (pc) ...).
6087 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6088 Jeff Law <law@redhat.com>
6090 PR rtl-optimization/49847
6091 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6092 are in different blocks.
6093 * doc/tm.texi (Condition Code Status): Update documention for
6094 relative locations of cc0-setter and cc0-user.
6096 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6099 * lra.c (lra_emit_add): Check SUBREG too.
6101 2014-02-27 Andreas Schwab <schwab@suse.de>
6103 * config/m68k/m68k.c (m68k_option_override): Disable
6104 -flive-range-shrinkage for classic m68k.
6105 (m68k_override_options_after_change): Likewise.
6107 2014-02-27 Marek Polacek <polacek@redhat.com>
6110 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6111 -Wmaybe-uninitialized.
6113 2014-02-27 Alan Modra <amodra@gmail.com>
6116 * reload1.c (emit_input_reload_insns): When reload_override_in,
6117 set old to rl->in_reg when rl->in_reg is a subreg.
6119 2014-02-26 Richard Biener <rguenther@suse.de>
6122 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6124 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6126 * common/config/i386/predicates.md (const1256_operand): Remove.
6127 (const2356_operand): New.
6128 (const_1_to_2_operand): Remove.
6129 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6130 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6131 (*avx512pf_gatherpf<mode>sf): Ditto.
6132 (avx512pf_gatherpf<mode>df): Ditto.
6133 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6134 (*avx512pf_gatherpf<mode>df): Ditto.
6135 (avx512pf_scatterpf<mode>sf): Ditto.
6136 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6137 (*avx512pf_scatterpf<mode>sf): Ditto.
6138 (avx512pf_scatterpf<mode>df): Ditto.
6139 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6140 (*avx512pf_scatterpf<mode>df): Ditto.
6141 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6143 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6145 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6146 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6147 (_mm512_mask_testn_epi64_mask): Move to ...
6148 * config/i386/avx512cdintrin.h: Here.
6149 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6150 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6151 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6152 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6153 TARGET_AVX512F from TARGET_AVX512CD.
6155 2014-02-26 Richard Biener <rguenther@suse.de>
6158 * ipa.c (walk_polymorphic_call_targets): Properly guard
6159 call to inline_update_overall_summary.
6161 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6164 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6165 and latches only if requested. Fix latch if it is removed.
6166 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6167 LOOPS_HAVE_PREHEADERS.
6169 2014-02-25 Andrew Pinski <apinski@cavium.com>
6171 * builtins.c (expand_builtin_thread_pointer): Create a new target
6172 when the target is NULL.
6174 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6176 PR rtl-optimization/60317
6177 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6178 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6179 * lra-assigns.c: Include params.h.
6180 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6181 other reload pseudos considerations.
6183 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6185 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6186 to use canonical form for nor<mode>3.
6188 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6191 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6194 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6196 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6197 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6198 (ix86_handle_option): Handle OPT_mprefetchwt1.
6199 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6200 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6202 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6203 OPTION_MASK_ISA_PREFETCHWT1.
6204 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6205 (PTA_PREFETCHWT1): New.
6206 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6207 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6208 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6209 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6210 (*prefetch_avx512pf_<mode>_: Change into ...
6211 (*prefetch_prefetchwt1_<mode>: This.
6212 * config/i386/i386.opt (mprefetchwt1): New.
6213 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6214 (_mm_prefetch): Handle intent to write.
6215 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6217 2014-02-25 Richard Biener <rguenther@suse.de>
6220 * emit-rtl.c (mem_attrs_htab): Remove.
6221 (mem_attrs_htab_hash): Likewise.
6222 (mem_attrs_htab_eq): Likewise.
6223 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6224 (init_emit_once): Do not allocate mem_attrs_htab.
6226 2014-02-25 Richard Biener <rguenther@suse.de>
6229 * lto-opts.c (lto_write_options): Output non-explicit conservative
6230 -fwrapv, -fno-trapv and -fno-strict-overflow.
6231 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6232 (run_gcc): And pass them through.
6234 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6236 * sel-sched.c (calculate_new_fences): New parameter ptime.
6237 Calculate it as a maximum over all fence cycles.
6238 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6239 Print the final schedule timing when sched_verbose.
6241 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6243 PR rtl-optimization/60292
6244 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6245 bit fot the fence instruction.
6247 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6249 * calls.h: Fix typo in comment.
6251 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6253 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6254 adjusting offsetable addresses.
6256 2014-02-24 Guozhi Wei <carrot@google.com>
6258 * sparseset.h (sparseset_pop): Fix the wrong index.
6260 2014-02-24 Walter Lee <walt@tilera.com>
6262 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6263 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6265 * common/config/tilegx/tilegx-common.c
6266 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6267 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6269 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6270 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6271 (tilegx_gimplify_va_arg_expr): Handle big endian.
6272 (tilegx_expand_unaligned_load): Ditto.
6273 (tilegx_expand_unaligned_store): Ditto.
6274 (TARGET_RETURN_IN_MSB): New.
6275 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6276 (TARGET_ENDIAN_DEFAULT): New.
6277 (TARGET_BIG_ENDIAN): Handle big endian.
6278 (BYTES_BIG_ENDIAN): Ditto.
6279 (WORDS_BIG_ENDIAN): Ditto.
6280 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6283 * config/tilegx/tilegx.md (extv): Handle big endian.
6285 (insn_st<n>): Ditto.
6286 (insn_st<n>_add<bitsuffix>): Ditto.
6287 (insn_stnt<n>): Ditto.
6288 (insn_stnt<n>_add<bitsuffix>):Ditto.
6289 (vec_interleave_highv8qi): Handle big endian.
6290 (vec_interleave_highv8qi_be): New.
6291 (vec_interleave_highv8qi_le): New.
6292 (insn_v1int_h): Handle big endian.
6293 (vec_interleave_lowv8qi): Handle big endian.
6294 (vec_interleave_lowv8qi_be): New.
6295 (vec_interleave_lowv8qi_le): New.
6296 (insn_v1int_l): Handle big endian.
6297 (vec_interleave_highv4hi): Handle big endian.
6298 (vec_interleave_highv4hi_be): New.
6299 (vec_interleave_highv4hi_le): New.
6300 (insn_v2int_h): Handle big endian.
6301 (vec_interleave_lowv4hi): Handle big endian.
6302 (vec_interleave_lowv4hi_be): New.
6303 (vec_interleave_lowv4hi_le): New.
6304 (insn_v2int_l): Handle big endian.
6305 (vec_interleave_highv2si): Handle big endian.
6306 (vec_interleave_highv2si_be): New.
6307 (vec_interleave_highv2si_le): New.
6308 (insn_v4int_h): Handle big endian.
6309 (vec_interleave_lowv2si): Handle big endian.
6310 (vec_interleave_lowv2si_be): New.
6311 (vec_interleave_lowv2si_le): New.
6312 (insn_v4int_l): Handle big endian.
6313 * config/tilegx/tilegx.opt (mbig-endian): New option.
6314 (mlittle-endian): New option.
6315 * doc/install.texi: Document tilegxbe-linux.
6316 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6318 2014-02-24 Martin Jambor <mjambor@suse.cz>
6321 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6322 there are no parameter descriptors.
6324 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6326 PR rtl-optimization/60268
6327 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6328 initialization to ...
6329 (sched_rgn_init): ... here.
6330 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6332 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6334 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6337 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6339 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6342 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6344 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6345 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6347 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6349 * config/microblaze/predicates.md: Add cmp_op predicate.
6350 * config/microblaze/microblaze.md: Add branch_compare instruction
6351 which uses cmp_op predicate and emits cmp insn before branch.
6352 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6353 to microblaze_expand_conditional_branch and consolidate logic.
6354 (microblaze_expand_conditional_branch): emit branch_compare
6355 insn instead of handling cmp op separate from branch insn.
6357 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6359 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6362 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6364 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6365 define_insn with define_expand and new define_insn
6366 *altivec_lve<VI_char>x_internal.
6367 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6368 and new define_insn *altivec_stve<VI_char>x_internal.
6369 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6371 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6373 (altivec_expand_stvex_be): New function.
6375 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6377 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6378 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6379 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6380 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6382 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6385 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6386 instead of emit_move_insn.
6388 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6390 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6392 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6393 gen_altivec_vsumsws.
6395 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6397 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6398 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6399 (altivec_lvxl_<mode>): New define_expand incorporating
6400 -maltivec=be semantics where needed.
6401 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6402 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6403 semantics where needed.
6404 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6405 (altivec_stvx_<mode>): New define_expand incorporating
6406 -maltivec=be semantics where needed.
6407 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6408 VM2 iterator instead of V4SI.
6409 (altivec_stvxl_<mode>): New define_expand incorporating
6410 -maltivec=be semantics where needed.
6411 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6412 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6413 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6414 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6415 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6416 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6417 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6418 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6419 ALTIVEC_BUILTIN_STVXL.
6420 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6421 (altivec_expand_stvx_be): Likewise.
6422 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6423 (altivec_expand_lvx_be): Likewise.
6424 (altivec_expand_stvx_be): Likewise.
6425 (altivec_expand_builtin): Add cases for
6426 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6427 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6428 (altivec_init_builtins): Add definitions for
6429 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6430 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6432 2014-02-21 Catherine Moore <clm@codesourcery.com>
6434 * doc/invoke.texi (mvirt, mno-virt): Document.
6435 * config/mips/mips.opt (mvirt): New option.
6436 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6438 2014-02-21 Richard Biener <rguenther@suse.de>
6440 PR tree-optimization/60276
6441 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6442 (STMT_VINFO_MIN_NEG_DIST): New macro.
6443 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6444 STMT_VINFO_MIN_NEG_DIST.
6445 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6446 made for negative dependence distances still hold.
6448 2014-02-21 Richard Biener <rguenther@suse.de>
6451 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6452 DECL_INITIAL for globals not in the current function context.
6454 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6456 PR tree-optimization/56490
6457 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6458 * tree-ssa-uninit.c: Include params.h.
6459 (compute_control_dep_chain): Add num_calls argument, return false
6460 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6461 num_calls to recursive call.
6462 (find_predicates): Change dep_chain into normal array,
6463 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6464 variable and adjust compute_control_dep_chain caller.
6465 (find_def_preds): Likewise.
6467 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6469 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6470 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6472 2014-02-21 Nick Clifton <nickc@redhat.com>
6474 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6475 (pushhi1): Likewise.
6476 (popqi1): Add mode to pre_dec.
6479 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6481 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6482 mode for mask of V8SFmode permutation.
6484 2014-02-20 Richard Henderson <rth@redhat.com>
6487 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6488 a new pseudo for OLDVAL.
6490 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6493 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6494 gen_reg_rtx if d->testing_p.
6495 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6496 if d->testing_p and we will certainly return true.
6497 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6500 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6502 * emit-rtl.c (gen_reg_rtx): Assert that
6503 crtl->emit.regno_pointer_align_length is non-zero.
6505 2014-02-20 Richard Henderson <rth@redhat.com>
6508 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6509 on failure the store back into EXPECT.
6511 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6512 Sandra Loosemore <sandra@codesourcery.com>
6514 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6515 * config/nios2/nios2.c (nios2_function_profiler): Add
6516 -fPIC (flag_pic == 2) support.
6517 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6518 (nios2_large_offset_p): New function.
6519 (nios2_unspec_reloc_p): Move up position, update to use
6520 nios2_large_offset_p.
6521 (nios2_unspec_address): Remove function.
6522 (nios2_unspec_offset): New function.
6523 (nios2_large_got_address): New function.
6524 (nios2_got_address): Add large offset support.
6525 (nios2_legitimize_tls_address): Update usage of removed and new
6527 (nios2_symbol_binds_local_p): New function.
6528 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6529 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6530 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6531 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6532 processing for (const (unspec ...)).
6533 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6535 2014-02-20 Richard Biener <rguenther@suse.de>
6537 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6538 doing the substitution.
6539 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6541 2014-02-20 Martin Jambor <mjambor@suse.cz>
6544 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6545 info when checking whether lattices are bottom.
6547 2014-02-20 Richard Biener <rguenther@suse.de>
6550 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6553 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6556 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6557 parameter specifying the scaling.
6558 (inline_call): Update.
6559 (want_inline_recursively): Guard division by zero.
6560 (recursive_inlining): Update.
6561 * ipa-inline.h (clone_inlined_nodes): Update.
6563 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6566 * config/i386/i386.c (classify_argument): Pass structures of size
6567 64 bytes or less in register.
6569 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6570 Kirill Yukhin <kirill.yukhin@intel.com>
6572 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6573 (_mm_rcp28_round_ss): Ditto.
6574 (_mm_rsqrt28_round_sd): Ditto.
6575 (_mm_rsqrt28_round_ss): Ditto.
6576 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6577 (_mm_rcp14_round_ss): Ditto.
6578 (_mm_rsqrt14_round_sd): Ditto.
6579 (_mm_rsqrt14_round_ss): Ditto.
6580 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6581 the first input operand, get rid of match_dup.
6582 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6584 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6586 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6587 operand as the first input operand, set type attribute.
6588 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6590 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6591 operand as the first input operand, set type attribute.
6593 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6595 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6598 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6601 * config/i386/i386.c (construct_container): Remove TFmode check
6602 for X86_64_INTEGER_CLASS.
6604 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6607 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6608 only when -Wpsabi is enabled.
6610 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6613 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6614 passing arrays in registers are the same as for structs, so remove the
6615 special case for them.
6617 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6619 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6620 destination type, extract only the valid bits if the source type is not
6621 integral and has a different mode.
6623 2014-02-19 Richard Biener <rguenther@suse.de>
6626 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6629 2014-02-19 Richard Biener <rguenther@suse.de>
6632 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6633 (ipa_modify_call_arguments): Emit an argument load explicitely and
6634 preserve virtual SSA form there and for the replacement call.
6635 Do not update SSA form nor free dominance info.
6637 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6639 * ipa.c (function_and_variable_visibility): Also clear WEAK
6640 flag when disolving COMDAT_GROUP.
6642 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6644 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6645 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6647 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6648 do more sanity checks.
6649 (detect_type_change): Return true when giving up early.
6650 (compute_complex_assign_jump_func): Fix type parameter of
6651 ipa_set_ancestor_jf.
6652 (compute_complex_ancestor_jump_func): Likewise.
6653 (update_jump_functions_after_inlining): Fix updating of
6655 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6657 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6659 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6660 inline clones when edge disappears.
6662 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6665 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6666 Split 64-bit moves into 2 patterns. Do not allow the use of
6667 direct move for TDmode in little endian, since the decimal value
6668 has little endian bytes within a word, but the 64-bit pieces are
6669 ordered in a big endian fashion, and normal subreg's of TDmode are
6671 (mov<mode>_64bit_dm): Likewise.
6672 (movtd_64bit_nodm): Likewise.
6674 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6676 PR tree-optimization/60174
6677 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6678 statement of an SSA_NAME that occurs in an abnormal PHI node.
6680 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6683 * final.c (SEEN_BB): Remove.
6684 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6685 (final_scan_insn): Don't force_source_line on second
6686 NOTE_INSN_BASIC_BLOCK.
6688 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6691 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6692 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6693 (type_natural_mode): Warn ABI change when %zmm register is not
6694 available for AVX512F vector value passing.
6696 2014-02-18 Kai Tietz <ktietz@redhat.com>
6699 * config/i386/i386.c (ix86_expand_prologue): Use value in
6700 rax register as displacement when restoring %r10 or %rax.
6701 Fix wrong offset when restoring both registers.
6703 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6705 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6706 assertion with conditional return.
6708 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6709 Uros Bizjak <ubizjak@gmail.com>
6712 * config/i386/driver-i386.c (host_detect_local_cpu): If
6713 YMM state is not saved by the OS, also clear has_f16c. Move
6714 CPUID 0x80000001 handling before YMM state saving checking.
6716 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6718 PR rtl-optimization/58960
6719 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6720 factored out from ...
6721 (sched_init): ... here.
6722 (free_global_sched_pressure_data): New, factored out from ...
6723 (sched_finish): ... here.
6724 * sched-int.h (free_global_sched_pressure_data): Declare.
6725 * sched-rgn.c (nr_regions_initial): New static global.
6726 (haifa_find_rgns): Initialize it.
6727 (schedule_region): Disable sched-pressure for the newly
6730 2014-02-17 Richard Biener <rguenther@suse.de>
6732 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6733 release SSA defs of pattern stmts.
6735 2014-02-17 Richard Biener <rguenther@suse.de>
6737 * tree-inline.c (expand_call_inline): Release the virtual
6738 operand defined by the call we are about to inline.
6740 2014-02-17 Richard Biener <rguenther@suse.de>
6742 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6744 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6745 Ilya Tocar <ilya.tocar@intel.com>
6747 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6748 arguments order in builtin.
6749 (_mm512_permutexvar_epi64): Ditto.
6750 (_mm512_mask_permutexvar_epi64): Ditto
6751 (_mm512_maskz_permutexvar_epi32): Ditto
6752 (_mm512_permutexvar_epi32): Ditto
6753 (_mm512_mask_permutexvar_epi32): Ditto
6755 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6757 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6758 (p8_vmrgow): Likewise.
6760 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6762 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6765 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6768 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6769 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6770 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6771 using direct move instructions on ISA 2.07. Also adjust
6772 instruction length for 64-bit.
6773 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6774 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6776 2014-02-15 Alan Modra <amodra@gmail.com>
6780 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6781 find_replacement on parts of insn rtl that might be reloaded.
6783 2014-02-15 Richard Biener <rguenther@suse.de>
6785 PR tree-optimization/60183
6786 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6787 (tree_ssa_phiprop): Calculate and free post-dominators.
6789 2014-02-14 Jeff Law <law@redhat.com>
6791 PR rtl-optimization/60131
6792 * ree.c (get_extended_src_reg): New function.
6793 (combine_reaching_defs): Use it rather than assuming location of REG.
6794 (find_and_remove_re): Verify first operand of extension is
6795 a REG before adding the insns to the copy list.
6797 2014-02-14 Roland McGrath <mcgrathr@google.com>
6799 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6800 * configure: Regenerated.
6801 * config.in: Regenerated.
6802 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6803 instead of ASM_SHORT.
6805 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6806 Richard Earnshaw <rearnsha@arm.com>
6808 PR rtl-optimization/59535
6809 * lra-constraints.c (process_alt_operands): Encourage alternative
6810 when unassigned pseudo class is superset of the alternative class.
6811 (inherit_reload_reg): Don't inherit when optimizing for code size.
6812 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6813 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6814 modes not less than 4 for Thumb1.
6816 2014-02-14 Kyle McMartin <kyle@redhat.com>
6819 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6821 2014-02-14 Richard Biener <rguenther@suse.de>
6823 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6824 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6825 Do not strip INDIRECT_REFs.
6827 2014-02-14 Richard Biener <rguenther@suse.de>
6830 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6831 DECL_FUNCTION_SPECIFIC_TARGET.
6832 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6833 * tree-streamer-out.c (pack_ts_target_option): Remove.
6834 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6835 (write_ts_function_decl_tree_pointers): Do not stream
6836 DECL_FUNCTION_SPECIFIC_TARGET.
6837 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6838 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6839 (lto_input_ts_function_decl_tree_pointers): Do not stream
6840 DECL_FUNCTION_SPECIFIC_TARGET.
6842 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6844 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6845 (get_initial_def_for_induction, vectorizable_induction): Ignore
6846 debug stmts when looking for exit_phi.
6847 (vectorizable_live_operation): Fix up condition.
6849 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6851 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6852 nreverse() because it changes the content of original tree list.
6854 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6856 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6857 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6859 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6861 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6862 GNU coding standards.
6864 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6867 * dwarf2out.c (gen_subprogram_die): Don't call
6868 add_calling_convention_attribute if subr_die is old_die.
6870 2014-02-13 Sharad Singhai <singhai@google.com>
6872 * doc/optinfo.texi: Fix order of nodes.
6874 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6876 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6877 operands[2], not operands[3].
6879 2014-02-13 Richard Biener <rguenther@suse.de>
6882 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6883 mention the possibility of an in-tree build.
6884 (CLooG): Update recommended version to 0.18.1, mention the
6885 possibility of an in-tree build and clarify that the ISL
6886 bundled with CLooG does not work.
6888 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6891 * expr.c (compress_float_constant): If x is a hard register,
6892 extend into a pseudo and then move to x.
6894 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6896 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6897 caused by bad second argument to warning_at() with -mhotpatch and
6898 nested functions (e.g. with gfortran).
6900 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6902 * opts.c (option_name): Remove "enabled by default" rider.
6904 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6906 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6908 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6909 Uros Bizjak <ubizjak@gmail.com>
6912 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6913 * configure: Regenerated.
6915 2014-02-12 Richard Biener <rguenther@suse.de>
6917 * vec.c (vec_prefix::calculate_allocation): Move as
6918 inline variant to vec.h.
6919 (vec_prefix::calculate_allocation_1): New out-of-line version.
6920 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6921 (vec_prefix::m_has_auto_buf): Rename to ...
6922 (vec_prefix::m_using_auto_storage): ... this.
6923 (vec_prefix::calculate_allocation): Inline the easy cases
6924 and dispatch to calculate_allocation_1 which doesn't need the
6926 (va_heap::reserve): Use gcc_checking_assert.
6927 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6928 m_using_auto_storage.
6929 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6931 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6932 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6933 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6935 2014-02-12 Richard Biener <rguenther@suse.de>
6937 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6938 when we found a dependence.
6940 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6942 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6944 (maybe_fold_stmt): ... into this new function.
6945 * omp-low.c (lower_omp): Update comment.
6947 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6950 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6953 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6955 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6956 identifiers in comments.
6957 (cortexa53_extra_costs): Likewise.
6958 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6959 (cortexa7_extra_costs): Likewise.
6960 (cortexa12_extra_costs): Likewise.
6961 (cortexa15_extra_costs): Likewise.
6962 (v7m_extra_costs): Likewise.
6964 2014-02-12 Richard Biener <rguenther@suse.de>
6967 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6968 of posix_memalign being successful.
6969 (lower_stmt): Restrict lowering of posix_memalign to when
6970 -ftree-bit-ccp is enabled.
6972 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6974 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6976 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6978 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6980 PR rtl-optimization/60116
6981 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6982 other_insn once the combination has been validated.
6984 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6987 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6989 * ipa-devirt.c: Include demangle.h
6990 (odr_violation_reported): New static variable.
6991 (add_type_duplicate): Update odr_violations.
6992 (maybe_record_node): Add completep parameter; update it.
6993 (record_target_from_binfo): Add COMPLETEP parameter;
6994 update it as needed.
6995 (possible_polymorphic_call_targets_1): Likewise.
6996 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6997 rename FINAL to COMPLETE.
6998 (record_targets_from_bases): Sanity check we found the binfo;
6999 fix COMPLETEP updating.
7000 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7001 parameter, fix computing of COMPLETEP.
7002 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7003 at LTO time do demangling.
7004 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7005 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7007 (gimple_get_virt_method_for_binfo): Likewise.
7008 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7009 gimple_get_virt_method_for_vtable): Update prototypes.
7011 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7014 * genautomata.c (add_presence_absence): Fix typo with
7015 {final_}presence_list.
7017 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7020 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7021 for VSX/Altivec vectors that land in GPR registers.
7023 2014-02-11 Richard Henderson <rth@redhat.com>
7024 Jakub Jelinek <jakub@redhat.com>
7027 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7028 around drhs if type conversion to lacc->type is not useless.
7030 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7032 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7034 (cortex-a57.cortex-a53): Likewise.
7035 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7037 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7039 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7042 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7044 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7045 add_options_for_arm_vfp3.
7047 2014-02-11 Jeff Law <law@redhat.com>
7050 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7051 object with an undesirable mode.
7053 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7056 * config/i386/sol2-9.h: New file.
7057 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7058 *-*-solaris2.9*): Use it.
7060 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7062 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7063 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7065 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7067 * config/microblaze/microblaze.c: Extend mcpu version format
7069 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7071 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7073 2014-02-10 Richard Henderson <rth@redhat.com>
7076 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7077 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7078 ms-abi vs -mno-accumulate-outgoing-args.
7079 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7080 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7083 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7086 * cfgexpand.c (expand_asm_operands): Attach source location to
7087 ASM_INPUT rtx objects.
7088 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7090 2014-02-10 Nick Clifton <nickc@redhat.com>
7092 * config/mn10300/mn10300.c (popcount): New function.
7093 (mn10300_expand_prologue): Include saved registers in stack usage
7096 2014-02-10 Jeff Law <law@redhat.com>
7099 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7100 when changing the SET_DEST of a prior insn to avoid an input reload.
7102 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7104 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7105 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7106 -mcall-openbsd, or -mcall-linux.
7107 (CC1_ENDIAN_BIG_SPEC): Remove.
7108 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7109 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7110 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7111 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7112 and %cc1_endian_default.
7113 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7115 2014-02-10 Richard Biener <rguenther@suse.de>
7117 PR tree-optimization/60115
7118 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7119 MEM_REF handling. Properly verify that the accesses are not
7120 out of the objects bound.
7122 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7124 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7127 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7129 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7130 proper constants and fix formatting.
7131 (possible_polymorphic_call_targets): Fix formatting.
7133 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7134 Ilya Tocar <ilya.tocar@intel.com>
7136 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7137 (_mm512_loadu_epi32): Renamed into...
7138 (_mm512_loadu_si512): This.
7139 (_mm512_storeu_epi32): Renamed into...
7140 (_mm512_storeu_si512): This.
7141 (_mm512_maskz_ceil_ps): Removed.
7142 (_mm512_maskz_ceil_pd): Ditto.
7143 (_mm512_maskz_floor_ps): Ditto.
7144 (_mm512_maskz_floor_pd): Ditto.
7145 (_mm512_floor_round_ps): Ditto.
7146 (_mm512_floor_round_pd): Ditto.
7147 (_mm512_ceil_round_ps): Ditto.
7148 (_mm512_ceil_round_pd): Ditto.
7149 (_mm512_mask_floor_round_ps): Ditto.
7150 (_mm512_mask_floor_round_pd): Ditto.
7151 (_mm512_mask_ceil_round_ps): Ditto.
7152 (_mm512_mask_ceil_round_pd): Ditto.
7153 (_mm512_maskz_floor_round_ps): Ditto.
7154 (_mm512_maskz_floor_round_pd): Ditto.
7155 (_mm512_maskz_ceil_round_ps): Ditto.
7156 (_mm512_maskz_ceil_round_pd): Ditto.
7157 (_mm512_expand_pd): Ditto.
7158 (_mm512_expand_ps): Ditto.
7159 * config/i386/i386.c (ix86_builtins): Remove
7160 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7161 (bdesc_args): Ditto.
7162 * config/i386/predicates.md (const1256_operand): New.
7163 (const_1_to_2_operand): Ditto.
7164 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7165 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7166 (*avx512pf_gatherpf<mode>sf): Ditto.
7167 (avx512pf_gatherpf<mode>df): Ditto.
7168 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7169 (*avx512pf_gatherpf<mode>df): Ditto.
7170 (avx512pf_scatterpf<mode>sf): Ditto.
7171 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7172 (*avx512pf_scatterpf<mode>sf): Ditto.
7173 (avx512pf_scatterpf<mode>df): Ditto.
7174 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7175 (*avx512pf_scatterpf<mode>df): Ditto.
7176 (avx512f_expand<mode>): Removed.
7177 (<shift_insn><mode>3<mask_name>): Change predicate type.
7179 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7181 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7182 not at the end of datarefs vector use ordered_remove to avoid
7183 reordering datarefs vector.
7186 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7187 mark local addressable non-static vars as GOVD_PRIVATE
7188 instead of GOVD_LOCAL.
7189 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7190 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7194 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7195 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7196 assume_aligned or alloc_align attributes.
7197 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7198 arguments. Handle also assume_aligned and alloc_align attributes.
7199 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7200 calls to functions with assume_aligned or alloc_align attributes.
7201 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7203 2014-02-08 Terry Guo <terry.guo@arm.com>
7205 * doc/invoke.texi: Document ARM -march=armv7e-m.
7207 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7209 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7210 flag on __cilkrts_rethrow builtin.
7213 * ipa-cp.c (determine_versionability): Fail at -O0
7214 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7215 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7218 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7221 * tree-inline.c (copy_forbidden): Fail for
7222 __attribute__((optimize (0))) functions.
7224 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7226 * varpool.c: Include pointer-set.h.
7227 (varpool_remove_unreferenced_decls): Variables in other partitions
7228 will not be output; be however careful to not lose information
7231 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7233 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7234 lookup in the vtable constructor.
7236 2014-02-07 Jeff Law <law@redhat.com>
7239 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7240 define_insn_and_split.
7242 * ipa-inline.c (inline_small_functions): Fix typos.
7244 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7246 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7247 (s390_can_use_return_insn): Declare.
7248 * config/s390/s390.h (EPILOGUE_USES): Define.
7249 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7251 (s390_chunkify_start): Handle return JUMP_LABELs.
7252 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7253 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7254 (s390_can_use_return_insn): New functions.
7255 (s390_fix_long_loop_prediction): Handle conditional returns.
7256 (TARGET_SET_UP_BY_PROLOGUE): Define.
7257 * config/s390/s390.md (ANY_RETURN): New code iterator.
7258 (*creturn, *csimple_return, return, simple_return): New patterns.
7260 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7262 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7263 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7264 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7265 REG_CFA_RESTORE list when deciding not to restore a register.
7267 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7269 * config/s390/s390.c: Include tree-pass.h and context.h.
7270 (s390_early_mach): New function, split out from...
7271 (s390_emit_prologue): ...here.
7272 (pass_data_s390_early_mach): New pass structure.
7273 (pass_s390_early_mach): New class.
7274 (s390_option_override): Create and register early_mach pass.
7275 Move to end of file.
7277 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7279 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7280 to match for the exit block.
7282 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7284 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7285 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7286 Reject misaligned operands.
7288 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7290 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7292 2014-02-07 Richard Biener <rguenther@suse.de>
7295 * gimple-low.c (lower_builtin_posix_memalign): New function.
7296 (lower_stmt): Call it to lower posix_memalign in a way
7297 to make alignment info accessible.
7299 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7302 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7303 __builtin_setjmp_receiver.
7305 2014-02-07 Richard Biener <rguenther@suse.de>
7308 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7309 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7310 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7311 Handle BUILT_IN_POSIX_MEMALIGN.
7312 (find_func_clobbers): Likewise.
7313 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7314 (call_may_clobber_ref_p_1): Likewise.
7316 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7319 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7322 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7325 * lto-cgraph.c (lto_output_node): Use
7326 symtab_get_symbol_partitioning_class.
7327 (lto_output_varpool_node): likewise.
7328 (symtab_get_symbol_partitioning_class): Move here from
7330 * cgraph.h (symbol_partitioning_class): Likewise.
7331 (symtab_get_symbol_partitioning_class): Declare.
7333 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7335 * ggc.h (ggc_internal_cleared_alloc): New macro.
7336 * vec.h (vec_safe_copy): Handle memory stats.
7337 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7338 * target-globals.c (save_target_globals): Likewise.
7340 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7343 * expr.c (emit_move_resolve_push): Export; be bit more selective
7344 on when to clear alias set.
7345 * expr.h (emit_move_resolve_push): Declare.
7346 * function.h (struct function): Add tail_call_marked.
7347 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7348 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7349 * config/i386/i386.md (TImode move expander): De not call
7351 (FP push expanders): Preserve memory attributes.
7352 * config/i386/sse.md (push<mode>1): Remove.
7353 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7354 (ix86_expand_push): Remove.
7355 * config/i386/mmx.md (push<mode>1): Remove.
7357 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7359 PR rtl-optimization/60030
7360 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7361 lopart with paradoxical subreg before shifting it up by hprec.
7363 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7365 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7366 Remove extra newline at end of file.
7367 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7368 (arm_issue_rate): Handle cortexa57.
7369 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7370 (cortex-a57.cortex-a53): Likewise.
7372 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7375 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7376 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7378 (arm_expand_prologue): Adjust all callers.
7379 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7380 info, registers also at the lowest numbered registers side. Use
7381 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7385 * var-tracking.c (adjust_mems): Before adding a SET to
7386 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7388 2014-02-06 Alan Modra <amodra@gmail.com>
7391 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7392 change SDmode to DDmode when lra_in_progress.
7394 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7397 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7398 free_data_ref on the dr first, and before goto again also set dr
7399 to the next dr. For simd_lane_access, free old datarefs[i] before
7400 overwriting it. For get_vectype_for_scalar_type failure, don't
7401 free_data_ref if simd_lane_access.
7403 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7406 * tree.h (opts_for_fn): New inline function.
7407 (opt_for_fn): Define.
7408 * config/i386/i386.c (ix86_function_regparm): Use
7409 opt_for_fn (decl, optimize) instead of optimize.
7411 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7413 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7414 for SYMBOL_REF in large memory model.
7416 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7418 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7420 (cortex-a57): Likewise.
7421 (cortex-a57.cortex-a53): Likewise.
7423 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7424 Kugan Vivekanandarajah <kuganv@linaro.org>
7426 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7428 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7430 2014-02-06 Richard Biener <rguenther@suse.de>
7432 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7433 set_loop_copy and initialize_original_copy_tables.
7435 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7437 * config/aarch64/aarch64-simd.md
7438 (aarch64_ashr_simddi): Change QI to SI.
7440 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7441 Jakub Jelinek <jakub@redhat.com>
7444 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7447 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7449 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7450 CODE_FOR_altivec_vpku[hw]um to
7451 CODE_FOR_altivec_vpku[hw]um_direct.
7452 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7453 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7454 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7455 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7457 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7459 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7460 generation for -maltivec=be.
7461 (altivec_vsumsws): Simplify redundant test.
7463 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7465 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7466 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7467 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7468 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7469 gen_altivec_vpkuwum.
7470 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7472 (altivec_vpks<VI_char>ss): Likewise.
7473 (altivec_vpks<VI_char>us): Likewise.
7474 (altivec_vpku<VI_char>us): Likewise.
7475 (altivec_vpku<VI_char>um): Likewise.
7476 (altivec_vpku<VI_char>um_direct): New (copy of
7477 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7479 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7480 target is little endian and -maltivec=be is not specified.
7481 (*altivec_vupkhs<VU_char>_direct): New (copy of
7482 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7483 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7484 target is little endian and -maltivec=be is not specified.
7485 (*altivec_vupkls<VU_char>_direct): New (copy of
7486 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7487 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7488 little endian and -maltivec=be is not specified.
7489 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7490 little endian and -maltivec=be is not specified.
7492 2014-02-05 Richard Henderson <rth@redhat.com>
7495 * combine-stack-adj.c: Revert r206943.
7496 * sched-int.h (struct deps_desc): Add last_args_size.
7497 * sched-deps.c (init_deps): Initialize it.
7498 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7499 contain REG_ARGS_SIZE notes.
7501 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7503 * lto-cgraph.c (asm_nodes_output): Make global.
7504 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7505 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7506 (driver_handle_option): Handle OPT_fwpa.
7508 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7511 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7512 a comment typo and formatting issue. If odr_hash hasn't been
7513 created, return vNULL and set *completep to false.
7516 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7517 bb with no successors.
7519 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7522 * doc/invoke.texi (-march): Clarify documentation for ARM.
7526 2014-02-05 Richard Biener <rguenther@suse.de>
7528 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7529 when not vectorizing because of too many alias checks.
7530 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7531 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7533 2014-02-05 Nick Clifton <nickc@redhat.com>
7535 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7536 accept extended registers in any mode when compiling for the MN10300.
7538 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7540 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7541 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7542 sanitization attributes.
7543 (can_inline_edge_p): Likewise.
7544 (sanitize_attrs_match_for_inline_p): New function.
7546 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7548 * ipa-prop.c (detect_type_change): Shor circuit testing of
7549 type changes on THIS pointer.
7551 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7554 * config/pa/pa.c (legitimize_tls_address): Return original address
7555 if not passed a SYMBOL_REF rtx.
7556 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7558 (pa_emit_move_sequence): Simplify TLS source operands.
7559 (pa_legitimate_constant_p): Reject all TLS constants.
7560 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7561 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7563 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7565 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7566 groups when we know they are controlled by LTO.
7567 * varasm.c (default_binds_local_p_1): If object is in other partition,
7568 it will be resolved locally.
7570 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7572 * config/host-linux.c (linux_gt_pch_use_address): Don't
7573 use SSIZE_MAX because it is not always defined.
7575 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7578 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7579 threshold for pseudo splitting.
7580 (update_ebb_live_info): Process call argument hard registers and
7581 hard registers from insn definition too.
7582 (max_small_class_regs_num): New constant.
7583 (inherit_in_ebb): Update live hard regs through EBBs. Update
7584 reloads_num only for small register classes. Don't split for
7587 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7590 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7593 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7595 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7598 2014-02-04 Marek Polacek <polacek@redhat.com>
7600 * gdbinit.in (pel): Define.
7602 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7604 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7607 2014-02-04 Richard Biener <rguenther@suse.de>
7610 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7611 in function context local.
7612 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7613 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7614 similar to LTO_imported_decl_ref.
7616 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7618 PR tree-optimization/60002
7619 * cgraphclones.c (build_function_decl_skip_args): Clear
7622 PR tree-optimization/60023
7623 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7624 false to gsi_replace.
7625 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7626 has been in some EH region and vec_stmt could throw, add
7627 vec_stmt into the same EH region.
7628 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7629 has no lhs, ignore it.
7630 * internal-fn.c (expand_MASK_LOAD): Likewise.
7633 * tree-inline.c (copy_forbidden): Fail for
7634 __attribute__((optimize (0))) functions.
7637 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7638 is set, copy one less argument.
7639 (expand_simd_clones): Don't subtract clone_info->inbranch
7640 from simd_clone_struct_alloc argument.
7642 PR rtl-optimization/57915
7643 * recog.c (simplify_while_replacing): If all unary/binary/relational
7644 operation arguments are constant, attempt to simplify those.
7647 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7648 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7650 2014-02-04 Richard Biener <rguenther@suse.de>
7652 PR tree-optimization/60012
7653 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7654 TBAA disambiguation to all DDRs.
7656 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7659 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7660 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7662 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7665 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7667 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7669 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7670 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7672 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7675 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7676 to figure out targets of polymorphic calls with known decl.
7677 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7678 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7679 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7680 (get_polymorphic_call_info): ... here.
7681 (get_polymorphic_call_info_from_invariant): New function.
7683 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7685 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7686 lookup via vtable pointer; check for type consistency
7687 and turn inconsitent facts into UNREACHABLE.
7688 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7689 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7690 type inconsistent querries; return UNREACHABLE instead.
7692 2014-02-03 Richard Henderson <rth@twiddle.net>
7695 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7696 already processed this node.
7697 (normalize_one_pred_1): Pass along mark_set.
7698 (normalize_one_pred): Create and destroy a pointer_set_t.
7699 (normalize_one_pred_chain): Likewise.
7701 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7703 PR gcov-profile/58602
7704 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7706 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7709 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7710 -fno-devirtualize; try to devirtualize by the knowledge of
7711 virtual table pointer given by aggregate propagation.
7712 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7713 (ipa_print_node_jump_functions): Dump also offset that
7714 is relevant for polymorphic calls.
7715 (determine_known_aggregate_parts): Add arg_type parameter; use it
7716 instead of determining the type from pointer type.
7717 (ipa_compute_jump_functions_for_edge): Update call of
7718 determine_known_aggregate_parts.
7719 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7720 (gimple_get_virt_method_for_binfo): ... here; simplify using
7721 vtable_pointer_value_to_vtable.
7722 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7723 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7724 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7725 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7727 (vtable_pointer_value_to_binfo): ... here.
7728 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7730 2014-02-03 Teresa Johnson <tejohnson@google.com>
7732 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7733 redef of outer loop index variable.
7735 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7739 * doc/extend.texi (Function Attributes): Typo.
7741 2014-02-03 Cong Hou <congh@google.com>
7743 PR tree-optimization/60000
7744 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7745 if the vectorized statement is a store. A store statement can only
7746 appear at the end of pattern statements.
7748 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7750 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7751 (ix86_option_override_internal): Default long double to 64-bit for
7752 32-bit Bionic and to 128-bit for 64-bit Bionic.
7754 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7755 TARGET_LONG_DOUBLE_128 is true.
7756 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7758 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7759 (mlong-double-64): Negate -mlong-double-128.
7760 (mlong-double-128): New option.
7762 * config/i386/i386-c.c (ix86_target_macros): Define
7763 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7765 * doc/invoke.texi: Document -mlong-double-128.
7767 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7769 PR rtl-optimization/60024
7770 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7772 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7774 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7776 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7778 PR rtl-optimization/57662
7779 * sel-sched.c (code_motion_path_driver): Do not mark already not
7780 existing blocks in the visiting bitmap.
7782 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7784 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7785 on the insn being emitted.
7787 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7788 Will Deacon <will.deacon@arm.com>
7790 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7792 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7794 * config/arm/arm-tables.opt: Regenerate.
7796 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7798 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7799 for vector types other than V16QImode.
7800 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7801 define_expand, and call altivec_expand_vec_perm_le when producing
7802 code with little endian element order.
7803 (*altivec_vperm_<mode>_internal): New insn having previous
7804 behavior of altivec_vperm_<mode>.
7805 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7806 altivec_expand_vec_perm_le when producing code with little endian
7808 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7809 behavior of altivec_vperm_<mode>_uns.
7811 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7813 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7814 (altivec_vsumsws): Add handling for -maltivec=be with a little
7816 (altivec_vsumsws_direct): New.
7817 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7818 gen_altivec_vsumsws.
7820 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7822 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7823 vtable_pointer_value_to_binfo): New functions.
7824 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7825 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7827 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7829 * config/nios2/nios2.md (load_got_register): Initialize GOT
7830 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7831 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7833 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7835 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7836 preserverd by passthrough, do not propagate the type.
7838 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7840 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7841 (mips_atomic_assign_expand_fenv): New function.
7842 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7844 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7846 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7847 (__builtin_mips_set_fcsr): Likewise.
7848 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7849 MIPS_USI_FTYPE_VOID.
7850 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7851 (mips16_expand_set_fcsr): Likewise.
7852 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7853 (mips16_set_fcsr_stub): Likewise.
7854 (mips16_get_fcsr_one_only_stub): New class.
7855 (mips16_set_fcsr_one_only_stub): Likewise.
7856 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7857 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7858 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7859 (hard_float): New availability predicate.
7860 (mips_builtins): Add get_fcsr and set_fcsr.
7861 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7862 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7863 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7864 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7865 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7868 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7870 * config/mips/mips.c (mips_one_only_stub): New class.
7871 (mips_need_mips16_rdhwr_p): Replace with...
7872 (mips16_rdhwr_stub): ...this new variable.
7873 (mips16_stub_call_address): New function.
7874 (mips16_rdhwr_one_only_stub): New class.
7875 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7876 (mips_output_mips16_rdhwr): Delete.
7877 (mips_finish_stub): New function.
7878 (mips_code_end): Use it to handle rdhwr stubs.
7880 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7883 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7884 when calculating size of integer atomic types.
7886 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7888 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7890 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7892 PR tree-optimization/60003
7893 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7894 * profile.c (branch_prob): Use gimple_call_builtin_p
7895 to check for BUILT_IN_SETJMP_RECEIVER.
7896 * tree-inline.c (copy_bb): Call notice_special_calls.
7898 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7901 * lra-constraints.c (process_alt_operands): Update reload_sum only
7904 2014-01-31 Richard Henderson <rth@redhat.com>
7907 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7908 until after else_eh is processed.
7910 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7912 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7913 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7914 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7915 in smmintrin.h, remove them.
7916 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7917 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7918 * config/i386/i386.md (ROUND_SAE): Fix value.
7919 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7920 (const48_operand): New.
7921 * config/i386/subst.md (round), (round_expand): Use
7922 const_4_or_8_to_11_operand.
7923 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7925 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7927 * config/i386/constraints.md (Yk): Swap meaning with k.
7928 * config/i386/i386.md (movhi_internal): Change Yk to k.
7929 (movqi_internal): Ditto.
7930 (*k<logic><mode>): Ditto.
7933 (kandn<mode>): Ditto.
7934 (*<code>hi_1): Ditto.
7935 (*<code>qi_1): Ditto.
7936 (kxnor<mode>): Ditto.
7937 (kortestzhi): Ditto.
7938 (kortestchi): Ditto.
7940 (*one_cmplhi2_1): Ditto.
7941 (*one_cmplqi2_1): Ditto.
7942 * config/i386/sse.md (): Change k to Yk.
7943 (avx512f_load<mode>_mask): Ditto.
7944 (avx512f_blendm<mode>): Ditto.
7945 (avx512f_store<mode>_mask): Ditto.
7946 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7947 (avx512f_storedqu<mode>_mask): Ditto.
7948 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7950 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7951 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7952 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7953 (avx512f_maskcmp<mode>3): Ditto.
7954 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7955 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7956 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7957 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7958 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7959 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7960 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7961 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7962 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7963 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7964 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7965 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7966 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7967 (vec_extract_lo_<mode>_maskm): Ditto.
7968 (vec_extract_hi_<mode>_maskm): Ditto.
7969 (avx512f_vternlog<mode>_mask): Ditto.
7970 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7971 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7972 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7973 (avx512f_<code>v8div16qi2_mask): Ditto.
7974 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7975 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7976 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7977 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7978 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7979 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7980 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7981 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7982 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7983 (avx512cd_maskb_vec_dupv8di): Ditto.
7984 (avx512cd_maskw_vec_dupv16si): Ditto.
7985 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7986 (avx512f_vpermi2var<mode>3_mask): Ditto.
7987 (avx512f_vpermi2var<mode>3_mask): Ditto.
7988 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7989 (*avx512f_gathersi<mode>): Ditto.
7990 (*avx512f_gathersi<mode>_2): Ditto.
7991 (*avx512f_gatherdi<mode>): Ditto.
7992 (*avx512f_gatherdi<mode>_2): Ditto.
7993 (*avx512f_scattersi<mode>): Ditto.
7994 (*avx512f_scatterdi<mode>): Ditto.
7995 (avx512f_compress<mode>_mask): Ditto.
7996 (avx512f_compressstore<mode>_mask): Ditto.
7997 (avx512f_expand<mode>_mask): Ditto.
7998 * config/i386/subst.md (mask): Change k to Yk.
7999 (mask_scalar_merge): Ditto.
8002 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8004 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8006 2014-01-31 Richard Biener <rguenther@suse.de>
8009 * builtins.c (fold_builtin_memory_op): Make sure to not
8010 use a floating-point mode or a boolean or enumeral type for
8013 2014-01-30 DJ Delorie <dj@redhat.com>
8015 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8016 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8017 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8018 whenever main() has an epilogue.
8020 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8022 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8023 unused variable "field".
8024 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8025 (vsx_mergeh_<mode>): Likewise.
8026 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8027 (altivec_vmrghh): Likewise.
8028 (altivec_vmrghw): Likewise.
8029 (altivec_vmrglb): Likewise.
8030 (altivec_vmrglh): Likewise.
8031 (altivec_vmrglw): Likewise.
8032 (altivec_vspltb): Add missing uses.
8033 (altivec_vsplth): Likewise.
8034 (altivec_vspltw): Likewise.
8035 (altivec_vspltsf): Likewise.
8037 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8040 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8041 frame related instructions.
8043 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8045 PR rtl-optimization/59959
8046 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8047 any reload of register whose subreg is invalid.
8049 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8051 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8052 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8053 Add missing return type - void.
8055 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8057 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8058 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8059 remove element index adjustment for endian (now handled in vsx.md
8061 (altivec_expand_vec_perm_const): Use
8062 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8063 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8064 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8065 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8066 define_expand and a new define_insn *altivec_vspltb_internal;
8067 adjust for -maltivec=be on a little endian target.
8068 (altivec_vspltb_direct): New.
8069 (altivec_vsplth): Divide into a define_expand and a new
8070 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8071 little endian target.
8072 (altivec_vsplth_direct): New.
8073 (altivec_vspltw): Divide into a define_expand and a new
8074 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8075 little endian target.
8076 (altivec_vspltw_direct): New.
8077 (altivec_vspltsf): Divide into a define_expand and a new
8078 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8079 a little endian target.
8081 2014-01-30 Richard Biener <rguenther@suse.de>
8083 PR tree-optimization/59993
8084 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8085 can propagate form the earlier stmt and avoid the transform
8086 when the intermediate result is needed.
8088 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8090 * README.Portability: Fix typo.
8092 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8094 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8095 comparison_operator with ordered_comparison_operator.
8097 2014-01-30 Nick Clifton <nickc@redhat.com>
8099 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8100 Rename to mn10300_store_multiple_regs.
8101 * config/mn10300/mn10300.c: Likewise.
8102 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8103 store_multiple_regs.
8104 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8105 Call mn10300_store_multiple_regs.
8107 2014-01-30 Nick Clifton <nickc@redhat.com>
8108 DJ Delorie <dj@redhat.com>
8110 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8111 %fp 2 to keep registers after it properly word-aligned.
8112 (rl78_alloc_physical_registers_umul): Handle the case where both
8113 input operands are the same.
8115 2014-01-30 Richard Biener <rguenther@suse.de>
8117 PR tree-optimization/59903
8118 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8121 2014-01-30 Jason Merrill <jason@redhat.com>
8124 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8127 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8129 2014-01-30 Richard Biener <rguenther@suse.de>
8131 PR tree-optimization/59951
8132 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8134 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8137 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8138 SFmode to DFmode case.
8140 2014-01-29 DJ Delorie <dj@redhat.com>
8142 * config/msp430/msp430.opt (-minrt): New.
8143 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8145 (ENDFILE_SPEC): Likewise.
8147 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8149 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8150 (estimate_function_body_sizes): Use it.
8152 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8155 * dwarf2out.c (is_cxx_auto): New.
8156 (is_base_type): Use it.
8157 (gen_type_die_with_usage): Likewise.
8159 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8161 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8162 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8163 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8164 -maltivec=be with LE targets.
8165 (vsx_mergeh_<mode>): Likewise.
8166 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8167 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8168 (altivec_vmrghb): Replace with define_expand and new
8169 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8170 (altivec_vmrghb_direct): New define_insn.
8171 (altivec_vmrghh): Replace with define_expand and new
8172 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8173 (altivec_vmrghh_direct): New define_insn.
8174 (altivec_vmrghw): Replace with define_expand and new
8175 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8176 (altivec_vmrghw_direct): New define_insn.
8177 (*altivec_vmrghsf): Adjust for endianness.
8178 (altivec_vmrglb): Replace with define_expand and new
8179 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8180 (altivec_vmrglb_direct): New define_insn.
8181 (altivec_vmrglh): Replace with define_expand and new
8182 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8183 (altivec_vmrglh_direct): New define_insn.
8184 (altivec_vmrglw): Replace with define_expand and new
8185 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8186 (altivec_vmrglw_direct): New define_insn.
8187 (*altivec_vmrglsf): Adjust for endianness.
8188 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8189 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8190 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8191 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8192 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8193 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8194 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8195 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8197 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8199 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8200 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8203 2014-01-29 Richard Biener <rguenther@suse.de>
8205 PR tree-optimization/58742
8206 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8207 associate_pointerplus_align.
8208 (associate_pointerplus_diff): New function.
8209 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8210 and associate_pointerplus_diff.
8212 2014-01-29 Richard Biener <rguenther@suse.de>
8214 * lto-streamer.h (LTO_major_version): Bump to 3.
8215 (LTO_minor_version): Reset to 0.
8217 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8219 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8220 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8221 (arm_file_start): Generate correct asm header for armv7ve.
8222 * config/arm/bpabi.h: Add multilib support for armv7ve.
8223 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8224 and cortex-a15 to armv7ve.
8225 * config/arm/t-aprofile: Add multilib support for armv7ve.
8226 * doc/invoke.texi: Document -march=armv7ve.
8228 2014-01-29 Richard Biener <rguenther@suse.de>
8230 PR tree-optimization/58742
8231 * tree-ssa-forwprop.c (associate_plusminus): Return true
8232 if we changed sth, defer EH cleanup to ...
8233 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8234 (simplify_mult): New function.
8236 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8239 PR tree-optimization/59920
8240 * tree.c (build_common_builtin_nodes): Remove
8241 __builtin_setjmp_dispatcher initialization.
8242 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8243 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8244 instead of gsi_after_labels + manually skipping debug stmts.
8245 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8246 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8247 * tree-inline.c (copy_edges_for_bb): Remove
8248 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8249 argument. Ignore computed_goto_p stmts. Don't call
8250 make_abnormal_goto_edges. If a call might need abnormal edges
8251 for non-local gotos, see if it already has an edge to
8252 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8253 with true argument, don't do anything then, otherwise add
8254 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8255 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8257 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8258 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8259 (lower_stmt): Don't set data->calls_builtin_setjmp.
8260 (lower_builtin_setjmp): Adjust comment.
8261 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8262 * tree-cfg.c (found_computed_goto): Remove.
8263 (factor_computed_gotos): Remove.
8264 (make_goto_expr_edges): Return bool, true for computed gotos.
8265 Don't call make_abnormal_goto_edges.
8266 (build_gimple_cfg): Don't set found_computed_goto, don't call
8267 factor_computed_gotos.
8268 (computed_goto_p): No longer static.
8269 (make_blocks): Don't set found_computed_goto.
8270 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8271 (make_edges): If make_goto_expr_edges returns true, push bb
8272 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8273 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8274 vector. Record mapping between bbs and OpenMP regions if there
8275 are any, adjust make_gimple_omp_edges caller. Call
8276 handle_abnormal_edges.
8277 (make_abnormal_goto_edges): Remove.
8278 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8279 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8280 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8281 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8282 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8283 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8284 filling *region also set *region_idx to (*region)->entry->index.
8287 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8288 For REGs set ORIGINAL_REGNO.
8290 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8292 * doc/md.texi: Mention that a target shouldn't implement
8293 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8296 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8298 PR tree-optimization/59594
8299 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8300 a copy of the datarefs vector rather than the vector itself.
8302 2014-01-28 Jason Merrill <jason@redhat.com>
8305 * dwarf2out.c (auto_die): New static.
8306 (gen_type_die_with_usage): Handle C++1y 'auto'.
8307 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8310 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8313 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8314 (SPEC_X32): Likewise.
8315 (SPEC_64): Likewise.
8316 * config/i386/i386.c (ix86_option_override_internal): Turn off
8317 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8319 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8320 * config/i386/i386.h (TARGET_16BIT): New macro.
8321 (TARGET_16BIT_P): Likewise.
8322 * config/i386/i386.opt: Add m16.
8323 * doc/invoke.texi: Document -m16.
8325 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8327 PR preprocessor/59935
8328 * input.c (location_get_source_line): Bail out on when line number
8329 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8331 2014-01-28 Richard Biener <rguenther@suse.de>
8333 PR tree-optimization/58742
8334 * tree-ssa-forwprop.c (associate_plusminus): Handle
8335 pointer subtraction of the form (T)(P + A) - (T)P.
8337 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8339 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8342 2014-01-28 Richard Biener <rguenther@suse.de>
8345 2014-01-28 Richard Biener <rguenther@suse.de>
8347 PR rtl-optimization/45364
8348 PR rtl-optimization/59890
8349 * var-tracking.c (local_get_addr_clear_given_value): Handle
8350 already cleared slot.
8351 (val_reset): Handle not allocated local_get_addr_cache.
8352 (vt_find_locations): Use post-order on the inverted CFG.
8354 2014-01-28 Richard Biener <rguenther@suse.de>
8356 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8358 2014-01-28 Richard Biener <rguenther@suse.de>
8360 PR rtl-optimization/45364
8361 PR rtl-optimization/59890
8362 * var-tracking.c (local_get_addr_clear_given_value): Handle
8363 already cleared slot.
8364 (val_reset): Handle not allocated local_get_addr_cache.
8365 (vt_find_locations): Use post-order on the inverted CFG.
8367 2014-01-28 Alan Modra <amodra@gmail.com>
8369 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8370 * configure.ac <recursive call for build != host>: Define
8371 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8372 and LD_FOR_BUILD too.
8373 * configure: Regenerate.
8375 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8377 * config/i386/i386.c (get_builtin_code_for_version): Separate
8378 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8379 Broadwell from Haswell.
8381 2014-01-27 Steve Ellcey <sellcey@mips.com>
8383 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8384 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8385 * config/mips/mips.c (mips_option_override): Change setting
8387 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8388 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8389 Change from Mask to Var.
8391 2014-01-27 Jeff Law <law@redhat.com>
8393 * ipa-inline.c (inline_small_functions): Fix typo.
8395 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8397 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8398 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8399 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8400 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8401 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8402 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8403 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8404 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8405 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8406 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8407 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8408 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8409 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8410 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8411 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8412 (_mm512_storeu_epi64): Ditto.
8413 (_mm512_cmpge_epi32_mask): Ditto.
8414 (_mm512_cmpge_epu32_mask): Ditto.
8415 (_mm512_cmpge_epi64_mask): Ditto.
8416 (_mm512_cmpge_epu64_mask): Ditto.
8417 (_mm512_cmple_epi32_mask): Ditto.
8418 (_mm512_cmple_epu32_mask): Ditto.
8419 (_mm512_cmple_epi64_mask): Ditto.
8420 (_mm512_cmple_epu64_mask): Ditto.
8421 (_mm512_cmplt_epi32_mask): Ditto.
8422 (_mm512_cmplt_epu32_mask): Ditto.
8423 (_mm512_cmplt_epi64_mask): Ditto.
8424 (_mm512_cmplt_epu64_mask): Ditto.
8425 (_mm512_cmpneq_epi32_mask): Ditto.
8426 (_mm512_cmpneq_epu32_mask): Ditto.
8427 (_mm512_cmpneq_epi64_mask): Ditto.
8428 (_mm512_cmpneq_epu64_mask): Ditto.
8429 (_mm512_expand_pd): Ditto.
8430 (_mm512_expand_ps): Ditto.
8431 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8432 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8433 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8434 * config/i386/i386.c (ix86_builtins): Add
8435 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8436 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8437 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8438 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8439 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8440 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8441 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8442 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8443 IX86_BUILTIN_PMOVUSQW512_MEM.
8444 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8445 __builtin_ia32_pmovsqd512mem_mask,
8446 __builtin_ia32_pmovqd512mem_mask,
8447 __builtin_ia32_pmovusqw512mem_mask,
8448 __builtin_ia32_pmovsqw512mem_mask,
8449 __builtin_ia32_pmovqw512mem_mask,
8450 __builtin_ia32_pmovusdw512mem_mask,
8451 __builtin_ia32_pmovsdw512mem_mask,
8452 __builtin_ia32_pmovdw512mem_mask,
8453 __builtin_ia32_pmovqb512mem_mask,
8454 __builtin_ia32_pmovusqb512mem_mask,
8455 __builtin_ia32_pmovsqb512mem_mask,
8456 __builtin_ia32_pmovusdb512mem_mask,
8457 __builtin_ia32_pmovsdb512mem_mask,
8458 __builtin_ia32_pmovdb512mem_mask.
8459 (bdesc_args): Add __builtin_ia32_expanddf512,
8460 __builtin_ia32_expandsf512.
8461 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8462 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8463 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8464 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8465 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8466 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8467 (avx512f_<code>v8div16qi2_mask_store): This.
8468 (avx512f_expand<mode>): New.
8470 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8472 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8474 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8475 (_mm512_prefetch_i32scatter_pd): Ditto.
8476 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8477 (_mm512_prefetch_i64scatter_pd): Ditto.
8478 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8479 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8480 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8481 (_mm512_prefetch_i32scatter_ps): Ditto.
8482 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8483 (_mm512_prefetch_i64scatter_ps): Ditto.
8484 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8485 * config/i386/i386-builtin-types.def: Define
8486 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8487 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8488 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8489 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8490 IX86_BUILTIN_SCATTERPFQPD.
8491 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8492 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8493 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8494 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8495 __builtin_ia32_scatterpfqps.
8496 (ix86_expand_builtin): Expand new built-ins.
8497 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8498 fix memory access data type.
8499 (*avx512pf_gatherpf<mode>_mask): Ditto.
8500 (*avx512pf_gatherpf<mode>): Ditto.
8501 (avx512pf_scatterpf<mode>): Ditto.
8502 (*avx512pf_scatterpf<mode>_mask): Ditto.
8503 (*avx512pf_scatterpf<mode>): Ditto.
8504 (GATHER_SCATTER_SF_MEM_MODE): New.
8505 (avx512pf_gatherpf<mode>df): Ditto.
8506 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8507 (*avx512pf_scatterpf<mode>df): Ditto.
8509 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8512 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8513 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8516 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8518 * common/config/arm/arm-common.c
8519 (arm_rewrite_mcpu): Handle multiple names.
8521 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8523 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8525 * gimple-builder.h (create_gimple_tmp): Delete.
8527 2014-01-27 Christian Bruel <christian.bruel@st.com>
8529 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8532 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8534 * config/pa/pa.md (call): Generate indirect long calls to non-local
8535 functions when outputing 32-bit code.
8536 (call_value): Likewise except for special call to buggy powf function.
8538 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8539 portable runtime and PIC indirect calls.
8540 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8541 and PIC call sequences. Use ldo instead of blr to set return register
8542 in PIC call sequence.
8544 2014-01-25 Walter Lee <walt@tilera.com>
8546 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8547 avoid clobbering a live register.
8549 2014-01-25 Walter Lee <walt@tilera.com>
8551 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8552 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8553 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8554 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8556 2014-01-25 Walter Lee <walt@tilera.com>
8558 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8559 arguments on even registers.
8560 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8562 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8563 (BIGGEST_ALIGNMENT): Ditto.
8564 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8566 2014-01-25 Walter Lee <walt@tilera.com>
8568 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8569 insns before bundling.
8570 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8572 2014-01-25 Walter Lee <walt@tilera.com>
8574 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8575 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8576 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8578 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8580 * config/mips/constraints.md (kl): Delete.
8581 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8582 define expands, using...
8583 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8584 instructions for MIPS16.
8585 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8586 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8588 2014-01-25 Walter Lee <walt@tilera.com>
8590 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8594 2014-01-25 Walter Lee <walt@tilera.com>
8596 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8597 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8599 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8601 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8604 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8606 * print-rtl.c (in_call_function_usage): New var.
8607 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8608 EXPR_LIST mode as mode and not as reg note name.
8611 * cfgloopmanip.c (copy_loop_info): If
8612 loop->warned_aggressive_loop_optimizations, make sure
8613 the flag is set in target loop too.
8615 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8617 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8619 * builtins.def: Likewise.
8620 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8621 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8622 * ira.c (ira_setup_eliminable_regset): Likewise.
8623 * omp-low.c (gate_expand_omp): Likewise.
8624 (execute_lower_omp): Likewise.
8625 (diagnose_sb_0): Likewise.
8626 (gate_diagnose_omp_blocks): Likewise.
8627 (simd_clone_clauses_extract): Likewise.
8630 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8632 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8633 correction for little endian...
8634 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8637 2014-01-24 Jeff Law <law@redhat.com>
8639 PR tree-optimization/59919
8640 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8641 for non-returning calls.
8643 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8645 * common/config/aarch64/aarch64-common.c
8646 (aarch64_rewrite_mcpu): Handle multiple names.
8647 * config/aarch64/aarch64.h
8648 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8650 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8652 * input.c (add_file_to_cache_tab): Handle the case where fopen
8655 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8658 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8659 from push operand if code of push isn't PRE_DEC.
8661 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8664 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8665 -mquad-memory-atomic. Update -mquad-memory documentation to say
8666 it is only used for non-atomic loads/stores.
8668 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8669 -mquad-memory or -mquad-memory-atomic switches.
8671 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8672 -mquad-memory-atomic to ISA 2.07 support.
8674 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8675 to separate support of normal quad word memory operations (ldq, stq)
8676 from the atomic quad word memory operations.
8678 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8679 support to separate non-atomic quad word operations from atomic
8680 quad word operations. Disable non-atomic quad word operations in
8681 little endian mode so that we don't have to swap words after the
8682 load and before the store.
8683 (quad_load_store_p): Add comment about atomic quad word support.
8684 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8685 options printed with -mdebug=reg.
8687 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8688 -mquad-memory-atomic as the test for whether we have quad word
8689 atomic instructions.
8690 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8691 or -mp8-vector are used, allow byte/half-word atomic operations.
8693 * config/rs6000/sync.md (load_lockedti): Insure that the address
8694 is a proper indexed or indirect address for the lqarx instruction.
8695 On little endian systems, swap the hi/lo registers after the lqarx
8697 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8698 insure the address is valid for the lqarx instruction.
8699 (store_conditionalti): Insure that the address is a proper indexed
8700 or indirect address for the stqcrx. instruction. On little endian
8701 systems, swap the hi/lo registers before doing the stqcrx.
8703 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8704 insure the address is valid for the stqcrx. instruction.
8706 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8707 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8708 type of quad memory support is available.
8710 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8713 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8714 there is a danger of looping.
8716 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8718 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8719 force flag_ira_loop_pressure if set via command line.
8721 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8723 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8724 (ashr_simd): New builtin handling DI mode.
8725 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8726 (aarch64_sshr_simddi): New match pattern.
8727 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8728 (vshrd_n_s64): Likewise.
8729 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8731 2014-01-23 Nick Clifton <nickc@redhat.com>
8733 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8734 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8735 favour of mcu specific scripts.
8736 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8739 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8740 Alex Velenko <Alex.Velenko@arm.com>
8742 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8743 (vaddv_s16): Likewise.
8744 (vaddv_s32): Likewise.
8745 (vaddv_u8): Likewise.
8746 (vaddv_u16): Likewise.
8747 (vaddv_u32): Likewise.
8748 (vaddvq_s8): Likewise.
8749 (vaddvq_s16): Likewise.
8750 (vaddvq_s32): Likewise.
8751 (vaddvq_s64): Likewise.
8752 (vaddvq_u8): Likewise.
8753 (vaddvq_u16): Likewise.
8754 (vaddvq_u32): Likewise.
8755 (vaddvq_u64): Likewise.
8756 (vaddv_f32): Likewise.
8757 (vaddvq_f32): Likewise.
8758 (vaddvq_f64): Likewise.
8759 (vmaxv_f32): Likewise.
8760 (vmaxv_s8): Likewise.
8761 (vmaxv_s16): Likewise.
8762 (vmaxv_s32): Likewise.
8763 (vmaxv_u8): Likewise.
8764 (vmaxv_u16): Likewise.
8765 (vmaxv_u32): Likewise.
8766 (vmaxvq_f32): Likewise.
8767 (vmaxvq_f64): Likewise.
8768 (vmaxvq_s8): Likewise.
8769 (vmaxvq_s16): Likewise.
8770 (vmaxvq_s32): Likewise.
8771 (vmaxvq_u8): Likewise.
8772 (vmaxvq_u16): Likewise.
8773 (vmaxvq_u32): Likewise.
8774 (vmaxnmv_f32): Likewise.
8775 (vmaxnmvq_f32): Likewise.
8776 (vmaxnmvq_f64): Likewise.
8777 (vminv_f32): Likewise.
8778 (vminv_s8): Likewise.
8779 (vminv_s16): Likewise.
8780 (vminv_s32): Likewise.
8781 (vminv_u8): Likewise.
8782 (vminv_u16): Likewise.
8783 (vminv_u32): Likewise.
8784 (vminvq_f32): Likewise.
8785 (vminvq_f64): Likewise.
8786 (vminvq_s8): Likewise.
8787 (vminvq_s16): Likewise.
8788 (vminvq_s32): Likewise.
8789 (vminvq_u8): Likewise.
8790 (vminvq_u16): Likewise.
8791 (vminvq_u32): Likewise.
8792 (vminnmv_f32): Likewise.
8793 (vminnmvq_f32): Likewise.
8794 (vminnmvq_f64): Likewise.
8796 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8798 * config/aarch64/aarch64-simd.md
8799 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8800 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8801 (*aarch64_mul3_elt<mode>): Likewise.
8802 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8803 (*aarch64_mul3_elt_to_64v2df): Likewise.
8804 (*aarch64_mla_elt<mode>): Likewise.
8805 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8806 (*aarch64_mls_elt<mode>): Likewise.
8807 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8808 (*aarch64_fma4_elt<mode>): Likewise.
8809 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8810 (*aarch64_fma4_elt_to_64v2df): Likewise.
8811 (*aarch64_fnma4_elt<mode>): Likewise.
8812 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8813 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8814 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8815 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8816 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8817 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8818 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8819 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8820 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8822 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8824 * config/aarch64/aarch64-simd.md
8825 (aarch64_be_checked_get_lane<mode>): New define_expand.
8826 * config/aarch64/aarch64-simd-builtins.def
8827 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8828 New builtin definition.
8829 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8830 Use new safe be builtin.
8832 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8834 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8836 (aarch64_be_st1<mode>): Likewise.
8837 (aarch_ld1<VALL:mode>): Define_expand modified.
8838 (aarch_st1<VALL:mode>): Likewise.
8839 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8840 (UNSPEC_ST1): Likewise.
8842 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8844 * config/microblaze/microblaze.md: Add trap insn and attribute
8846 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8848 PR preprocessor/58580
8849 * input.h (location_get_source_line): Take an additional line_size
8851 (void diagnostics_file_cache_fini): Declare new function.
8852 * input.c (struct fcache): New type.
8853 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8854 New static constants.
8855 (diagnostic_file_cache_init, total_lines_num)
8856 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8857 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8858 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8859 (get_next_line, read_next_line, goto_next_line, read_line_num):
8860 New static function definitions.
8861 (diagnostic_file_cache_fini): New function.
8862 (location_get_source_line): Take an additional output line_len
8863 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8865 * diagnostic.c (diagnostic_finish): Call
8866 diagnostic_file_cache_fini.
8867 (adjust_line): Take an additional input parameter for the length
8868 of the line, rather than calculating it with strlen.
8869 (diagnostic_show_locus): Adjust the use of
8870 location_get_source_line and adjust_line with respect to their new
8871 signature. While displaying a line now, do not stop at the first
8872 null byte. Rather, display the zero byte as a space and keep
8873 going until we reach the size of the line.
8874 * Makefile.in: Add vec.o to OBJS-libcommon
8876 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8877 Ilya Tocar <ilya.tocar@intel.com>
8879 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8880 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8881 (__builtin_ia32_kmov16): Ditto.
8882 * config/i386/i386.md (UNSPEC_KMOV): New.
8885 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8887 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8888 (_mm512_storeu_si512): Ditto.
8890 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8893 * rtl.h (get_referenced_operands): Declare.
8894 * recog.c (get_referenced_operands): New function.
8895 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8896 operands have been referenced when recording LO_SUM references.
8898 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8900 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8902 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8904 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8905 Enable for generic and recent AMD targets.
8907 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8909 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8910 ARG_SIZE note when adjustment was eliminated.
8912 2014-01-22 Jeff Law <law@redhat.com>
8914 PR tree-optimization/59597
8915 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8916 in file. Accept new argument REGISTERING and use it to modify
8917 dump output appropriately.
8918 (register_jump_thread): Corresponding changes.
8919 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8920 thread paths involving joiner blocks. Add code to dump cancelled
8921 jump threading paths.
8923 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8925 PR rtl-optimization/59477
8926 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8927 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8929 2014-01-22 Tom Tromey <tromey@redhat.com>
8931 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8933 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8935 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8937 PR rtl-optimization/59896
8938 * lra-constraints.c (process_alt_operands): Check unused note for
8939 matched operands of insn with no output reloads.
8941 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8943 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8944 (mips_move_from_gpr_cost): Likewise.
8946 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8948 PR rtl-optimization/59858
8949 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8950 ira_class_hard_regs_num.
8951 (process_alt_operands): Increase reject for dying matched operand.
8953 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8956 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8957 smaller than size, perform several stores or loads and stores
8958 at dst + count - size to store or copy all of size bytes, rather
8959 than just last modesize bytes.
8961 2014-01-20 DJ Delorie <dj@redhat.com>
8963 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8964 that CLOBBERs are REGs before propogating their values.
8966 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8969 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8970 (cgraph_inline_failed_type): New function.
8971 * cgraph.h (DEFCIFCODE): Add type.
8972 (cgraph_inline_failed_type_t): New enum.
8973 (cgraph_inline_failed_type): New prototype.
8974 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8975 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8976 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8977 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8978 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8979 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8980 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8981 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8982 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8983 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8984 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8985 OPTIMIZATION_MISMATCH.
8986 * tree-inline.c (expand_call_inline): Emit errors during
8987 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8989 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8992 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8993 mode attribute in insn output.
8995 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8997 * output.h (output_constant): Delete.
8998 * varasm.c (output_constant): Make private.
9000 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9002 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9004 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9007 * tree.h (fold_builtin_strcat): New prototype.
9008 * builtins.c (fold_builtin_strcat): No longer static. Add len
9009 argument, if non-NULL, don't call c_strlen. Optimize
9010 directly into __builtin_memcpy instead of __builtin_strcpy.
9011 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9012 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9014 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9016 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9017 for SImode_address_operand operands, having only a REG argument.
9019 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9021 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9022 loader name using mbig-endian.
9023 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9025 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9027 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9031 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9033 * config/aarch64/aarch64-protos.h
9034 (aarch64_cannot_change_mode_class_ptr): Declare.
9035 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9036 aarch64_cannot_change_mode_class_ptr): New.
9037 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9038 backend hook aarch64_cannot_change_mode_class.
9040 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9042 * common/config/aarch64/aarch64-common.c
9043 (aarch64_handle_option): Don't handle any option order logic here.
9044 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9045 selected_cpu, warn on architecture version mismatch.
9046 (aarch64_override_options): Fix parsing order for option strings.
9048 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9049 Iain Sandoe <iain@codesourcery.com>
9052 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9053 warning. Amend comment to reflect current functionality.
9055 2014-01-20 Richard Biener <rguenther@suse.de>
9058 * builtins.c (fold_builtin_strcat): Remove case better handled
9059 by tree-ssa-strlen.c.
9061 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9063 * config/aarch64/aarch64.opt
9064 (mcpu, march, mtune): Make case-insensitive.
9066 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9069 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9070 if operands[1] is a REG or ZERO_EXTEND of a REG.
9072 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9074 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9076 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9078 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9079 long non-pic millicode calls.
9081 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9083 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9085 2014-01-19 Kito Cheng <kito@0xlab.org>
9087 * builtins.c (expand_movstr): Check movstr expand done or fail.
9089 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9090 H.J. Lu <hongjiu.lu@intel.com>
9093 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9094 to DImode for zero-extended addresses.
9096 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9098 PR rtl-optimization/57763
9099 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9100 on the new indirect jump_insn and increment LABEL_NUSES (label).
9102 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9106 * config.gcc (x86_archs): New variable.
9107 (x86_64_archs): Likewise.
9108 (x86_cpus): Likewise.
9109 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9110 --with-arch/--with-cpu= options.
9111 Support --with-arch=/--with-cpu={nehalem,westmere,
9112 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9114 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9116 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9117 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9119 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9121 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9123 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9126 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9127 clear cpp_get_options (parse_in)->warn_unused_macros for
9128 ix86_target_macros_internal with cpp_define.
9130 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9132 * jump.c (delete_related_insns): Keep (use (insn))s.
9133 * reorg.c (redundant_insn): Check for barriers too.
9135 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9137 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9139 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9141 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9142 call to $$dyncall when TARGET_LONG_CALLS is true.
9144 2014-01-17 Jeff Law <law@redhat.com>
9146 * ree.c (combine_set_extension): Temporarily disable test for
9147 changing number of hard registers.
9149 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9152 * ipa-inline-analysis.c (inline_free_summary):
9153 Do not free summary of aliases.
9155 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9158 * gimplify.c (gimplify_expr): Use create_tmp_var
9159 instead of create_tmp_var_raw. If cond doesn't have
9160 integral type, don't add the IFN_ANNOTATE builtin at all.
9162 2014-01-17 Martin Jambor <mjambor@suse.cz>
9165 * ipa-cp.c (prev_edge_clone): New variable.
9166 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9167 Also resize prev_edge_clone vector.
9168 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9169 (ipcp_edge_removal_hook): New function.
9170 (ipcp_driver): Register ipcp_edge_removal_hook.
9172 2014-01-17 Andrew Pinski <apinski@cavium.com>
9173 Steve Ellcey <sellcey@mips.com>
9176 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9179 2014-01-17 Jeff Law <law@redhat.com>
9182 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9183 so that pass_ccp runs first.
9185 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9187 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9188 (ix86_adjust_cost): Use !TARGET_XXX.
9189 (do_reorder_for_imul): Likewise.
9190 (swap_top_of_ready_list): Likewise.
9191 (ix86_sched_reorder): Likewise.
9193 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9195 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9196 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9197 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9198 (intel_memset): New. Duplicate slm_memset.
9199 (intel_cost): New. Duplicate slm_cost.
9200 (m_INTEL): New macro.
9201 (processor_target_table): Add "intel".
9202 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9203 with PROCESSOR_INTEL for "intel".
9204 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9205 PROCESSOR_SILVERMONT.
9206 (ix86_issue_rate): Likewise.
9207 (ix86_adjust_cost): Likewise.
9208 (ia32_multipass_dfa_lookahead): Likewise.
9209 (swap_top_of_ready_list): Likewise.
9210 (ix86_sched_reorder): Likewise.
9211 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9212 instead of TARGET_OPT_AGU.
9213 * config/i386/i386.h (TARGET_INTEL): New.
9214 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9215 (processor_type): Add PROCESSOR_INTEL.
9216 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9217 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9219 2014-01-17 Marek Polacek <polacek@redhat.com>
9222 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9225 2014-01-17 Richard Biener <rguenther@suse.de>
9227 PR tree-optimization/46590
9228 * opts.c (default_options_table): Add entries for
9229 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9230 all enabled at -O1 but not for -Og.
9231 * common.opt (fbranch-count-reg): Remove Init(1).
9232 (fmove-loop-invariants): Likewise.
9233 (ftree-pta): Likewise.
9235 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9237 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9238 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9239 decls to at least the GCC 4.8 used alignments.
9242 * tree-nested.c (convert_nonlocal_reference_stmt,
9243 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9244 of GIMPLE_BIND stmts, adjust associated decls.
9246 2014-01-17 Richard Biener <rguenther@suse.de>
9248 PR tree-optimization/46590
9249 * vec.h (vec<>::bseach): New member function implementing
9250 binary search according to C89 bsearch.
9251 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9252 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9253 bitmap pointer again. Make accesses_in_loop a flat array.
9254 (mem_ref_obstack): New global.
9255 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9256 (mark_ref_stored): Likewise.
9257 (ref_indep_loop_p_2): Likewise.
9258 (set_ref_stored_in_loop): New helper function.
9259 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9260 (memref_free): Adjust.
9261 (record_mem_ref_loc): Simplify.
9262 (gather_mem_refs_stmt): Adjust.
9263 (sort_locs_in_loop_postorder_cmp): New function.
9264 (analyze_memory_references): Sort accesses_in_loop after
9265 loop postorder number.
9266 (find_ref_loc_in_loop_cmp): New function.
9267 (for_all_locs_in_loop): Find relevant cluster of locs in
9268 accesses_in_loop and iterate without recursion.
9269 (execute_sm): Avoid uninit warning.
9270 (struct ref_always_accessed): Simplify.
9271 (ref_always_accessed::operator ()): Likewise.
9272 (ref_always_accessed_p): Likewise.
9273 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9274 loop postorder numbers here.
9275 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9278 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9281 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9282 on decls for which assemble_alias has been called.
9284 2014-01-17 Nick Clifton <nickc@redhat.com>
9286 * config/msp430/msp430.opt: (mcpu): New option.
9287 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9288 (msp430_option_override): Parse target_cpu. If the MCU name
9289 matches a generic string, clear target_mcu.
9290 (msp430_attr): Allow numeric interrupt values up to 63.
9291 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9292 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9294 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9296 * config/msp430/msp430.md (popm): Use %J rather than %I.
9297 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9298 (addhi_cy_i): Use immediate_operand for operand 2.
9299 * doc/invoke.texi: Document -mcpu option.
9301 2014-01-17 Richard Biener <rguenther@suse.de>
9303 PR rtl-optimization/38518
9304 * df.h (df_analyze_loop): Declare.
9305 * df-core.c: Include cfgloop.h.
9306 (df_analyze_1): Split out main part of df_analyze.
9307 (df_analyze): Adjust.
9308 (loop_inverted_post_order_compute): New function.
9309 (loop_post_order_compute): Likewise.
9310 (df_analyze_loop): New function avoiding whole-function
9312 * loop-invariant.c (find_defs): Use df_analyze_loop.
9313 (find_invariants): Adjust.
9314 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9316 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9318 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9319 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9321 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9323 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9324 traversal when removing references.
9326 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9329 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9331 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9334 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9335 pushing a reload for an autoinc when we had previously reloaded an
9336 inner part of the address.
9338 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9340 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9342 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9343 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9344 when not giving up or versioning for alias only because of
9346 (vect_analyze_data_ref_dependences): Set to true.
9347 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9349 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9350 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9354 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9357 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9358 operand 0 predicate for gathers, use a new pseudo as subtarget.
9360 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9363 * lra-constraints.c (process_alt_operands): Add printing debug info.
9364 Check absence of input/output reloads for matched operands too.
9366 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9368 PR rtl-optimization/59835
9369 * ira.c (ira_init_register_move_cost): Increase cost for
9372 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9374 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9376 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9379 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9380 non-register objects. Use gen_(high/low)part more consistently.
9383 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9386 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9387 endian support, remove tests for WORDS_BIG_ENDIAN.
9388 (p8_mfvsrd_3_<mode>): Likewise.
9389 (reload_gpr_from_vsx<mode>): Likewise.
9390 (reload_gpr_from_vsxsf): Likewise.
9391 (p8_mfvsrd_4_disf): Likewise.
9393 2014-01-16 Richard Biener <rguenther@suse.de>
9395 PR rtl-optimization/46590
9396 * lcm.c (compute_antinout_edge): Use postorder iteration.
9397 (compute_laterin): Use inverted postorder iteration.
9399 2014-01-16 Nick Clifton <nickc@redhat.com>
9402 * varasm.c (output_constant): Return the number of bytes actually
9404 (output_constructor_array_range): Update the field size with the
9405 number of bytes emitted by output_constant.
9406 (output_constructor_regular_field): Likewise. Also do not
9407 complain if the total number of bytes emitted is now greater
9408 than the expected fieldpos.
9409 * output.h (output_constant): Update prototype and descriptive comment.
9411 2014-01-16 Marek Polacek <polacek@redhat.com>
9414 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9415 it is error_mark_node.
9417 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9419 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9420 VALID_AVX256_REG_OR_OI_MODE.
9422 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9424 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9425 current procedure should be profiled.
9427 2014-01-15 Andrew Pinski <apinski@cavium.com>
9429 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9430 of moving from/to the STACK_REG register class.
9432 2014-01-15 Richard Henderson <rth@redhat.com>
9435 * reginfo.c (global_regs_decl): Globalize.
9436 * rtl.h (global_regs_decl): Declare.
9437 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9438 reserved via global_regs.
9440 2014-01-15 Teresa Johnson <tejohnson@google.com>
9442 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9444 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9446 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9447 and vmulosh rather than call gen_vec_widen_smult_*.
9448 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9449 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9450 (vec_widen_smult_even_v16qi): Likewise.
9451 (vec_widen_umult_even_v8hi): Likewise.
9452 (vec_widen_smult_even_v8hi): Likewise.
9453 (vec_widen_umult_odd_v16qi): Likewise.
9454 (vec_widen_smult_odd_v16qi): Likewise.
9455 (vec_widen_umult_odd_v8hi): Likewise.
9456 (vec_widen_smult_odd_v8hi): Likewise.
9457 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9458 vmuloub rather than call gen_vec_widen_umult_*.
9459 (vec_widen_umult_lo_v16qi): Likewise.
9460 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9461 vmulosb rather than call gen_vec_widen_smult_*.
9462 (vec_widen_smult_lo_v16qi): Likewise.
9463 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9464 rather than call gen_vec_widen_umult_*.
9465 (vec_widen_umult_lo_v8hi): Likewise.
9466 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9467 rather than call gen_vec_widen_smult_*.
9468 (vec_widen_smult_lo_v8hi): Likewise.
9470 2014-01-15 Jeff Law <law@redhat.com>
9472 PR tree-optimization/59747
9473 * ree.c (find_and_remove_re): Properly handle case where a second
9474 eliminated extension requires widening a copy created for elimination
9475 of a prior extension.
9476 (combine_set_extension): Ensure that the number of hard regs needed
9477 for a destination register does not change when we widen it.
9479 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9481 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9482 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9483 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9484 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9485 (avr-*-rtems*): Likewise.
9486 (bfin*-rtems*): Likewise.
9487 (moxie-*-rtems*): Likewise.
9488 (h8300-*-rtems*): Likewise.
9489 (i[34567]86-*-rtems*): Likewise.
9490 (lm32-*-rtems*): Likewise.
9491 (m32r-*-rtems*): Likewise.
9492 (m68k-*-rtems*): Likewise.
9493 (microblaze*-*-rtems*): Likewise.
9494 (mips*-*-rtems*): Likewise.
9495 (powerpc-*-rtems*): Likewise.
9496 (sh-*-rtems*): Likewise.
9497 (sparc-*-rtems*): Likewise.
9498 (sparc64-*-rtems*): Likewise.
9499 (v850-*-rtems*): Likewise.
9500 (m32c-*-rtems*): Likewise.
9502 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9504 PR rtl-optimization/59511
9505 * ira.c (ira_init_register_move_cost): Use memory costs for some
9506 cases of register move cost calculations.
9507 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9508 instead of BB frequency.
9509 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9510 * lra-assigns.c (find_hard_regno_for): Ditto.
9512 2014-01-15 Richard Biener <rguenther@suse.de>
9514 PR tree-optimization/59822
9515 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9516 (vectorizable_load): Use it to hoist defs of uses of invariant
9517 loads out of the loop.
9519 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9520 Kugan Vivekanandarajah <kuganv@linaro.org>
9523 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9526 2014-01-15 Richard Biener <rguenther@suse.de>
9528 PR rtl-optimization/59802
9529 * lcm.c (compute_available): Use inverted postorder to seed
9530 the initial worklist.
9532 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9535 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9536 ADDR_REGS for invalid symrefs in non-PIC code.
9538 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9541 * builtins.c (determine_block_size): Initialize *probable_max_size
9542 even if len_rtx is CONST_INT.
9544 2014-01-14 Andrew Pinski <apinski@cavium.com>
9546 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9547 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9548 (cortexa53_tunings): Likewise.
9549 (aarch64_sched_issue_rate): New function.
9550 (TARGET_SCHED_ISSUE_RATE): Define.
9552 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9554 * ira-costs.c (find_costs_and_classes): Add missed
9555 ira_init_register_move_cost_if_necessary.
9557 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9560 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9562 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9565 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9566 to indicate if type is used for function return value. Warn ABI
9567 change if the vector mode isn't available for function return value.
9568 (ix86_function_arg_advance): Pass false to type_natural_mode.
9569 (ix86_function_arg): Likewise.
9570 (ix86_gimplify_va_arg): Likewise.
9571 (function_arg_32): Don't warn ABI change.
9572 (ix86_function_value): Pass true to type_natural_mode.
9573 (ix86_return_in_memory): Likewise.
9574 (ix86_struct_value_rtx): Removed.
9575 (TARGET_STRUCT_VALUE_RTX): Likewise.
9577 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9579 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9580 converting a conditional jump into a conditional return.
9582 2014-01-14 Richard Biener <rguenther@suse.de>
9584 PR tree-optimization/58921
9585 PR tree-optimization/59006
9586 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9587 hoisting invariant stmts.
9588 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9589 invariant loads on the preheader edge if possible.
9591 2014-01-14 Joey Ye <joey.ye@arm.com>
9593 * doc/plugin.texi (Building GCC plugins): Update to C++.
9595 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9597 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9598 (_mm_rcp28_round_ss): Ditto.
9599 (_mm_rsqrt28_round_sd): Ditto.
9600 (_mm_rsqrt28_round_ss): Ditto.
9601 (_mm_rcp28_sd): Ditto.
9602 (_mm_rcp28_ss): Ditto.
9603 (_mm_rsqrt28_sd): Ditto.
9604 (_mm_rsqrt28_ss): Ditto.
9605 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9606 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9607 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9608 (IX86_BUILTIN_RCP28SD): Ditto.
9609 (IX86_BUILTIN_RCP28SS): Ditto.
9610 (IX86_BUILTIN_RSQRT28SD): Ditto.
9611 (IX86_BUILTIN_RSQRT28SS): Ditto.
9612 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9613 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9614 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9615 (ix86_expand_special_args_builtin): Expand new FTYPE.
9616 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9617 (srcp14<mode>): Make insn unary.
9618 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9619 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9620 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9621 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9622 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9623 Fix rounding: make it SAE only.
9624 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9626 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9628 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9629 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9630 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9631 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9632 (round_saeonly_mask_scalar_operand4): Ditto.
9633 (round_saeonly_mask_scalar_op3): Ditto.
9634 (round_saeonly_mask_scalar_op4): Ditto.
9636 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9638 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9639 Implement -maltivec=be for vec_insert and vec_extract.
9641 2014-01-10 DJ Delorie <dj@redhat.com>
9643 * config/msp430/msp430.md (call_internal): Don't allow memory
9644 references with SP as the base register.
9645 (call_value_internal): Likewise.
9646 * config/msp430/constraints.md (Yc): New. For memory references
9647 that don't use SP as a base register.
9649 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9650 "an integer without a # prefix"
9651 * config/msp430/msp430.md (epilogue_helper): Use it.
9653 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9656 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9657 AVX512F gather builtins.
9658 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9659 on gather decls with INTEGER_TYPE masktype.
9660 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9661 directly into the builtin rather than hoisting it before loop.
9663 PR tree-optimization/59387
9664 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9665 (scev_const_prop): If folded_casts and type has undefined overflow,
9666 use force_gimple_operand instead of force_gimple_operand_gsi and
9667 for each added stmt if it is assign with
9668 arith_code_with_undefined_signed_overflow, call
9669 rewrite_to_defined_overflow.
9670 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9671 gimple-fold.h instead.
9672 (arith_code_with_undefined_signed_overflow,
9673 rewrite_to_defined_overflow): Moved to ...
9674 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9675 rewrite_to_defined_overflow): ... here. No longer static.
9676 Include gimplify-me.h.
9677 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9678 rewrite_to_defined_overflow): New prototypes.
9680 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9682 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9684 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9686 * builtins.c (get_object_alignment_2): Minor tweak.
9687 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9689 2014-01-13 Christian Bruel <christian.bruel@st.com>
9691 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9692 optimized non constant lengths.
9694 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9697 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9698 load as __atomic_load_N if possible.
9700 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9702 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9704 (rs6000_expand_builtin): Adjust call.
9706 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9709 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9710 * config/rs6000/rs6000.c: Include target-globals.h.
9711 (rs6000_set_current_function): Instead of doing target_reinit
9712 unconditionally, use save_target_globals_default_opts and
9713 restore_target_globals.
9715 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9717 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9718 (rs6000_expand_builtin): Handle mffs and mtfsf.
9719 (rs6000_init_builtins): Define mffs and mtfsf.
9720 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9721 (rs6000_mffs): New pattern.
9722 (rs6000_mtfsf): New pattern.
9724 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9726 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9727 Start narrowing with START. Apply candidate-use pair
9728 and check overall cost in narrowing.
9729 (iv_ca_prune): Pass new argument.
9731 2014-01-10 Jeff Law <law@redhat.com>
9734 * ree.c (combine_reaching_defs): Ensure the defining statement
9735 occurs before the extension when optimizing extensions with
9736 different source and destination hard registers.
9738 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9741 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9742 vtables into the type inheritance graph.
9744 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9746 PR rtl-optimization/59754
9747 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9748 modes in the REGNO != REGNO case.
9750 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9752 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9754 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9756 PR tree-optimization/59745
9757 * tree-predcom.c (tree_predictive_commoning_loop): Call
9758 free_affine_expand_cache if giving up because components is NULL.
9760 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9761 GC in payload of target_globals struct instead of allocating them on
9762 the heap and the larger structs separately using GC.
9763 * target-globals.h (struct target_globals): Make regs, hard_regs,
9764 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9765 of GTY((skip)) and change type to void *.
9766 (reset_target_globals): Cast loads from those fields to corresponding
9769 2014-01-10 Steve Ellcey <sellcey@mips.com>
9772 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9773 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9774 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9776 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9779 * aarch64-modes.def (CC_Zmode): New flags mode.
9780 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9781 represents an equality.
9782 (aarch64_get_condition_code): Handle CC_Zmode.
9783 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9785 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9787 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9788 extraction in good case.
9790 2014-01-10 Richard Biener <rguenther@suse.de>
9792 PR tree-optimization/59374
9793 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9794 checking after SLP discovery. Mark stmts not participating
9795 in any SLP instance properly.
9797 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9799 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9800 when handling a SET rtx.
9802 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9804 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9805 (cortex-a57): Likewise.
9806 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9808 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9810 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9811 non-iwmmxt builtins.
9813 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9817 * ipa-devirt.c record_target_from_binfo): Take as argument
9818 stack of binfos and lookup matching one for virtual inheritance.
9819 (possible_polymorphic_call_targets_1): Update.
9821 2014-01-10 Huacai Chen <chenhc@lemote.com>
9823 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9824 kernel strings for Loongson-2E/2F/3A.
9826 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9829 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9830 is_gimple_call before calling gimple_call_internal_p.
9832 2014-01-09 Steve Ellcey <sellcey@mips.com>
9834 * Makefile.in (TREE_FLOW_H): Remove.
9835 (TREE_SSA_H): Add file names from tree-flow.h.
9836 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9837 * tree.h: Remove tree-flow.h reference.
9838 * hash-table.h: Remove tree-flow.h reference.
9839 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9840 reference with tree-ssa-loop.h.
9842 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9844 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9845 default element-order behavior for -maltivec.
9846 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9847 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9848 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9849 when targeting big endian, at least for now.
9850 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9852 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9855 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9856 var satisfies use_register_for_decl, just take into account type
9857 alignment, rather than decl alignment.
9859 PR tree-optimization/59622
9860 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9861 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9862 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9863 Don't devirtualize for inplace at all. For targets.length () == 1,
9864 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9866 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9868 * config/i386/i386.md (cpu): Remove the unused btver1.
9870 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9872 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9874 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9877 * tree-core.h (struct target_globals): New forward declaration.
9878 (struct tree_target_option): Add globals field.
9879 * tree.h (TREE_TARGET_GLOBALS): Define.
9880 (prepare_target_option_nodes_for_pch): New prototype.
9881 * target-globals.h (struct target_globals): Define even if
9883 * tree.c (prepare_target_option_node_for_pch,
9884 prepare_target_option_nodes_for_pch): New functions.
9885 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9886 * config/i386/i386.c: Include target-globals.h.
9887 (ix86_set_current_function): Instead of doing target_reinit
9888 unconditionally, use save_target_globals_default_opts and
9889 restore_target_globals.
9891 2014-01-09 Richard Biener <rguenther@suse.de>
9893 PR tree-optimization/59715
9894 * tree-cfg.h (split_critical_edges): Declare.
9895 * tree-cfg.c (split_critical_edges): Export.
9896 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9898 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9900 * cfgexpand.c (expand_stack_vars): Optionally disable
9901 asan stack protection.
9902 (expand_used_vars): Likewise.
9903 (partition_stack_vars): Likewise.
9904 * asan.c (asan_emit_stack_protection): Optionally disable
9905 after return stack usage.
9906 (instrument_derefs): Optionally disable memory access instrumentation.
9907 (instrument_builtin_call): Likewise.
9908 (instrument_strlen_call): Likewise.
9909 (asan_protect_global): Optionally disable global variables protection.
9910 * doc/invoke.texi: Added doc for new options.
9911 * params.def: Added new options.
9912 * params.h: Likewise.
9914 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9916 PR rtl-optimization/59724
9917 * ifcvt.c (cond_exec_process_if_block): Don't call
9918 flow_find_head_matching_sequence with 0 longest_match.
9919 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9920 non-active insns if !stop_after.
9921 (try_head_merge_bb): Revert 2014-01-07 changes.
9923 2014-01-08 Jeff Law <law@redhat.com>
9925 * ree.c (get_sub_rtx): New function, extracted from...
9926 (merge_def_and_ext): Here.
9927 (combine_reaching_defs): Use get_sub_rtx.
9929 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9931 * cgraph.h (varpool_variable_node): Do not choke on null node.
9933 2014-01-08 Catherine Moore <clm@codesourcery.com>
9935 * config/mips/mips.md (simple_return): Attempt to use JRC
9937 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9939 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9941 PR rtl-optimization/59137
9942 * reorg.c (steal_delay_list_from_target): Call update_block for
9944 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9946 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9948 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9949 two duplicate entries.
9951 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9954 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9956 * config/mips/mips.c (mips_truncated_op_cost): New function.
9957 (mips_rtx_costs): Adjust test for BADDU.
9958 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9960 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9962 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9963 (*baddu_si): ...this new pattern.
9965 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9968 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9970 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9973 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9975 (expand_expr, expand_normal): Adjust.
9976 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9977 inner_reference_p. Use inner_reference_p to expand inner references.
9978 (store_expr): Adjust.
9979 * cfgexpand.c (expand_call_stmt): Adjust.
9981 2014-01-08 Rong Xu <xur@google.com>
9983 * gcov-io.c (gcov_var): Move from gcov-io.h.
9984 (gcov_position): Ditto.
9985 (gcov_is_error): Ditto.
9986 (gcov_rewrite): Ditto.
9987 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9988 only part to libgcc/libgcov.h.
9990 2014-01-08 Marek Polacek <polacek@redhat.com>
9993 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9995 2014-01-08 Marek Polacek <polacek@redhat.com>
9998 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10000 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10002 PR rtl-optimization/59649
10003 * stor-layout.c (get_mode_bounds): For BImode return
10004 0 and STORE_FLAG_VALUE.
10006 2014-01-08 Richard Biener <rguenther@suse.de>
10008 PR middle-end/59630
10009 * gimple.h (is_gimple_builtin_call): Remove.
10010 (gimple_builtin_call_types_compatible_p): New.
10011 (gimple_call_builtin_p): New overload.
10012 * gimple.c (is_gimple_builtin_call): Remove.
10013 (validate_call): Rename to ...
10014 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10015 check return types.
10016 (validate_type): New static function.
10017 (gimple_call_builtin_p): New overload and adjust.
10018 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10019 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10020 (gimple_fold_stmt_to_constant_1): Likewise.
10021 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10023 2014-01-08 Richard Biener <rguenther@suse.de>
10025 PR middle-end/59471
10026 * gimplify.c (gimplify_expr): Gimplify register-register type
10027 VIEW_CONVERT_EXPRs to separate stmts.
10029 2014-01-07 Jeff Law <law@redhat.com>
10031 PR middle-end/53623
10032 * ree.c (combine_set_extension): Handle case where source
10033 and destination registers in an extension insn are different.
10034 (combine_reaching_defs): Allow source and destination registers
10035 in extension to be different under limited circumstances.
10036 (add_removable_extension): Remove restriction that the
10037 source and destination registers in the extension are the same.
10038 (find_and_remove_re): Emit a copy from the extension's
10039 destination to its source after the defining insn if
10040 the source and destination registers are different.
10042 PR middle-end/59285
10043 * ifcvt.c (merge_if_block): If we are merging a block with more than
10044 one successor with a block with no successors, remove any BARRIER
10045 after the second block.
10047 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10049 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10051 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10054 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10055 for 14-bit register offsets when INT14_OK_STRICT is false.
10057 2014-01-07 Roland Stigge <stigge@antcom.de>
10058 Michael Meissner <meissner@linux.vnet.ibm.com>
10061 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10062 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10064 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10066 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10069 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10071 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10072 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10073 rtx is const0_rtx or not.
10075 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10078 * target-globals.c (save_target_globals): Remove this_fn_optab
10080 * toplev.c: Include optabs.h.
10081 (target_reinit): Temporarily restore the global options if another
10082 set of options are in force.
10084 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10086 PR rtl-optimization/58668
10087 * cfgcleanup.c (flow_find_cross_jump): Don't count
10088 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10089 to determine what is counted.
10090 (flow_find_head_matching_sequence): Use active_insn_p to determine
10092 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10094 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10095 determine what is counted.
10097 PR tree-optimization/59643
10098 * tree-predcom.c (split_data_refs_to_components): If one dr is
10099 read and one write, determine_offset fails and the write isn't
10100 in the bad component, just put the read into the bad component.
10102 2014-01-07 Mike Stump <mikestump@comcast.net>
10103 Jakub Jelinek <jakub@redhat.com>
10106 * tree-core.h (struct tree_optimization_option): Change optabs
10107 type from unsigned char * to void *.
10108 * optabs.c (init_tree_optimization_optabs): Adjust
10109 TREE_OPTIMIZATION_OPTABS initialization.
10111 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10114 * config/i386/i386.h (struct machine_function): Add
10115 no_drap_save_restore field.
10116 * config/i386/i386.c (ix86_save_reg): Use
10117 !cfun->machine->no_drap_save_restore instead of
10118 crtl->stack_realign_needed.
10119 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10120 this function clears frame_pointer_needed. Set
10121 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10122 and DRAP reg is needed.
10124 2014-01-06 Marek Polacek <polacek@redhat.com>
10127 * doc/implement-c.texi: Mention that other integer types are
10128 permitted as bit-field types in strictly conforming mode.
10130 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10132 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10133 is newly allocated.
10135 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10137 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10139 2014-01-06 Martin Jambor <mjambor@suse.cz>
10142 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10144 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10146 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10150 * var-tracking.c (add_stores): Preserve the value of the source even if
10151 we don't record the store.
10153 2014-01-06 Terry Guo <terry.guo@arm.com>
10155 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10157 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10160 * config/darwin.c (darwin_function_section): Adjust return values to
10161 correspond to optimisation changes made in r206070.
10163 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10165 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10166 from prefetch_block tune setting.
10167 (nocona_cost): Correct size of prefetch block to 64.
10169 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10171 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10172 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10173 used to save the static chain register in the computation of the offset
10174 from which the FP registers need to be restored.
10176 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10178 PR tree-optimization/59519
10179 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10180 ICE if get_current_def (current_new_name) is already non-NULL, as long
10181 as it is a phi result of some other phi in *new_exit_bb that has
10184 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10185 or vmovdqu* for misaligned_operand.
10186 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10187 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10188 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10189 aligned_mem for AVX512F masked aligned load and store builtins and for
10190 non-temporal moves.
10192 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10194 PR tree-optimization/59651
10195 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10196 Address range for negative step should be added by TYPE_SIZE_UNIT.
10198 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10200 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10201 overlapping registers also for registers other than the stack pointer.
10203 2014-01-03 Marek Polacek <polacek@redhat.com>
10206 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10209 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10212 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10215 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10216 (push splitter): Use <P:MODE_SIZE> instead of
10217 GET_MODE_SIZE (<P:MODE>mode).
10218 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10219 (mov -1, reg peephole2): Likewise.
10220 * config/i386/sse.md (*mov<mode>_internal,
10221 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10222 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10223 *<code><mode>3, *andnot<mode>3<mask_name>,
10224 <mask_codefor><code><mode>3<mask_name>): Likewise.
10225 * config/i386/subst.md (mask_mode512bit_condition,
10226 sd_mask_mode512bit_condition): Likewise.
10228 2014-01-02 Xinliang David Li <davidxl@google.com>
10230 PR tree-optimization/59303
10231 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10232 (dump_predicates): Better output format.
10233 (pred_equal_p): New function.
10234 (is_neq_relop_p): Ditto.
10235 (is_neq_zero_form_p): Ditto.
10236 (pred_expr_equal_p): Ditto.
10237 (pred_neg_p): Ditto.
10238 (simplify_pred): Ditto.
10239 (simplify_preds_2): Ditto.
10240 (simplify_preds_3): Ditto.
10241 (simplify_preds_4): Ditto.
10242 (simplify_preds): Ditto.
10243 (push_pred): Ditto.
10244 (push_to_worklist): Ditto.
10245 (get_pred_info_from_cmp): Ditto.
10246 (is_degenerated_phi): Ditto.
10247 (normalize_one_pred_1): Ditto.
10248 (normalize_one_pred): Ditto.
10249 (normalize_one_pred_chain): Ditto.
10250 (normalize_preds): Ditto.
10251 (normalize_cond_1): Remove function.
10252 (normalize_cond): Ditto.
10253 (is_gcond_subset_of): Ditto.
10254 (is_subset_of_any): Ditto.
10255 (is_or_set_subset_of): Ditto.
10256 (is_and_set_subset_of): Ditto.
10257 (is_norm_cond_subset_of): Ditto.
10258 (pred_chain_length_cmp): Ditto.
10259 (convert_control_dep_chain_into_preds): Type change.
10260 (find_predicates): Ditto.
10261 (find_def_preds): Ditto.
10262 (destroy_predicates_vecs): Ditto.
10263 (find_matching_predicates_in_rest_chains): Ditto.
10264 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10265 (is_pred_expr_subset): Ditto.
10266 (is_pred_chain_subset_of): Ditto.
10267 (is_included_in): Ditto.
10268 (is_superset_of): Ditto.
10270 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10272 Update copyright years.
10274 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10276 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10277 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10278 config/arc/arc.md, config/arc/arc.opt,
10279 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10280 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10281 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10282 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10283 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10284 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10285 vtable-verify.c, vtable-verify.h: Use the standard form for the
10288 2014-01-02 Tobias Burnus <burnus@net-b.de>
10290 * gcc.c (process_command): Update copyright notice dates.
10291 * gcov-dump.c: Ditto.
10293 * doc/cpp.texi: Bump @copying's copyright year.
10294 * doc/cppinternals.texi: Ditto.
10295 * doc/gcc.texi: Ditto.
10296 * doc/gccint.texi: Ditto.
10297 * doc/gcov.texi: Ditto.
10298 * doc/install.texi: Ditto.
10299 * doc/invoke.texi: Ditto.
10301 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10303 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10305 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10307 * config/i386/sse.md (*mov<mode>_internal): Guard
10308 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10310 PR rtl-optimization/59647
10311 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10312 new_rtx into UNSIGNED_FLOAT rtxes.
10314 Copyright (C) 2014 Free Software Foundation, Inc.
10316 Copying and distribution of this file, with or without modification,
10317 are permitted in any medium without royalty provided the copyright
10318 notice and this notice are preserved.