1 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6 * wide-int.cc (UTItype): Define.
7 (UDWtype): Define for appropriate W_TYPE_SIZE.
9 2014-05-09 Richard Biener <rguenther@suse.de>
11 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
12 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
13 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
14 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
15 ssa_propagate): Adjust.
17 2014-05-08 Jeff Law <law@redhat.com>
19 PR tree-optimization/61009
20 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
21 tri-state rather than a boolean. When a block is too big to
22 thread through, inform caller via negative return value.
23 (thread_across_edge): If a block was too big for normal threading,
24 then it's too big for a joiner too, so remove temporary equivalences
25 and return immediately.
27 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
28 Matthias Klose <doko@ubuntu.com>
31 * optc-gen.awk: Fix option handling for -Wunused-parameter.
33 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
36 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
38 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
41 * config/alpha/alpha.c: Include gimple-iterator.h.
42 (alpha_gimple_fold_builtin): New function. Move
43 ALPHA_BUILTIN_UMULH folding from ...
44 (alpha_fold_builtin): ... here.
45 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
47 2014-05-08 Wei Mi <wmi@google.com>
50 * config/i386/i386.c (ix86_compute_frame_layout): Update
51 preferred_stack_boundary for call, expanded from tls descriptor.
52 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
53 to depend on SP register.
54 (*tls_local_dynamic_base_32_gnu): Ditto.
55 (*tls_local_dynamic_32_once): Ditto.
56 (tls_global_dynamic_64_<mode>): Set
57 ix86_tls_descriptor_calls_expanded_in_cfun.
58 (tls_local_dynamic_base_64_<mode>): Ditto.
59 (tls_global_dynamic_32): Set
60 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
61 to depend on SP register.
62 (tls_local_dynamic_base_32): Ditto.
64 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
66 * config/arm/arm_neon.h: Update comment.
67 * config/arm/neon-docgen.ml: Delete.
68 * config/arm/neon-gen.ml: Delete.
69 * doc/arm-neon-intrinsics.texi: Update comment.
71 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
73 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
75 (vand, vorr, veor, vorn, vbic): Remove.
76 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
78 (neon_vsub_unspec): Likewise.
79 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
81 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
83 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
93 (vaddq_s16): Likewise.
94 (vaddq_s32): Likewise.
95 (vaddq_s64): Likewise.
96 (vaddq_f32): Likewise.
98 (vaddq_u16): Likewise.
99 (vaddq_u32): Likewise.
100 (vaddq_u64): Likewise.
102 (vmul_s16): Likewise.
103 (vmul_s32): Likewise.
104 (vmul_f32): Likewise.
106 (vmul_u16): Likewise.
107 (vmul_u32): Likewise.
109 (vmulq_s8): Likewise.
110 (vmulq_s16): Likewise.
111 (vmulq_s32): Likewise.
112 (vmulq_f32): Likewise.
113 (vmulq_u8): Likewise.
114 (vmulq_u16): Likewise.
115 (vmulq_u32): Likewise.
117 (vsub_s16): Likewise.
118 (vsub_s32): Likewise.
119 (vsub_f32): Likewise.
121 (vsub_u16): Likewise.
122 (vsub_u32): Likewise.
123 (vsub_s64): Likewise.
124 (vsub_u64): Likewise.
125 (vsubq_s8): Likewise.
126 (vsubq_s16): Likewise.
127 (vsubq_s32): Likewise.
128 (vsubq_s64): Likewise.
129 (vsubq_f32): Likewise.
130 (vsubq_u8): Likewise.
131 (vsubq_u16): Likewise.
132 (vsubq_u32): Likewise.
133 (vsubq_u64): Likewise.
135 (vand_s16): Likewise.
136 (vand_s32): Likewise.
138 (vand_u16): Likewise.
139 (vand_u32): Likewise.
140 (vand_s64): Likewise.
141 (vand_u64): Likewise.
142 (vandq_s8): Likewise.
143 (vandq_s16): Likewise.
144 (vandq_s32): Likewise.
145 (vandq_s64): Likewise.
146 (vandq_u8): Likewise.
147 (vandq_u16): Likewise.
148 (vandq_u32): Likewise.
149 (vandq_u64): Likewise.
151 (vorr_s16): Likewise.
152 (vorr_s32): Likewise.
154 (vorr_u16): Likewise.
155 (vorr_u32): Likewise.
156 (vorr_s64): Likewise.
157 (vorr_u64): Likewise.
158 (vorrq_s8): Likewise.
159 (vorrq_s16): Likewise.
160 (vorrq_s32): Likewise.
161 (vorrq_s64): Likewise.
162 (vorrq_u8): Likewise.
163 (vorrq_u16): Likewise.
164 (vorrq_u32): Likewise.
165 (vorrq_u64): Likewise.
167 (veor_s16): Likewise.
168 (veor_s32): Likewise.
170 (veor_u16): Likewise.
171 (veor_u32): Likewise.
172 (veor_s64): Likewise.
173 (veor_u64): Likewise.
174 (veorq_s8): Likewise.
175 (veorq_s16): Likewise.
176 (veorq_s32): Likewise.
177 (veorq_s64): Likewise.
178 (veorq_u8): Likewise.
179 (veorq_u16): Likewise.
180 (veorq_u32): Likewise.
181 (veorq_u64): Likewise.
183 (vbic_s16): Likewise.
184 (vbic_s32): Likewise.
186 (vbic_u16): Likewise.
187 (vbic_u32): Likewise.
188 (vbic_s64): Likewise.
189 (vbic_u64): Likewise.
190 (vbicq_s8): Likewise.
191 (vbicq_s16): Likewise.
192 (vbicq_s32): Likewise.
193 (vbicq_s64): Likewise.
194 (vbicq_u8): Likewise.
195 (vbicq_u16): Likewise.
196 (vbicq_u32): Likewise.
197 (vbicq_u64): Likewise.
199 (vorn_s16): Likewise.
200 (vorn_s32): Likewise.
202 (vorn_u16): Likewise.
203 (vorn_u32): Likewise.
204 (vorn_s64): Likewise.
205 (vorn_u64): Likewise.
206 (vornq_s8): Likewise.
207 (vornq_s16): Likewise.
208 (vornq_s32): Likewise.
209 (vornq_s64): Likewise.
210 (vornq_u8): Likewise.
211 (vornq_u16): Likewise.
212 (vornq_u32): Likewise.
213 (vornq_u64): Likewise.
215 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
217 * wide-int.cc (UTItype): Define.
218 (UDWtype): Define for appropriate W_TYPE_SIZE.
220 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
222 PR tree-optimization/59100
223 * tree-ssa-phiopt.c: Include tree-inline.h.
224 (neutral_element_p, absorbing_element_p): New functions.
225 (value_replacement): Handle conditional binary operations with a
226 neutral or absorbing element.
228 2014-05-08 Richard Biener <rguenther@suse.de>
230 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
231 folding the expression.
232 (valueize_expr): Remove.
233 (visit_reference_op_load): Do not valueize the result of
235 (simplify_binary_expression): Likewise.
236 (simplify_unary_expression): Likewise.
238 2014-05-08 Richard Biener <rguenther@suse.de>
240 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
241 looking at TYPE_ARG_TYPES.
243 2014-05-08 Richard Biener <rguenther@suse.de>
245 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
246 pointer propagation special-case.
248 2014-05-08 Bin Cheng <bin.cheng@arm.com>
250 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
251 core part of address expressions.
253 2014-05-08 Alan Modra <amodra@gmail.com>
256 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
257 loads and stores when -mno-strict-align at any alignment.
258 (expand_block_clear): Similarly. Also correct calculation of
261 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
264 * tree-complex.c (expand_complex_move): Keep line info when expanding
266 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
267 of complex expression. Use new argument to display correct location
268 for values coming from phi statement.
269 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
270 (warn_uninitialized_phi): Pass location of phi argument to
272 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
273 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
275 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
277 * config/rs6000/predicates.md (indexed_address_mem): New.
278 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
279 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
280 fpstore_ux, fpstore_u.
281 (sign_extend, indexed, update): New.
282 (cell_micro): Adjust.
283 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
284 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
285 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
286 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
287 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
288 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
289 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
290 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
291 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
292 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
293 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
294 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
295 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
296 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
297 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
299 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
300 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
301 *vsx_extract_<mode>_store): Adjust.
302 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
303 is_cracked_insn, insn_must_be_first_in_group,
304 insn_must_be_last_in_group): Adjust.
306 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
308 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
309 ppc440-fpstore): Adjust.
310 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
311 ppc476-fpstore): Adjust.
312 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
313 ppc601-fpstore): Adjust.
314 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
316 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
318 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
319 ppc7450-fpstore): Adjust.
320 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
321 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
322 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
324 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
325 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
326 cell-fpstore, cell-fpstore-update): Adjust.
327 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
328 ppce300c3_store, ppce300c3_fpstore): Adjust.
329 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
330 e500mc_fpstore): Adjust.
331 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
332 e500mc64_store, e500mc64_fpstore): Adjust.
333 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
334 e5500_fpstore): Adjust.
335 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
336 e6500_fpstore): Adjust.
337 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
339 * config/rs6000/power4.md (power4-load, power4-load-ext,
340 power4-load-ext-update, power4-load-ext-update-indexed,
341 power4-load-update-indexed, power4-load-update, power4-fpload,
342 power4-fpload-update, power4-store, power4-store-update,
343 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
345 * config/rs6000/power5.md (power5-load, power5-load-ext,
346 power5-load-ext-update, power5-load-ext-update-indexed,
347 power5-load-update-indexed, power5-load-update, power5-fpload,
348 power5-fpload-update, power5-store, power5-store-update,
349 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
351 * config/rs6000/power6.md (power6-load, power6-load-ext,
352 power6-load-update, power6-load-update-indexed,
353 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
354 power6-fpload-update, power6-store, power6-store-update,
355 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
357 * config/rs6000/power7.md (power7-load, power7-load-ext,
358 power7-load-update, power7-load-update-indexed,
359 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
360 power7-fpload-update, power7-store, power7-store-update,
361 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
363 * config/rs6000/power8.md (power8-load, power8-load-update,
364 power8-load-ext, power8-load-ext-update, power8-fpload,
365 power8-fpload-update, power8-store, power8-store-update-indexed,
366 power8-fpstore, power8-fpstore-update): Adjust.
367 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
369 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
370 titan_lsu_store, titan_lsu_fpstore): Adjust.
371 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
373 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
376 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
377 unrolled byte insns. Emit address increments after move insns.
379 2014-05-07 David Malcolm <dmalcolm@redhat.com>
381 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
382 const_gimple, rather than a gimple.
383 (gimple_call_builtin_p): Likewise, for the three variants.
385 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
386 (gimple_call_builtin_p): Likewise, for the three variants.
388 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
390 PR tree-optimization/61095
391 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
393 2014-05-07 Richard Biener <rguenther@suse.de>
395 PR tree-optimization/61034
396 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
397 (maybe_skip_until): Use translate to take into account
398 lattices when trying to do disambiguations.
399 (get_continuation_for_phi_1): Likewise.
400 (get_continuation_for_phi): Adjust for added translate arguments.
401 (walk_non_aliased_vuses): Likewise.
402 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
403 (walk_non_aliased_vuses): Likewise.
404 (call_may_clobber_ref_p_1): Declare.
405 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
406 calls. Stop early if we are only supposed to disambiguate.
407 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
409 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
411 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
412 Emit an error when the function has arguments.
414 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
416 * cfgloop.h (unswitch_loops): Remove.
417 * doc/passes.texi: Remove references to loop-unswitch.c
418 * timevar.def (TV_LOOP_UNSWITCH): Remove.
420 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
422 * tree-vect-data-refs.c (vect_grouped_load_supported): New
423 check for loads group of length 3.
424 (vect_permute_load_chain): New permutations for loads group of
426 * tree-vect-stmts.c (vect_model_load_cost): Change cost
427 of vec_perm_shuffle for the new permutations.
429 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
431 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
432 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
433 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
434 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
435 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
436 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
437 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
438 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
440 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
442 * loop-unswitch.c: Delete.
444 2014-05-07 Richard Biener <rguenther@suse.de>
446 * config.gcc: Always set need_64bit_hwint to yes.
448 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
450 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
451 of using optimize_size.
453 2014-05-06 Mike Stump <mikestump@comcast.net>
455 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
457 2014-05-06 Joseph Myers <joseph@codesourcery.com>
459 * config/i386/sse.md (*mov<mode>_internal)
460 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
461 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
462 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
463 (*<code><mode>3, *andnot<mode>3<mask_name>)
464 (<mask_codefor><code><mode>3<mask_name>): Only consider
465 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
467 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
470 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
472 * lra-constraints.c (valid_address_p): Move earlier in file.
473 Add a constraint argument to the address_info version.
474 (satisfies_memory_constraint_p): New function.
475 (satisfies_address_constraint_p): Likewise.
476 (process_alt_operands, curr_insn_transform): Use them.
477 (process_address): Pass the constraint to valid_address_p when
478 checking address operands.
480 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
482 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
483 to their respective blocks. Fix inadvertent use of "node".
485 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
487 * emit-rtl.c (init_derived_machine_modes): New functionm, split
489 (init_emit_once): ...here.
490 * rtl.h (init_derived_machine_modes): Declare.
491 * toplev.c (do_compile): Call it even if no_backend.
493 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
494 Mike Stump <mikestump@comcast.net>
495 Richard Sandiford <rdsandiford@googlemail.com>
496 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
498 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
499 (rtx_equal_for_memref_p): Update comment.
500 (adjust_offset_for_component_ref): Use wide-int interfaces.
501 * builtins.c (get_object_alignment_2): Likewise.
502 (c_readstr): Likewise.
503 (target_char_cast): Add comment.
504 (determine_block_size): Use wide-int interfaces.
505 (expand_builtin_signbit): Likewise.
506 (fold_builtin_int_roundingfn): Likewise.
507 (fold_builtin_bitop): Likewise.
508 (fold_builtin_bswap): Likewise.
509 (fold_builtin_logarithm): Use signop.
510 (fold_builtin_pow): Likewise.
511 (fold_builtin_memory_op): Use wide-int interfaces.
512 (fold_builtin_object_size): Likewise.
513 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
514 nb_iterations_estimate.
515 (record_niter_bound): Use wide-int interfaces.
516 (get_estimated_loop_iterations_int): Likewise.
517 (get_estimated_loop_iterations): Likewise.
518 (get_max_loop_iterations): Likewise.
519 * cfgloop.h: Include wide-int.h.
520 (struct nb_iter_bound): Change bound to widest_int.
521 (struct loop): Change nb_iterations_upper_bound and
522 nb_iterations_estimate to widest_int.
523 (record_niter_bound): Switch to use widest_int.
524 (get_estimated_loop_iterations): Likewise.
525 (get_max_loop_iterations): Likewise.
526 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
528 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
529 * combine.c (try_combine): Likewise.
530 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
531 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
533 (aarch64_float_const_representable_p): Likewise.
534 * config/arc/arc.c: Include wide-int.h.
535 (arc_can_use_doloop_p): Use wide-int interfaces.
536 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
537 (vfp3_const_double_index): Likewise.
538 * config/avr/avr.c (avr_out_round): Likewise.
539 (avr_fold_builtin): Likewise.
540 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
541 (bfin_can_use_doloop_p): Likewise.
542 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
543 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
544 * config/i386/i386.c: Include wide-int.h.
545 (ix86_data_alignment): Use wide-int interfaces.
546 (ix86_local_alignment): Likewise.
547 (ix86_emit_swsqrtsf): Update real_from_integer.
548 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
549 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
550 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
551 (zero_constant): Likewise.
552 (input_operand): Likewise.
553 (splat_input_operand): Likewise.
554 (non_logical_cint_operand): Change const_double to const_wide_int.
555 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
556 (easy_altivec_constant): Remove comment.
557 (paired_expand_vector_init): Use CONSTANT_P.
558 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
559 (rs6000_emit_move): Update checks.
560 (rs6000_aggregate_candidate): Use wide-int interfaces.
561 (rs6000_expand_ternop_builtin): Likewise.
562 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
563 (rs6000_assemble_integer): Likewise.
564 (rs6000_hash_constant): Likewise.
565 (output_toc): Likewise.
566 (rs6000_rtx_costs): Likewise.
567 (rs6000_emit_swrsqrt); Update call to real_from_integer.
568 * config/rs6000/rs6000-c.c: Include wide-int.h.
569 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
570 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
571 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
572 Handle CONST_WIDE_INT.
573 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
574 Use tree_fits_uhwi_p.
575 * config/sparc/sparc.c: Include wide-int.h.
576 (sparc_fold_builtin): Use wide-int interfaces.
577 * config/vax/vax.c: Include wide-int.h.
578 (vax_float_literal): Use real_from_integer.
579 * coretypes.h (struct hwivec_def): New.
582 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
583 (equiv_constant): Handle CONST_WIDE_INT.
584 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
585 (cselib_hash_rtx): Handle CONST_WIDE_INT.
586 * dbxout.c (stabstr_U): Use wide-int interfaces.
587 (dbxout_type): Update to use cst_fits_shwi_p.
588 * defaults.h (LOG2_BITS_PER_UNIT): Define.
589 (TARGET_SUPPORTS_WIDE_INT): Add default.
590 * dfp.c: Include wide-int.h.
591 (decimal_real_to_integer2): Use wide-int interfaces and rename to
592 decimal_real_to_integer.
593 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
594 decimal_real_to_integer.
595 * doc/generic.texi (Constant expressions): Update for wide_int.
596 * doc/rtl.texi (const_double): Likewise.
597 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
598 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
599 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
600 (REAL_VALUE_FROM_INT): Remove.
601 (TARGET_SUPPORTS_WIDE_INT): New.
602 * doc/tm.texi: Regenerate.
603 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
604 * double-int.h: Include wide-int.h.
605 (struct wi::int_traits): New.
606 * dwarf2out.c (get_full_len): New.
607 (dw_val_equal_p): Add case dw_val_class_wide_int.
608 (size_of_loc_descr): Likewise.
609 (output_loc_operands): Likewise.
610 (insert_double): Remove.
611 (insert_wide_int): New.
613 (print_die): Add case dw_val_class_wide_int.
614 (attr_checksum): Likewise.
615 (attr_checksum_ordered): Likewise.
616 (same_dw_val_p): Likewise.
617 (size_of_die): Likewise.
618 (value_format): Likewise.
619 (output_die): Likewise.
620 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
622 (clz_loc_descriptor): Use wide-int interfaces.
623 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
624 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
625 (round_up_to_align): Use wide-int interfaces.
626 (field_byte_offset): Likewise.
627 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
628 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
629 CONST_DOUBLE handling. Use wide-int interfaces.
630 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
631 (gen_enumeration_type_die): Use add_AT_wide.
632 (hash_loc_operands): Add case dw_val_class_wide_int.
633 (compare_loc_operands): Likewise.
634 * dwarf2out.h: Include wide-int.h.
636 (enum dw_val_class): Add dw_val_class_wide_int.
637 (struct dw_val_struct): Add val_wide.
638 * emit-rtl.c (const_wide_int_htab): New.
639 (const_wide_int_htab_hash): New.
640 (const_wide_int_htab_eq): New.
641 (lookup_const_wide_int): New.
642 (const_double_htab_hash): Use wide-int interfaces.
643 (const_double_htab_eq): Likewise.
644 (rtx_to_double_int): Conditionally compile for wide-int.
645 (immed_double_int_const): Rename to immed_wide_int_const and
647 (immed_double_const): Conditionally compile for wide-int.
648 (init_emit_once): Use wide-int interfaces.
649 * explow.c (plus_constant): Likewise.
650 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
651 (lshift_value): Use wide-int interfaces.
652 (expand_mult): Likewise.
653 (choose_multiplier): Likewise.
654 (expand_smod_pow2): Likewise.
655 (make_tree): Likewise.
656 * expr.c (convert_modes): Consolidate handling of constants.
657 Use wide-int interfaces.
658 (emit_group_load_1): Add note.
659 (store_expr): Update comment.
660 (get_inner_reference): Use wide-int interfaces.
661 (expand_constructor): Update comment.
662 (expand_expr_real_2): Use wide-int interfaces.
663 (expand_expr_real_1): Likewise.
664 (reduce_to_bit_field_precision): Likewise.
665 (const_vector_from_tree): Likewise.
666 * final.c: Include wide-int-print.h.
667 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
668 * fixed-value.c: Include wide-int.h.
669 (fixed_from_string): Use wide-int interfaces.
670 (fixed_to_decimal): Likewise.
671 (fixed_convert_from_real): Likewise.
672 (real_convert_from_fixed): Likewise.
673 * fold-const.h (mem_ref_offset): Return an offset_int.
674 (div_if_zero_remainder): Remove code parameter.
675 * fold-const.c (div_if_zero_remainder): Remove code parameter.
676 Use wide-int interfaces.
677 (may_negate_without_overflow_p): Use wide-int interfaces.
678 (negate_expr_p): Likewise.
679 (fold_negate_expr): Likewise.
680 (int_const_binop_1): Likewise.
681 (const_binop): Likewise.
682 (fold_convert_const_int_from_int): Likewise.
683 (fold_convert_const_int_from_real): Likewise.
684 (fold_convert_const_int_from_fixed): Likewise.
685 (fold_convert_const_fixed_from_int): Likewise.
686 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
687 (sign_bit_p): Use wide-int interfaces.
688 (make_range_step): Likewise.
689 (build_range_check): Likewise. Pass an integer of the correct type
690 instead of using integer_one_node.
691 (range_predecessor): Pass an integer of the correct type instead
692 of using integer_one_node.
693 (range_successor): Likewise.
694 (merge_ranges): Likewise.
695 (unextend): Use wide-int interfaces.
696 (extract_muldiv_1): Likewise.
697 (fold_div_compare): Likewise.
698 (fold_single_bit_test): Likewise.
699 (fold_sign_changed_comparison): Likewise.
700 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
701 (fold_plusminus_mult_expr): Use wide-int interfaces.
702 (native_encode_int): Likewise.
703 (native_interpret_int): Likewise.
704 (fold_unary_loc): Likewise.
705 (pointer_may_wrap_p): Likewise.
706 (size_low_cst): Likewise.
707 (mask_with_tz): Likewise.
708 (fold_binary_loc): Likewise.
709 (fold_ternary_loc): Likewise.
710 (multiple_of_p): Likewise.
711 (tree_call_nonnegative_warnv_p): Update calls to
712 tree_int_cst_min_precision and real_from_integer.
713 (fold_negate_const): Use wide-int interfaces.
714 (fold_abs_const): Likewise.
715 (fold_relational_const): Use tree_int_cst_lt.
716 (round_up_loc): Use wide-int interfaces.
717 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
718 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
719 * gengtype.c: Remove include of double-int.h.
720 (do_typedef): Use wide-int interfaces.
721 (open_base_files): Add wide-int.h.
722 (main): Add offset_int and widest_int typedefs.
723 * gengtype-lex.l: Handle "^".
724 (CXX_KEYWORD): Add "static".
725 * gengtype-parse.c (require3): New.
726 (require_template_declaration): Handle constant template arguments
727 and nested templates.
728 * gengtype-state.c: Don't include "double-int.h".
729 * genpreds.c (write_one_predicate_function): Update comment.
730 (write_tm_constrs_h): Add check for hval and lval use in
732 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
733 (add_to_sequence): Likewise.
734 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
735 and const_double_operand.
736 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
738 * gimple-fold.c (get_base_constructor): Likewise.
739 (fold_array_ctor_reference): Likewise.
740 (fold_nonarray_ctor_reference): Likewise.
741 (fold_const_aggregate_ref_1): Likewise.
742 (gimple_val_nonnegative_real_p): Likewise.
743 (gimple_fold_indirect_ref): Likewise.
744 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
745 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
746 (struct slsr_cand_d): Change index to be widest_int.
747 (struct incr_info_d): Change incr to be widest_int.
748 (alloc_cand_and_find_basis): Use wide-int interfaces.
749 (slsr_process_phi): Likewise.
750 (backtrace_base_for_ref): Likewise. Return a widest_int.
751 (restructure_reference): Take a widest_int instead of a double_int.
752 (slsr_process_ref): Use wide-int interfaces.
753 (create_mul_ssa_cand): Likewise.
754 (create_mul_imm_cand): Likewise.
755 (create_add_ssa_cand): Likewise.
756 (create_add_imm_cand): Take a widest_int instead of a double_int.
757 (slsr_process_add): Use wide-int interfaces.
758 (slsr_process_cast): Likewise.
759 (slsr_process_copy): Likewise.
760 (dump_candidate): Likewise.
761 (dump_incr_vec): Likewise.
762 (replace_ref): Likewise.
763 (cand_increment): Likewise. Return a widest_int.
764 (cand_abs_increment): Likewise.
765 (replace_mult_candidate): Take a widest_int instead of a double_int.
766 (replace_unconditional_candidate): Use wide-int interfaces.
767 (incr_vec_index): Take a widest_int instead of a double_int.
768 (create_add_on_incoming_edge): Likewise.
769 (create_phi_basis): Use wide-int interfaces.
770 (replace_conditional_candidate): Likewise.
771 (record_increment): Take a widest_int instead of a double_int.
772 (record_phi_increments): Use wide-int interfaces.
773 (phi_incr_cost): Take a widest_int instead of a double_int.
774 (lowest_cost_path): Likewise.
775 (total_savings): Likewise.
776 (analyze_increments): Use wide-int interfaces.
777 (ncd_with_phi): Take a widest_int instead of a double_int.
778 (ncd_of_cand_and_phis): Likewise.
779 (nearest_common_dominator_for_cands): Likewise.
780 (insert_initializers): Use wide-int interfaces.
781 (all_phi_incrs_profitable): Likewise.
782 (replace_one_candidate): Likewise.
783 (replace_profitable_candidates): Likewise.
784 * godump.c: Include wide-int-print.h.
785 (go_output_typedef): Use wide-int interfaces.
786 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
787 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
788 (build_loop_iteration_domains): Likewise.
789 * hooks.h: Include wide-int.h rather than double-int.h.
790 (hook_bool_dint_dint_uint_bool_true): Delete.
791 (hook_bool_wint_wint_uint_bool_true): Declare.
792 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
793 (hook_bool_wint_wint_uint_bool_true): New.
794 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
796 (ubsan_expand_si_overflow_mul_check): Likewise.
797 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
798 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
799 (get_ancestor_addr_info): Likewise.
800 (ipa_modify_call_arguments): Likewise.
801 * loop-doloop.c (doloop_modify): Likewise.
802 (doloop_optimize): Likewise.
803 * loop-iv.c (iv_number_of_iterations): Likewise.
804 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
805 (unroll_loop_constant_iterations): Likewise.
806 (decide_unroll_runtime_iterations): Likewise.
807 (unroll_loop_runtime_iterations): Likewise.
808 (decide_peel_simple): Likewise.
809 (decide_unroll_stupid): Likewise.
810 * lto-streamer-in.c (streamer_read_wi): Add.
811 (input_cfg): Use wide-int interfaces.
812 (lto_input_tree_1): Likewise.
813 * lto-streamer-out.c (streamer_write_wi): Add.
814 (hash_tree): Use wide-int interfaces.
815 (output_cfg): Likewise.
816 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
817 (GTFILES): Add wide-int.h and signop.h.
818 (TAGS): Look for .cc files too.
819 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
820 * optabs.c (expand_subword_shift): Likewise.
821 (expand_doubleword_shift): Likewise.
822 (expand_absneg_bit): Likewise.
823 (expand_copysign_absneg): Likewise.
824 (expand_copysign_bit): Likewise.
825 * postreload.c (reload_cse_simplify_set): Likewise.
826 * predict.c (predict_iv_comparison): Likewise.
827 * pretty-print.h: Include wide-int-print.h.
829 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
830 * print-tree.c: Include wide-int-print.h.
831 (print_node_brief): Use wide-int interfaces.
832 (print_node): Likewise.
833 * read-rtl.c (validate_const_wide_int): New.
834 (read_rtx_code): Add CONST_WIDE_INT case.
835 * real.c: Include wide-int.h.
836 (real_to_integer2): Delete.
837 (real_to_integer): New function, returning a wide_int.
838 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
839 (ten_to_ptwo): Update call to real_from_integer.
840 (real_digit): Likewise.
841 * real.h: Include signop.h, wide-int.h and insn-modes.h.
842 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
843 (REAL_VALUE_TO_INT): Delete.
844 (real_to_integer): Declare a wide-int form.
845 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
846 * recog.c (const_int_operand): Improve comment.
847 (const_scalar_int_operand): New.
848 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
849 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
850 (split_double): Likewise.
851 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
852 (rtx_size): Likewise.
853 (rtx_alloc_stat_v): New.
854 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
855 (cwi_output_hex): New.
856 (iterative_hash_rtx): Handle CONST_WIDE_INT.
857 (cwi_check_failed_bounds): New.
858 * rtl.def (CONST_WIDE_INT): New.
859 * rtl.h: Include <utility> and wide-int.h.
860 (struct hwivec_def): New.
861 (CWI_GET_NUM_ELEM): New.
862 (CWI_PUT_NUM_ELEM): New.
863 (struct rtx_def): Add num_elem and hwiv.
864 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
865 (CASE_CONST_UNIQUE): Likewise.
866 (CASE_CONST_ANY): Likewise.
867 (CONST_SCALAR_INT_P): Likewise.
868 (CONST_WIDE_INT_P): New.
871 (cwi_check_failed_bounds): New.
874 (CONST_WIDE_INT_VEC) New.
875 (CONST_WIDE_INT_NUNITS) New.
876 (CONST_WIDE_INT_ELT) New.
877 (rtx_mode_t): New type.
878 (wi::int_traits <rtx_mode_t>): New.
880 (wi::min_value): New.
881 (wi::max_value): New.
883 (const_wide_int_alloc): New.
884 (immed_wide_int_const): New.
885 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
886 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
887 * signop.h: New file.
888 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
889 (simplify_const_unary_operation): Use wide-int interfaces.
890 (simplify_binary_operation_1): Likewise.
891 (simplify_const_binary_operation): Likewise.
892 (simplify_const_relational_operation): Likewise.
893 (simplify_immed_subreg): Likewise.
894 * stmt.c (expand_case): Likewise.
895 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
896 signop rather than a bool.
897 * stor-layout.c (layout_type): Use wide-int interfaces.
898 (initialize_sizetypes): Update calls to
899 set_min_and_max_values_for_integral_type.
900 (set_min_and_max_values_for_integral_type): Take a signop rather
901 than a bool. Use wide-int interfaces.
902 (fixup_signed_type): Update accordingly. Remove
903 HOST_BITS_PER_DOUBLE_INT limit.
904 (fixup_unsigned_type): Likewise.
905 * system.h (STATIC_CONSTANT_P): New.
906 (STATIC_ASSERT): New.
907 * target.def (can_use_doloop_p): Take widest_ints rather than
909 * target.h: Include wide-int.h rather than double-int.h.
910 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
912 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
913 rather than INT_CST_LT_UNSIGNED.
914 (can_use_doloop_if_innermost): Take widest_ints rather than
916 * tree-affine.c: Include wide-int-print.h.
917 (double_int_ext_for_comb): Delete.
918 (wide_int_ext_for_comb): New.
919 (aff_combination_zero): Use wide-int interfaces.
920 (aff_combination_const): Take a widest_int instead of a double_int.
921 (aff_combination_elt): Use wide-int interfaces.
922 (aff_combination_scale): Take a widest_int instead of a double_int.
923 (aff_combination_add_elt): Likewise.
924 (aff_combination_add_cst): Likewise.
925 (aff_combination_add): Use wide-int interfaces.
926 (aff_combination_convert): Likewise.
927 (tree_to_aff_combination): Likewise.
928 (add_elt_to_tree): Take a widest_int instead of a double_int.
929 (aff_combination_to_tree): Use wide-int interfaces.
930 (aff_combination_remove_elt): Likewise.
931 (aff_combination_add_product): Take a widest_int instead of
933 (aff_combination_mult): Use wide-int interfaces.
934 (aff_combination_expand): Likewise.
935 (double_int_constant_multiple_p): Delete.
936 (wide_int_constant_multiple_p): New.
937 (aff_combination_constant_multiple_p): Take a widest_int pointer
938 instead of a double_int pointer.
939 (print_aff): Use wide-int interfaces.
940 (get_inner_reference_aff): Take a widest_int pointer
941 instead of a double_int pointer.
942 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
943 * tree-affine.h: Include wide-int.h.
944 (struct aff_comb_elt): Change type of coef to widest_int.
945 (struct affine_tree_combination): Change type of offset to widest_int.
946 (double_int_ext_for_comb): Delete.
947 (wide_int_ext_for_comb): New.
948 (aff_combination_const): Use widest_int instead of double_int.
949 (aff_combination_scale): Likewise.
950 (aff_combination_add_elt): Likewise.
951 (aff_combination_constant_multiple_p): Likewise.
952 (get_inner_reference_aff): Likewise.
953 (aff_comb_cannot_overlap_p): Likewise.
954 (aff_combination_zero_p): Use wide-int interfaces.
955 * tree.c: Include tree.h.
956 (init_ttree): Use make_int_cst.
957 (tree_code_size): Removed code for INTEGER_CST case.
958 (tree_size): Add INTEGER_CST case.
959 (make_node_stat): Update comment.
960 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
961 (build_int_cst_type): Use wide-int interfaces.
962 (double_int_to_tree): Likewise.
963 (double_int_fits_to_tree_p): Delete.
964 (force_fit_type_double): Delete.
965 (force_fit_type): New.
966 (int_cst_hash_hash): Use wide-int interfaces.
967 (int_cst_hash_eq): Likewise.
968 (build_int_cst_wide): Delete.
969 (wide_int_to_tree): New.
970 (cache_integer_cst): Use wide-int interfaces.
971 (build_low_bits_mask): Likewise.
972 (cst_and_fits_in_hwi): Likewise.
973 (real_value_from_int_cst): Likewise.
974 (make_int_cst_stat): New.
975 (integer_zerop): Use wide_int interfaces.
976 (integer_onep): Likewise.
977 (integer_all_onesp): Likewise.
978 (integer_pow2p): Likewise.
979 (integer_nonzerop): Likewise.
980 (tree_log2): Likewise.
981 (tree_floor_log2): Likewise.
982 (tree_ctz): Likewise.
983 (int_size_in_bytes): Likewise.
984 (mem_ref_offset): Return an offset_int rather than a double_int.
985 (build_type_attribute_qual_variant): Use wide_int interfaces.
986 (type_hash_eq): Likewise
987 (tree_int_cst_equal): Likewise.
988 (tree_int_cst_lt): Delete.
989 (tree_int_cst_compare): Likewise.
990 (tree_fits_shwi_p): Use wide_int interfaces.
991 (tree_fits_uhwi_p): Likewise.
992 (tree_int_cst_sign_bit): Likewise.
993 (tree_int_cst_sgn): Likewise.
994 (tree_int_cst_min_precision): Take a signop rather than a bool.
995 (simple_cst_equal): Use wide_int interfaces.
996 (compare_tree_int): Likewise.
997 (iterative_hash_expr): Likewise.
998 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1000 (get_type_static_bounds): Use wide_int interfaces.
1001 (tree_int_cst_elt_check_failed): New.
1002 (build_common_tree_nodes): Reordered to set prec before filling in
1004 (int_cst_value): Check cst_and_fits_in_hwi.
1005 (widest_int_cst_value): Use wide_int interfaces.
1006 (upper_bound_in_type): Likewise.
1007 (lower_bound_in_type): Likewise.
1008 (num_ending_zeros): Likewise.
1009 (drop_tree_overflow): Likewise.
1010 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1011 (gen_conditions_for_pow_cst_base): Likewise.
1012 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1013 (group_case_labels_stmt): Use wide-int interfaces.
1014 (verify_gimple_assign_binary): Likewise.
1015 (print_loop): Likewise.
1016 * tree-chrec.c (tree_fold_binomial): Likewise.
1017 * tree-core.h (struct tree_base): Add int_length.
1018 (struct tree_int_cst): Change rep of value.
1019 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1020 (dr_may_alias_p): Likewise.
1021 (max_stmt_executions_tree): Likewise.
1022 * tree.def (INTEGER_CST): Update comment.
1023 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1024 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1025 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1026 (dequeue_and_dump): Use wide-int interfaces.
1027 * tree.h: Include wide-int.h.
1028 (NULL_TREE): Moved to earlier loc in file.
1029 (TREE_INT_CST_ELT_CHECK): New.
1030 (tree_int_cst_elt_check_failed): New.
1032 (TREE_INT_CST): Delete.
1033 (TREE_INT_CST_LOW): Use wide-int interfaces.
1034 (TREE_INT_CST_HIGH): Delete.
1035 (TREE_INT_CST_NUNITS): New.
1036 (TREE_INT_CST_EXT_NUNITS): Likewise.
1037 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1038 (TREE_INT_CST_ELT): Likewise.
1039 (INT_CST_LT): Delete.
1040 (tree_int_cst_elt_check): New (two forms).
1041 (type_code_size): Update comment.
1042 (make_int_cst_stat, make_int_cst): New.
1043 (tree_to_double_int): Delete.
1044 (double_int_fits_to_tree_p): Delete.
1045 (force_fit_type_double): Delete.
1046 (build_int_cstu): Replace with out-of-line function.
1047 (build_int_cst_wide): Delete.
1048 (tree_int_cst_lt): Define inline.
1049 (tree_int_cst_le): New.
1050 (tree_int_cst_compare): Define inline.
1051 (tree_int_cst_min_precision): Take a signop rather than a bool.
1052 (wi::int_traits <const_tree>): New.
1053 (wi::int_traits <tree>): New.
1054 (wi::extended_tree): New.
1055 (wi::int_traits <wi::extended_tree>): New.
1056 (wi::to_widest): New.
1057 (wi::to_offset): New.
1058 (wi::fits_to_tree_p): New.
1059 (wi::min_value): New.
1060 (wi::max_value): New.
1061 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1062 (copy_tree_body_r): Likewise.
1063 * tree-object-size.c (compute_object_offset): Likewise.
1064 (addr_object_size): Likewise.
1065 * tree-predcom.c: Include wide-int-print.h.
1066 (struct dref_d): Change type of offset to widest_int.
1067 (dump_dref): Call wide-int printer.
1068 (aff_combination_dr_offset): Use wide-int interfaces.
1069 (determine_offset): Take a widest_int pointer rather than a
1071 (split_data_refs_to_components): Use wide-int interfaces.
1072 (suitable_component_p): Likewise.
1073 (order_drefs): Likewise.
1074 (add_ref_to_chain): Likewise.
1075 (valid_initializer_p): Likewise.
1076 (determine_roots_comp): Likewise.
1077 * tree-pretty-print.c: Include wide-int-print.h.
1078 (dump_generic_node): Use wide-int interfaces.
1079 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1080 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1081 (move_fixed_address_to_symbol): Likewise.
1082 (move_hint_to_base): Likewise.
1083 (move_pointer_to_base): Likewise.
1084 (move_variant_to_index): Likewise.
1085 (most_expensive_mult_to_index): Likewise.
1086 (addr_to_parts): Likewise.
1087 (copy_ref_info): Likewise.
1088 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1089 (indirect_refs_may_alias_p): Likewise.
1090 (stmt_kills_ref_p_1): Likewise.
1091 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1092 * tree-ssa-ccp.c: Update comment at top of file. Include
1094 (struct prop_value_d): Change type of mask to widest_int.
1095 (extend_mask): New function.
1096 (dump_lattice_value): Use wide-int interfaces.
1097 (get_default_value): Likewise.
1098 (set_constant_value): Likewise.
1099 (set_value_varying): Likewise.
1100 (valid_lattice_transition): Likewise.
1101 (set_lattice_value): Likewise.
1102 (value_to_double_int): Delete.
1103 (value_to_wide_int): New.
1104 (get_value_from_alignment): Use wide-int interfaces.
1105 (get_value_for_expr): Likewise.
1106 (do_dbg_cnt): Likewise.
1107 (ccp_finalize): Likewise.
1108 (ccp_lattice_meet): Likewise.
1109 (bit_value_unop_1): Use widest_ints rather than double_ints.
1110 (bit_value_binop_1): Likewise.
1111 (bit_value_unop): Use wide-int interfaces.
1112 (bit_value_binop): Likewise.
1113 (bit_value_assume_aligned): Likewise.
1114 (evaluate_stmt): Likewise.
1115 (ccp_fold_stmt): Likewise.
1116 (visit_cond_stmt): Likewise.
1117 (ccp_visit_stmt): Likewise.
1118 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1119 (constant_pointer_difference): Likewise.
1120 (associate_pointerplus): Likewise.
1121 (combine_conversions): Likewise.
1122 * tree-ssa-loop.h: Include wide-int.h.
1123 (struct tree_niter_desc): Change type of max to widest_int.
1124 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1125 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1126 (remove_redundant_iv_tests): Likewise.
1127 (canonicalize_loop_induction_variables): Likewise.
1128 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1129 (constant_multiple_of): Take a widest_int pointer instead of
1130 a double_int pointer.
1131 (get_computation_aff): Use wide-int interfaces.
1132 (ptr_difference_cost): Likewise.
1133 (difference_cost): Likewise.
1134 (get_loop_invariant_expr_id): Likewise.
1135 (get_computation_cost_at): Likewise.
1136 (iv_elimination_compare_lt): Likewise.
1137 (may_eliminate_iv): Likewise.
1138 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1139 instead of double_int.
1140 (max_loop_iterations): Likewise.
1141 (max_stmt_executions): Likewise.
1142 (estimated_stmt_executions): Likewise.
1143 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1144 (split_to_var_and_offset): Use wide-int interfaces.
1145 (determine_value_range): Likewise.
1146 (bound_difference_of_offsetted_base): Likewise.
1147 (bounds_add): Take a widest_int instead of a double_int.
1148 (number_of_iterations_ne_max): Use wide-int interfaces.
1149 (number_of_iterations_ne): Likewise.
1150 (number_of_iterations_lt_to_ne): Likewise.
1151 (assert_loop_rolls_lt): Likewise.
1152 (number_of_iterations_lt): Likewise.
1153 (number_of_iterations_le): Likewise.
1154 (number_of_iterations_cond): Likewise.
1155 (number_of_iterations_exit): Likewise.
1156 (finite_loop_p): Likewise.
1157 (derive_constant_upper_bound_assign): Likewise.
1158 (derive_constant_upper_bound): Return a widest_int.
1159 (derive_constant_upper_bound_ops): Likewise.
1160 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1161 (record_estimate): Take a widest_int rather than a double_int.
1162 (record_nonwrapping_iv): Use wide-int interfaces.
1163 (double_int_cmp): Delete.
1164 (wide_int_cmp): New.
1165 (bound_index): Take a widest_int rather than a double_int.
1166 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1167 (maybe_lower_iteration_bound): Likewise.
1168 (estimate_numbers_of_iterations_loop): Likewise.
1169 (estimated_loop_iterations): Take a widest_int pointer than than
1170 a double_int pointer.
1171 (estimated_loop_iterations_int): Use wide-int interfaces.
1172 (max_loop_iterations): Take a widest_int pointer than than
1173 a double_int pointer.
1174 (max_loop_iterations_int): Use wide-int interfaces.
1175 (max_stmt_executions): Take a widest_int pointer than than
1176 a double_int pointer.
1177 (estimated_stmt_executions): Likewise.
1178 (n_of_executions_at_most): Use wide-int interfaces.
1179 (scev_probably_wraps_p): Likewise.
1180 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1182 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1184 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1185 double_ints. Adjust for trailing_wide_ints <3> representation.
1186 (set_nonzero_bits): Likewise.
1187 (get_range_info): Return wide_ints rather than double_ints.
1188 Adjust for trailing_wide_ints <3> representation.
1189 (get_nonzero_bits): Likewise.
1190 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1192 * tree-ssanames.h (struct range_info_def): Replace min, max and
1193 nonzero_bits with a trailing_wide_ints <3>.
1194 (set_range_info): Use wide_int_refs rather than double_ints.
1195 (set_nonzero_bits): Likewise.
1196 (get_range_info): Return wide_ints rather than double_ints.
1197 (get_nonzero_bits): Likewise.
1198 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1199 * tree-ssa-pre.c (phi_translate_1): Likewise.
1200 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1201 (acceptable_pow_call): Likewise.
1202 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1204 (vn_reference_fold_indirect): Likewise.
1205 (vn_reference_maybe_forwprop_address): Likewise.
1206 (valueize_refs_1): Likewise.
1207 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1208 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1209 tree_int_cst_lt and tree_int_cst_le.
1210 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1212 (streamer_alloc_tree): Likewise.
1213 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1214 (streamer_write_tree_header): Likewise.
1215 (streamer_write_integer_cst): Likewise.
1216 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1217 (build_constructors): Likewise.
1218 (array_value_type): Likewise.
1219 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1220 (vect_check_gather): Likewise.
1221 * tree-vect-generic.c (build_replicated_const): Likewise.
1222 (expand_vector_divmod): Likewise.
1223 * tree-vect-loop.c (vect_transform_loop): Likewise.
1224 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1225 (vect_do_peeling_for_alignment): Likewise.
1226 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1227 * tree-vrp.c: Include wide-int.h.
1228 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1229 (extract_range_from_assert): Use wide-int interfaces.
1230 (vrp_int_const_binop): Likewise.
1231 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1232 double_int pointers.
1233 (ranges_from_anti_range): Use wide-int interfaces.
1234 (quad_int_cmp): Delete.
1235 (quad_int_pair_sort): Likewise.
1236 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1237 (extract_range_from_unary_expr_1): Likewise.
1238 (adjust_range_with_scev): Likewise.
1239 (masked_increment): Take and return wide_ints rather than double_ints.
1240 (register_edge_assert_for_2): Use wide-int interfaces.
1241 (check_array_ref): Likewise.
1242 (search_for_addr_array): Likewise.
1243 (maybe_set_nonzero_bits): Likewise.
1244 (union_ranges): Pass an integer of the correct type instead of
1245 using integer_one_node.
1246 (intersect_ranges): Likewise.
1247 (simplify_truth_ops_using_ranges): Likewise.
1248 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1249 (range_fits_type_p): Likewise.
1250 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1252 (simplify_conversion_using_ranges): Use wide-int interfaces.
1253 (simplify_float_conversion_using_ranges): Likewise.
1254 (vrp_finalize): Likewise.
1255 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1256 (gimple_stringops_transform): Likewise.
1257 * varasm.c (decode_addr_const): Likewise.
1258 (const_hash_1): Likewise.
1259 (const_rtx_hash_1): Likewise
1260 (output_constant): Likewise.
1261 (array_size_for_constructor): Likewise.
1262 (output_constructor_regular_field): Likewise.
1263 (output_constructor_bitfield): Likewise.
1264 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1265 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1267 * gencheck.c: Define BITS_PER_UNIT.
1270 * wide-int-print.cc: New.
1271 * wide-int-print.h: New.
1273 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1275 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1277 2014-05-06 Richard Biener <rguenther@suse.de>
1279 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1280 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1281 (TODO_verify_all): Adjust.
1282 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1283 TODO_verify_stmts and TODO_verify_rtl_sharing.
1284 * bb-reorder.c: Likewise.
1285 * cfgexpand.c: Likewise.
1286 * cprop.c: Likewise.
1288 * function.c: Likewise.
1289 * fwprop.c: Likewise.
1291 * gimple-ssa-isolate-paths.c: Likewise.
1292 * gimple-ssa-strength-reduction.c: Likewise.
1293 * ipa-split.c: Likewise.
1294 * loop-init.c: Likewise.
1295 * loop-unroll.c: Likewise.
1296 * lower-subreg.c: Likewise.
1297 * modulo-sched.c: Likewise.
1298 * postreload-gcse.c: Likewise.
1299 * predict.c: Likewise.
1300 * recog.c: Likewise.
1301 * sched-rgn.c: Likewise.
1302 * store-motion.c: Likewise.
1303 * tracer.c: Likewise.
1304 * trans-mem.c: Likewise.
1305 * tree-call-cdce.c: Likewise.
1306 * tree-cfg.c: Likewise.
1307 * tree-cfgcleanup.c: Likewise.
1308 * tree-complex.c: Likewise.
1309 * tree-eh.c: Likewise.
1310 * tree-emutls.c: Likewise.
1311 * tree-if-conv.c: Likewise.
1312 * tree-into-ssa.c: Likewise.
1313 * tree-loop-distribution.c: Likewise.
1314 * tree-object-size.c: Likewise.
1315 * tree-parloops.c: Likewise.
1316 * tree-pass.h: Likewise.
1317 * tree-sra.c: Likewise.
1318 * tree-ssa-ccp.c: Likewise.
1319 * tree-ssa-copy.c: Likewise.
1320 * tree-ssa-copyrename.c: Likewise.
1321 * tree-ssa-dce.c: Likewise.
1322 * tree-ssa-dom.c: Likewise.
1323 * tree-ssa-dse.c: Likewise.
1324 * tree-ssa-forwprop.c: Likewise.
1325 * tree-ssa-ifcombine.c: Likewise.
1326 * tree-ssa-loop-ch.c: Likewise.
1327 * tree-ssa-loop-ivcanon.c: Likewise.
1328 * tree-ssa-loop.c: Likewise.
1329 * tree-ssa-math-opts.c: Likewise.
1330 * tree-ssa-phiopt.c: Likewise.
1331 * tree-ssa-phiprop.c: Likewise.
1332 * tree-ssa-pre.c: Likewise.
1333 * tree-ssa-reassoc.c: Likewise.
1334 * tree-ssa-sink.c: Likewise.
1335 * tree-ssa-strlen.c: Likewise.
1336 * tree-ssa-tail-merge.c: Likewise.
1337 * tree-ssa-uncprop.c: Likewise.
1338 * tree-switch-conversion.c: Likewise.
1339 * tree-tailcall.c: Likewise.
1340 * tree-vect-generic.c: Likewise.
1341 * tree-vectorizer.c: Likewise.
1342 * tree-vrp.c: Likewise.
1344 * var-tracking.c: Likewise.
1345 * bt-load.c: Likewise.
1346 * cfgcleanup.c: Likewise.
1347 * combine-stack-adj.c: Likewise.
1348 * combine.c: Likewise.
1349 * compare-elim.c: Likewise.
1350 * config/epiphany/resolve-sw-modes.c: Likewise.
1351 * config/i386/i386.c: Likewise.
1352 * config/mips/mips.c: Likewise.
1353 * config/s390/s390.c: Likewise.
1354 * config/sh/sh_treg_combine.cc: Likewise.
1355 * config/sparc/sparc.c: Likewise.
1358 * final.c: Likewise.
1359 * ifcvt.c: Likewise.
1360 * mode-switching.c: Likewise.
1361 * passes.c: Likewise.
1362 * postreload.c: Likewise.
1364 * reg-stack.c: Likewise.
1365 * regcprop.c: Likewise.
1366 * regrename.c: Likewise.
1369 2014-05-06 Richard Biener <rguenther@suse.de>
1372 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1373 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1375 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1378 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1380 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1381 Tom de Vries <tom@codesourcery.com>
1383 * target.def (call_fusage_contains_non_callee_clobbers): New
1385 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1387 (@node Miscellaneous Register Hooks): New node.
1388 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1389 * doc/tm.texi: Regenerate.
1391 2014-05-05 Marek Polacek <polacek@redhat.com>
1394 * opts.c (common_handle_option): Call error_at instead of warning_at.
1396 2014-05-05 Richard Biener <rguenther@suse.de>
1398 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1399 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1400 under the TODO_verify_il umbrella.
1402 2014-05-05 Richard Biener <rguenther@suse.de>
1404 * passes.c (execute_function_todo): Move TODO_verify_flow under
1405 the TODO_verify_ul umbrella.
1407 2014-05-05 Richard Biener <rguenther@suse.de>
1410 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1411 X & CST away from a CST that is the mask of a mode.
1413 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1415 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1416 int argument to enum machine_mode.
1417 (picochip_class_max_nregs): Ditto.
1418 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1419 (picochip_class_max_nregs): Ditto.
1421 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1423 * target.def: Add new target hook.
1424 * doc/tm.texi: Regenerate.
1425 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1426 * targhooks.c (default_keep_leaf_when_profiled): New function.
1428 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1429 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1431 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1433 PR tree-optimization/60363
1434 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1435 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1436 (update_destination_phis): New parameter.
1437 (create_edge_and_update_destination_phis): Ditto.
1438 (ssa_fix_duplicate_block_edges): Pass new arguments.
1439 (thread_single_edge): Ditto.
1441 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1443 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1444 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1445 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1446 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1447 Use RS6000_BTM_HARD_FLOAT.
1448 (BU_MISC_2): Likewise.
1449 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1450 RS6000_BTM_HARD_FLOAT.
1451 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1453 (rs6000_invalid_builtin): Add hard floating builtin support.
1454 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1455 hard float builtins.
1456 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1458 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1460 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1461 Add missing function* argument.
1463 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1465 * lra-constraints.c (valid_address_p): Move earlier in file.
1466 Add a constraint argument to the address_info version.
1467 (satisfies_memory_constraint_p): New function.
1468 (satisfies_address_constraint_p): Likewise.
1469 (process_alt_operands, curr_insn_transform): Use them.
1470 (process_address): Pass the constraint to valid_address_p when
1471 checking address operands.
1473 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1475 * config/mips/mips.c (mips_isa_rev): New variable.
1476 (mips_set_architecture): Set it.
1477 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1479 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1480 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1481 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1482 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1483 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1484 conditions in terms of mips_isa_rev.
1485 (mips_isa_rev): Declare.
1487 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1489 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1490 (prob_unlikely, prob_likely): Make variables const.
1492 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1494 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1496 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1498 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1500 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1502 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1503 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1504 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1506 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1507 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1509 Replace usage of ROUND_REG with sh_round_reg.
1510 Use CEIL instead of ROUND_ADVANCE.
1512 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1515 * config/sh/sh.c: Include stdlib headers before everything else.
1517 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1519 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1520 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1521 (gimplify_adjust_omp_clauses): Simd region is never
1522 directly nested in combined parallel. Instead, for linear
1523 with copyin/copyout, if in combined for simd loop, make decl
1524 firstprivate/lastprivate on OMP_FOR.
1525 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1526 expand_omp_for_static_chunk): When setting endvar, also set
1527 fd->loop.v to the same value.
1529 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1531 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1533 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1535 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1538 2014-05-02 Marek Polacek <polacek@redhat.com>
1540 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1542 2014-05-02 Kito Cheng <kito@0xlab.org>
1544 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1545 to a C expression marco.
1546 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1547 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1548 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1549 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1550 HONOR_REG_ALLOC_ORDER.
1551 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1553 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1555 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1557 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1559 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1561 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1563 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1564 (convert_scalar_cond_reduction): Likewise.
1565 (predicate_scalar_phi): Add recognition and transformation
1566 of simple conditioanl reduction to be vectorizable.
1568 2014-05-01 Marek Polacek <polacek@redhat.com>
1571 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1573 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1575 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1576 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1577 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1578 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1579 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1580 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1581 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1582 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1584 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1586 * config/arc/arc.opt (mlra): Move comment above option name
1587 to avoid mis-parsing as language options.
1589 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1591 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1592 * config/sol2.h: ... here.
1593 * config/sol2-10.h: Remove.
1595 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1596 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1597 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1598 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1599 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1600 * config/sol2.h: ... here.
1601 (SECTION_NAME_FORMAT): Don't redefine.
1602 (STARTFILE_ARCH32_SPEC): Rename to ...
1603 (STARTFILE_ARCH_SPEC): ... this.
1604 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1605 * config/sparc/sol2.h: ... here.
1606 (SECTION_NAME_FORMAT): Don't undef.
1607 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1608 (SUBTARGET_EXTRA_SPECS): Remove.
1609 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1611 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1612 (MD_STARTFILE_PREFIX): Remove.
1613 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1614 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1615 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1616 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1617 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1618 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1619 * config/i386/sol2.h: ... here.
1620 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1621 * config/i386/sol2-bi.h: Remove.
1622 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1623 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1625 * config/i386/t-sol2-64: Rename to ...
1626 * config/i386/t-sol2: ... this.
1627 * config/sparc/t-sol2-64: Rename to ...
1628 * config/sparc/t-sol2: ... this.
1630 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1631 sol2_tm_file_head, sol2_tm_file_tail.
1632 Include ${cpu_type}/sol2.h before sol2.h.
1634 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1635 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1636 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1637 Reflect i386/t-sol2-64 renaming.
1638 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1639 Reflect sparc/t-sol2-64 renaming.
1641 2014-04-30 Richard Biener <rguenther@suse.de>
1643 * passes.c (execute_function_todo): Move TODO_verify_stmts
1644 and TODO_verify_ssa under the TODO_verify_il umbrella.
1645 * tree-ssa.h (verify_ssa): Adjust prototype.
1646 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1647 we should verify SSA operands.
1648 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1649 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1650 whether we should verify whether not throwing stmts have EH info.
1651 * graphite-scop-detection.c (create_sese_edges): Adjust.
1652 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1653 * tree-eh.c (lower_try_finally_switch): Do not add the
1654 default case label twice.
1656 2014-04-30 Marek Polacek <polacek@redhat.com>
1658 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1659 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1660 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1661 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1663 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1665 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1666 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1667 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1668 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1669 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1670 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1671 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1672 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1674 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1676 * tree-cfg.c (dump_function_to_file): Dump the return type of
1677 functions, in a line to itself before the function body, mimicking
1678 the layout of a C function.
1680 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1682 PR tree-optimization/60971
1683 * tree-tailcall.c (process_assignment): Reject conversions which
1686 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1688 * calls.c (initialize_argument_information): Always treat
1689 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1690 (expand_call): Likewise.
1691 (emit_library_call_calue_1): Likewise.
1692 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1693 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1696 2014-04-29 Nick Clifton <nickc@redhat.com>
1698 * config/msp430/msp430.md (umulsidi): Fix typo.
1699 (mulhisi3): Enable even inside interrupt handlers.
1700 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1701 bigger return address pushed in large mode.
1703 2014-04-29 Nick Clifton <nickc@redhat.com>
1705 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1706 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1708 * config/m32r/m32r.c (init_reg_tables): Likewise.
1709 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1710 enum to hold the modes.
1712 2014-04-29 Richard Biener <rguenther@suse.de>
1714 * dominance.c (free_dominance_info): Add overload with
1716 (dom_info_state): Likewise.
1717 (dom_info_available_p): Likewise.
1718 * basic-block.h (free_dominance_info, dom_info_state,
1719 dom_info_available_p): Declare overloads.
1720 * passes.c (execute_function_todo): Verify that verifiers
1721 don't change dominator info state. Drop dominator info
1722 for IPA pass invocations.
1723 * cgraph.c (release_function_body): Restore asserts that
1724 dominator information is released.
1726 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1728 * doc/invoke.texi: Fix typo.
1729 * tree-vrp.c: Fix typos.
1730 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1732 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1734 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1736 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1738 * config/aarch64/aarch64-builtins.c
1739 (aarch64_types_storestruct_lane_qualifiers): New.
1740 (TYPES_STORESTRUCT_LANE): Likewise.
1741 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1742 (st3_lane): Likewise.
1743 (st4_lane): Likewise.
1744 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1745 (vec_store_lanesci_lane<mode>): Likewise.
1746 (vec_store_lanesxi_lane<mode>): Likewise.
1747 (aarch64_st2_lane<VQ:mode>): Likewise.
1748 (aarch64_st3_lane<VQ:mode>): Likewise.
1749 (aarch64_st4_lane<VQ:mode>): Likewise.
1750 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1751 * config/aarch64/arm_neon.h
1752 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1753 use new macro arguments.
1754 (__ST3_LANE_FUNC): Likewise.
1755 (__ST4_LANE_FUNC): Likewise.
1756 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1757 (V_THREE_ELEM): Likewise.
1758 (V_FOUR_ELEM): Likewise.
1760 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1762 * doc/gimple.texi: Replace the description of the now-defunct
1763 union gimple_statement_d with a diagram showing the
1764 gimple_statement_base class hierarchy and its relationships to
1765 the GSS_ and GIMPLE_ enums.
1767 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1769 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1770 * config/aarch64/aarch64.c
1771 (aarch64_cannot_change_mode_class): Weaken conditions.
1772 (aarch64_modes_tieable_p): New.
1773 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1775 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1777 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1778 (loadsync_<mode>): Change mode.
1779 (load_quadpti, store_quadpti): New.
1780 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1781 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1783 2014-04-28 Martin Jambor <mjambor@suse.cz>
1785 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1786 same alias type as the original statement.
1787 (subreplacement_assignment_data): New type.
1788 (handle_unscalarized_data_in_subtree): New type of parameter,
1789 generate new memory accesses with same alias type as the original
1791 (load_assign_lhs_subreplacements): Likewise.
1792 (sra_modify_constructor_assign): Generate new memory accesses with
1793 same alias type as the original statement.
1795 2014-04-28 Richard Biener <rguenther@suse.de>
1797 * tree-pass.h (TODO_verify_il): Define.
1798 (TODO_verify_all): Complete properly.
1799 * passes.c (execute_function_todo): Move existing loop-closed
1800 SSA verification under TODO_verify_il.
1801 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1802 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1803 Fix tree sharing issue.
1805 2014-04-28 Richard Biener <rguenther@suse.de>
1808 * builtins.def (DEF_C11_BUILTIN): Add.
1809 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1810 * coretypes.h (enum function_class): Add function_c11_misc.
1811 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1812 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1813 (call_may_clobber_ref_p_1): Likewise.
1814 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1815 (mark_all_reaching_defs_necessary_1): Likewise.
1816 (propagate_necessity): Likewise.
1817 (eliminate_unnecessary_stmts): Likewise.
1818 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1820 2014-04-28 Richard Biener <rguenther@suse.de>
1822 * tree-vrp.c (vrp_var_may_overflow): Remove.
1823 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1824 with overflow immediately bump to one before that value and
1825 let iteration figure out overflow status.
1827 2014-04-28 Richard Biener <rguenther@suse.de>
1829 * configure.ac: Do valgrind header checks unconditionally.
1830 Add --enable-valgrind-annotations.
1831 * system.h: Guard valgrind header inclusion with
1832 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1833 * alloc-pool.c (pool_alloc, pool_free): Use
1834 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1835 to guard possibly dead code.
1836 * config.in: Regenerated.
1837 * configure: Likewise.
1839 2014-04-28 Jeff Law <law@redhat.com>
1841 PR tree-optimization/60902
1842 * tree-ssa-threadedge.c
1843 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1844 over real defs when invalidating outputs from statements that do not
1845 produce useful outputs for threading.
1847 2014-04-28 Richard Biener <rguenther@suse.de>
1849 PR tree-optimization/60979
1850 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1851 SCOPs that end in a block with a successor with abnormal
1854 2014-04-28 Richard Biener <rguenther@suse.de>
1856 * tree-pass.h (execute_pass_list): Adjust prototype.
1857 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1858 (do_per_function): Change callback signature, push all actual
1859 work to the callbals.
1860 (do_per_function_toporder): Likewise.
1861 (execute_function_dump): Adjust.
1862 (execute_function_todo): Likewise.
1863 (clear_last_verified): Likewise.
1864 (verify_curr_properties): Likewise.
1865 (update_properties_after_pass): Likewise.
1866 (execute_pass_list_1): Split out from ...
1867 (execute_pass_list): ... here. Adjust.
1868 (execute_ipa_pass_list): Likewise.
1869 * cgraphunit.c (cgraph_add_new_function): Adjust.
1870 (analyze_function): Likewise.
1871 (expand_function): Likewise.
1872 * cgraph.c (release_function_body): Free dominance info
1873 here instead of asserting it was magically freed elsewhere.
1875 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1877 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1878 * configure: Regenerate.
1879 * config/sparc/sparc.opt (muser-mode): New option.
1880 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1882 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1883 * doc/invoke.texi (SPARC options): Document -muser-mode.
1885 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1887 * cselib.c (find_slot_memmode): Delete.
1888 (cselib_hasher): Change compare_type to a struct.
1889 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1891 (preserve_constants_and_equivs): Adjust for new compare_type.
1892 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1893 (wrap_constant): Delete.
1894 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1896 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1898 * doc/install.texi (Building with profile feedback): Remove
1901 2014-04-26 Tom de Vries <tom@codesourcery.com>
1903 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1906 2014-04-25 Cary Coutant <ccoutant@google.com>
1909 * dwarf2out.c (should_move_die_to_comdat): A type definition
1910 can contain a subprogram definition, but don't move it to a
1912 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1913 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1915 (clone_tree_hash): Rename to...
1916 (clone_tree_partial): ...this; change callers. Copy
1917 DW_TAG_subprogram DIEs as declarations.
1918 (copy_decls_walk): Don't copy children of a declaration into a
1921 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1924 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1927 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1929 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1930 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1932 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1935 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1937 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1939 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1941 PR tree-optimization/60930
1942 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1943 creating a multiply candidate by folding two constant
1944 multiplicands when the result overflows.
1946 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1948 PR tree-optimization/60960
1949 * tree-vect-generic.c (expand_vector_operation): Only call
1950 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1952 2014-04-25 Tom de Vries <tom@codesourcery.com>
1954 * expr.c (clobber_reg_mode): New function.
1955 * expr.h (clobber_reg): New function.
1957 2014-04-25 Tom de Vries <tom@codesourcery.com>
1959 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1962 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1963 Tom de Vries <tom@codesourcery.com>
1965 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1967 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1968 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1969 new argument to find_all_hard_reg_sets call.
1971 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1973 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1974 Use HOST_WIDE_INT_C for mask literal.
1975 (aarch_rev16_shleft_mask_imm_p): Likewise.
1977 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1980 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1982 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1984 PR preprocessor/56540
1985 * config/i386/i386-c.c (ix86_target_macros): Define
1986 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1988 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1990 * configure.ac (tga_func): Remove.
1991 (LIB_TLS_SPEC): Remove.
1992 * configure: Regenerate.
1993 * config.in: Regenerate.
1994 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1996 2014-04-25 Richard Biener <rguenther@suse.de>
1999 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2000 call stmt use/clobber sets during stmt walk instead of
2001 walking the possibly incomplete set of caller edges.
2003 2014-04-25 Richard Biener <rguenther@suse.de>
2006 * passes.c (apply_ipa_transforms): Inline into only caller ...
2007 (execute_one_pass): ... here. Properly bring in function
2008 bodies for nodes we want to apply IPA transforms to.
2010 2014-04-24 Cong Hou <congh@google.com>
2012 PR tree-optimization/60896
2013 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2014 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2015 (vect_mark_pattern_stmts): Set the def type of all statements in
2016 PATTERN_DEF_SEQ as vect_internal_def.
2018 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2020 * doc/extend.texi (PowerPC Built-in Functions): Document new
2021 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2022 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2024 * config/rs6000/predicates.md (const_0_to_3_operand): New
2025 predicate to match 0..3 integer constants.
2027 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2028 to support adding miscellaneous builtin functions.
2029 (BU_DFP_MISC_2): Likewise.
2030 (BU_P7_MISC_1): Likewise.
2031 (BU_P7_MISC_2): Likewise.
2032 (BU_P8V_MISC_3): Likewise.
2033 (BU_MISC_1): Likewise.
2034 (BU_MISC_2): Likewise.
2035 (DIVWE): Add extended divide builtin functions.
2038 (DIVWEUO): Likewise.
2042 (DIVDEUO): Likewise.
2043 (DXEX): Add decimal floating-point builtin functions.
2046 (DDEDPDQ): Likewise.
2048 (DENBCDQ): Likewise.
2055 (CDTBCD): Add new BCD builtin functions.
2059 (BCDADD_LT): Likewise.
2060 (BCDADD_EQ): Likewise.
2061 (BCDADD_GT): Likewise.
2062 (BCDADD_OV): Likewise.
2064 (BCDSUB_LT): Likewise.
2065 (BCDSUB_EQ): Likewise.
2066 (BCDSUB_GT): Likewise.
2067 (BCDSUB_OV): Likewise.
2068 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2069 (UNPACK_TD): Likewise.
2070 (PACK_TF): Likewise.
2071 (UNPACK_TF): Likewise.
2072 (UNPACK_TF_0): Likewise.
2073 (UNPACK_TF_1): Likewise.
2074 (PACK_V1TI): Likewise.
2075 (UNPACK_V1TI): Likewise.
2077 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2078 support for decimal floating point builtin functions.
2079 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2080 functions that take constant arguments.
2081 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2082 (rs6000_init_builtins): Setup long double, _Decimal64, and
2083 _Decimal128 types for new builtin functions.
2084 (builtin_function_type): Set the unsigned flags appropriately for
2085 the new builtin functions.
2086 (rs6000_opt_masks): Add support for decimal floating point builtin
2089 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2090 floating point builtin functions.
2091 (RS6000_BTM_COMMON): Likewise.
2092 (RS6000_BTI_long_double): Likewise.
2093 (RS6000_BTI_dfloat64): Likewise.
2094 (RS6000_BTI_dfloat128): Likewise.
2095 (long_double_type_internal_node): Likewise.
2096 (dfloat64_type_internal_node): Likewise.
2097 (dfloat128_type_internal_node): Likewise.
2099 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2100 2.07 bcd arithmetic instructions.
2101 (UNSPEC_BCDSUB): Likewise.
2102 (UNSPEC_BCD_OVERFLOW): Likewise.
2103 (UNSPEC_BCD_ADD_SUB): Likewise.
2104 (bcd_add_sub): Likewise.
2105 (BCD_TEST): Likewise.
2106 (bcd<bcd_add_sub>): Likewise.
2107 (bcd<bcd_add_sub>_test): Likewise.
2108 (bcd<bcd_add_sub>_test2): Likewise.
2109 (bcd<bcd_add_sub>_<code>): Likewise.
2110 (peephole2 for combined bcd ops): Likewise.
2112 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2113 decimal floating point builtin functions.
2114 (UNSPEC_DENBCD): Likewise.
2115 (UNSPEC_DXEX): Likewise.
2116 (UNSPEC_DIEX): Likewise.
2117 (UNSPEC_DSCLI): Likewise.
2118 (UNSPEC_DSCRI): Likewise.
2119 (D64_D128): Likewise.
2120 (dfp_suffix): Likewise.
2121 (dfp_ddedpd_<mode>): Likewise.
2122 (dfp_denbcd_<mode>): Likewise.
2123 (dfp_dxex_<mode>): Likewise.
2124 (dfp_diex_<mode>): Likewise.
2125 (dfp_dscli_<mode>): Likewise.
2126 (dfp_dscri_<mode>): Likewise.
2128 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2130 (UNSPEC_CDTBCD): Likewise.
2131 (UNSPEC_CBCDTD): Likewise.
2132 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2133 (UNSPEC_DIVEO): Likewise.
2134 (UNSPEC_DIVEU): Likewise.
2135 (UNSPEC_DIVEUO): Likewise.
2136 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2137 pack/unpack 128-bit types.
2138 (UNSPEC_PACK_128BIT): Likewise.
2139 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2140 (udiv<mode>3): Use idiv_ldiv mode attribute.
2141 (div<mode>3): Likewise.
2142 (addg6s): Add new BCD builtin functions.
2145 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2146 (div_extend): Likewise.
2147 (div<div_extend>_<mode>"): Likewise.
2148 (FP128_64): Add support for new builtin functions to pack/unpack
2150 (unpack<mode>): Likewise.
2151 (unpacktf_0): Likewise.
2152 (unpacktf_1): Likewise.
2153 (unpack<mode>_dm): Likewise.
2154 (unpack<mode>_nodm): Likewise.
2155 (pack<mode>): Likewise.
2156 (unpackv1ti): Likewise.
2157 (packv1ti): Likewise.
2159 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2161 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2164 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2166 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2167 * gimplify.c (omp_is_private): Change last argument's type to int.
2168 Only diagnose lastprivate if the simd argument is 1, only diagnose
2169 linear if the simd argument is 2.
2170 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2171 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2172 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2173 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2174 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2175 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2176 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2177 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2178 * tree-nested.c (convert_nonlocal_omp_clauses,
2179 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2181 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2184 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2187 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2189 * flag-types.h (enum ivar_visibility): Add.
2191 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2193 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2194 function * argument.
2196 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2198 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2200 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2201 Tom de Vries <tom@codesourcery.com>
2203 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2204 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2206 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2207 * emit-rtl.c (try_split): Same.
2209 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2210 Tom de Vries <tom@codesourcery.com>
2212 * common.opt (fuse-caller-save): New option.
2214 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2216 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2217 elements for big-endian.
2219 2014-04-24 Richard Biener <rguenther@suse.de>
2221 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2222 during TER and instead use the sepops interface for expanding
2223 non-GIMPLE_SINGLE_RHS.
2225 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2227 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2228 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2230 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2232 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2233 assembler 64-bit option.
2234 * configure: Regenerate.
2236 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2238 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2239 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2240 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2241 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2242 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2244 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2246 * config/aarch64/aarch64-builtins.c
2247 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2248 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2249 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2250 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2252 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2253 (Vrevsuff): New mode attribute.
2255 2014-04-24 Terry Guo <terry.guo@arm.com>
2257 * config/arm/arm.h (machine_function): Define variable
2258 after_arm_reorg here.
2259 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2260 (arm_split_constant): Update the way to access variable
2263 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2265 2014-04-23 Tom de Vries <tom@codesourcery.com>
2267 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2269 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2271 * is-a.h: Update comments to reflect the following changes to the
2272 "pointerness" of the API, making the template parameter match the
2273 return type, allowing use of is-a.h with typedefs of pointers.
2274 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2275 that the return type matches the parameter to the is_a_helper.
2277 (dyn_cast): Likewise.
2279 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2280 pointer from the is-a.h API.
2282 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2283 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2285 (is_a_helper <varpool_node>::test): Likewise, convert to...
2286 (is_a_helper <varpool_node *>::test): ...this.
2288 (varpool_first_variable): Update for removal of implicit pointer
2289 from the is-a.h API.
2290 (varpool_next_variable): Likewise.
2291 (varpool_first_static_initializer): Likewise.
2292 (varpool_next_static_initializer): Likewise.
2293 (varpool_first_defined_variable): Likewise.
2294 (varpool_next_defined_variable): Likewise.
2295 (cgraph_first_defined_function): Likewise.
2296 (cgraph_next_defined_function): Likewise.
2297 (cgraph_first_function): Likewise.
2298 (cgraph_next_function): Likewise.
2299 (cgraph_first_function_with_gimple_body): Likewise.
2300 (cgraph_next_function_with_gimple_body): Likewise.
2301 (cgraph_alias_target): Likewise.
2302 (varpool_alias_target): Likewise.
2303 (cgraph_function_or_thunk_node): Likewise.
2304 (varpool_variable_node): Likewise.
2305 (symtab_real_symbol_p): Likewise.
2306 * cgraphunit.c (referred_to_p): Likewise.
2307 (analyze_functions): Likewise.
2308 (handle_alias_pairs): Likewise.
2309 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2310 * gimple-ssa.h (gimple_vuse_op): Likewise.
2311 (gimple_vdef_op): Likewise.
2312 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2313 * gimple.c (gimple_build_asm_1): Likewise.
2314 (gimple_build_try): Likewise.
2315 (gimple_build_resx): Likewise.
2316 (gimple_build_eh_dispatch): Likewise.
2317 (gimple_build_omp_for): Likewise.
2318 (gimple_omp_for_set_clauses): Likewise.
2320 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2321 (is_a_helper <gimple_statement_asm *>::test): ...this.
2322 (is_a_helper <gimple_statement_bind>::test): Convert to...
2323 (is_a_helper <gimple_statement_bind *>::test): ...this.
2324 (is_a_helper <gimple_statement_call>::test): Convert to...
2325 (is_a_helper <gimple_statement_call *>::test): ...this.
2326 (is_a_helper <gimple_statement_catch>::test): Convert to...
2327 (is_a_helper <gimple_statement_catch *>::test): ...this.
2328 (is_a_helper <gimple_statement_resx>::test): Convert to...
2329 (is_a_helper <gimple_statement_resx *>::test): ...this.
2330 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2331 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2332 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2333 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2334 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2335 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2336 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2337 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2338 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2339 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2340 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2341 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2342 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2343 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2344 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2345 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2346 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2347 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2348 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2349 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2350 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2351 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2352 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2353 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2354 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2355 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2356 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2357 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2358 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2359 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2360 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2361 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2362 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2363 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2364 (is_a_helper <gimple_statement_phi>::test): Convert to...
2365 (is_a_helper <gimple_statement_phi *>::test): ...this.
2366 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2367 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2368 (is_a_helper <gimple_statement_try>::test): Convert to...
2369 (is_a_helper <gimple_statement_try *>::test): ...this.
2370 (is_a_helper <gimple_statement_wce>::test): Convert to...
2371 (is_a_helper <gimple_statement_wce *>::test): ...this.
2372 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2373 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2374 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2375 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2376 (is_a_helper <const gimple_statement_call>::test): Convert to...
2377 (is_a_helper <const gimple_statement_call *>::test): ...this.
2378 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2379 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2380 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2381 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2382 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2383 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2384 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2385 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2386 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2388 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2390 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2392 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2394 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2395 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2396 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2398 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2399 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2401 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2402 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2403 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2404 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2405 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2406 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2408 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2409 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2410 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2411 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2413 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2414 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2415 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2416 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2417 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2418 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2419 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2420 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2421 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2422 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2423 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2424 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2425 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2426 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2427 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2428 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2430 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2432 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2433 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2435 (gimple_use_ops): Update for removal of implicit pointer from the
2437 (gimple_set_use_ops): Likewise.
2438 (gimple_vuse): Likewise.
2439 (gimple_vdef): Likewise.
2440 (gimple_vuse_ptr): Likewise.
2441 (gimple_vdef_ptr): Likewise.
2442 (gimple_set_vuse): Likewise.
2443 (gimple_set_vdef): Likewise.
2444 (gimple_omp_return_set_lhs): Likewise.
2445 (gimple_omp_return_lhs): Likewise.
2446 (gimple_omp_return_lhs_ptr): Likewise.
2447 (gimple_call_fntype): Likewise.
2448 (gimple_call_set_fntype): Likewise.
2449 (gimple_call_set_internal_fn): Likewise.
2450 (gimple_call_use_set): Likewise.
2451 (gimple_call_clobber_set): Likewise.
2452 (gimple_bind_vars): Likewise.
2453 (gimple_bind_set_vars): Likewise.
2454 (gimple_bind_body_ptr): Likewise.
2455 (gimple_bind_set_body): Likewise.
2456 (gimple_bind_add_stmt): Likewise.
2457 (gimple_bind_block): Likewise.
2458 (gimple_bind_set_block): Likewise.
2459 (gimple_asm_ninputs): Likewise.
2460 (gimple_asm_noutputs): Likewise.
2461 (gimple_asm_nclobbers): Likewise.
2462 (gimple_asm_nlabels): Likewise.
2463 (gimple_asm_input_op): Likewise.
2464 (gimple_asm_input_op_ptr): Likewise.
2465 (gimple_asm_output_op): Likewise.
2466 (gimple_asm_output_op_ptr): Likewise.
2467 (gimple_asm_set_output_op): Likewise.
2468 (gimple_asm_clobber_op): Likewise.
2469 (gimple_asm_set_clobber_op): Likewise.
2470 (gimple_asm_label_op): Likewise.
2471 (gimple_asm_set_label_op): Likewise.
2472 (gimple_asm_string): Likewise.
2473 (gimple_catch_types): Likewise.
2474 (gimple_catch_types_ptr): Likewise.
2475 (gimple_catch_handler_ptr): Likewise.
2476 (gimple_catch_set_types): Likewise.
2477 (gimple_catch_set_handler): Likewise.
2478 (gimple_eh_filter_types): Likewise.
2479 (gimple_eh_filter_types_ptr): Likewise.
2480 (gimple_eh_filter_failure_ptr): Likewise.
2481 (gimple_eh_filter_set_types): Likewise.
2482 (gimple_eh_filter_set_failure): Likewise.
2483 (gimple_eh_must_not_throw_fndecl): Likewise.
2484 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2485 (gimple_eh_else_n_body_ptr): Likewise.
2486 (gimple_eh_else_e_body_ptr): Likewise.
2487 (gimple_eh_else_set_n_body): Likewise.
2488 (gimple_eh_else_set_e_body): Likewise.
2489 (gimple_try_eval_ptr): Likewise.
2490 (gimple_try_cleanup_ptr): Likewise.
2491 (gimple_try_set_eval): Likewise.
2492 (gimple_try_set_cleanup): Likewise.
2493 (gimple_wce_cleanup_ptr): Likewise.
2494 (gimple_wce_set_cleanup): Likewise.
2495 (gimple_phi_capacity): Likewise.
2496 (gimple_phi_num_args): Likewise.
2497 (gimple_phi_result): Likewise.
2498 (gimple_phi_result_ptr): Likewise.
2499 (gimple_phi_set_result): Likewise.
2500 (gimple_phi_arg): Likewise.
2501 (gimple_phi_set_arg): Likewise.
2502 (gimple_resx_region): Likewise.
2503 (gimple_resx_set_region): Likewise.
2504 (gimple_eh_dispatch_region): Likewise.
2505 (gimple_eh_dispatch_set_region): Likewise.
2506 (gimple_omp_critical_name): Likewise.
2507 (gimple_omp_critical_name_ptr): Likewise.
2508 (gimple_omp_critical_set_name): Likewise.
2509 (gimple_omp_for_clauses): Likewise.
2510 (gimple_omp_for_clauses_ptr): Likewise.
2511 (gimple_omp_for_set_clauses): Likewise.
2512 (gimple_omp_for_collapse): Likewise.
2513 (gimple_omp_for_index): Likewise.
2514 (gimple_omp_for_index_ptr): Likewise.
2515 (gimple_omp_for_set_index): Likewise.
2516 (gimple_omp_for_initial): Likewise.
2517 (gimple_omp_for_initial_ptr): Likewise.
2518 (gimple_omp_for_set_initial): Likewise.
2519 (gimple_omp_for_final): Likewise.
2520 (gimple_omp_for_final_ptr): Likewise.
2521 (gimple_omp_for_set_final): Likewise.
2522 (gimple_omp_for_incr): Likewise.
2523 (gimple_omp_for_incr_ptr): Likewise.
2524 (gimple_omp_for_set_incr): Likewise.
2525 (gimple_omp_for_pre_body_ptr): Likewise.
2526 (gimple_omp_for_set_pre_body): Likewise.
2527 (gimple_omp_parallel_clauses): Likewise.
2528 (gimple_omp_parallel_clauses_ptr): Likewise.
2529 (gimple_omp_parallel_set_clauses): Likewise.
2530 (gimple_omp_parallel_child_fn): Likewise.
2531 (gimple_omp_parallel_child_fn_ptr): Likewise.
2532 (gimple_omp_parallel_set_child_fn): Likewise.
2533 (gimple_omp_parallel_data_arg): Likewise.
2534 (gimple_omp_parallel_data_arg_ptr): Likewise.
2535 (gimple_omp_parallel_set_data_arg): Likewise.
2536 (gimple_omp_task_clauses): Likewise.
2537 (gimple_omp_task_clauses_ptr): Likewise.
2538 (gimple_omp_task_set_clauses): Likewise.
2539 (gimple_omp_task_child_fn): Likewise.
2540 (gimple_omp_task_child_fn_ptr): Likewise.
2541 (gimple_omp_task_set_child_fn): Likewise.
2542 (gimple_omp_task_data_arg): Likewise.
2543 (gimple_omp_task_data_arg_ptr): Likewise.
2544 (gimple_omp_task_set_data_arg): Likewise.
2545 (gimple_omp_taskreg_clauses): Likewise.
2546 (gimple_omp_taskreg_clauses_ptr): Likewise.
2547 (gimple_omp_taskreg_set_clauses): Likewise.
2548 (gimple_omp_taskreg_child_fn): Likewise.
2549 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2550 (gimple_omp_taskreg_set_child_fn): Likewise.
2551 (gimple_omp_taskreg_data_arg): Likewise.
2552 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2553 (gimple_omp_taskreg_set_data_arg): Likewise.
2554 (gimple_omp_task_copy_fn): Likewise.
2555 (gimple_omp_task_copy_fn_ptr): Likewise.
2556 (gimple_omp_task_set_copy_fn): Likewise.
2557 (gimple_omp_task_arg_size): Likewise.
2558 (gimple_omp_task_arg_size_ptr): Likewise.
2559 (gimple_omp_task_set_arg_size): Likewise.
2560 (gimple_omp_task_arg_align): Likewise.
2561 (gimple_omp_task_arg_align_ptr): Likewise.
2562 (gimple_omp_task_set_arg_align): Likewise.
2563 (gimple_omp_single_clauses): Likewise.
2564 (gimple_omp_single_clauses_ptr): Likewise.
2565 (gimple_omp_single_set_clauses): Likewise.
2566 (gimple_omp_target_clauses): Likewise.
2567 (gimple_omp_target_clauses_ptr): Likewise.
2568 (gimple_omp_target_set_clauses): Likewise.
2569 (gimple_omp_target_child_fn): Likewise.
2570 (gimple_omp_target_child_fn_ptr): Likewise.
2571 (gimple_omp_target_set_child_fn): Likewise.
2572 (gimple_omp_target_data_arg): Likewise.
2573 (gimple_omp_target_data_arg_ptr): Likewise.
2574 (gimple_omp_target_set_data_arg): Likewise.
2575 (gimple_omp_teams_clauses): Likewise.
2576 (gimple_omp_teams_clauses_ptr): Likewise.
2577 (gimple_omp_teams_set_clauses): Likewise.
2578 (gimple_omp_sections_clauses): Likewise.
2579 (gimple_omp_sections_clauses_ptr): Likewise.
2580 (gimple_omp_sections_set_clauses): Likewise.
2581 (gimple_omp_sections_control): Likewise.
2582 (gimple_omp_sections_control_ptr): Likewise.
2583 (gimple_omp_sections_set_control): Likewise.
2584 (gimple_omp_for_set_cond): Likewise.
2585 (gimple_omp_for_cond): Likewise.
2586 (gimple_omp_atomic_store_set_val): Likewise.
2587 (gimple_omp_atomic_store_val): Likewise.
2588 (gimple_omp_atomic_store_val_ptr): Likewise.
2589 (gimple_omp_atomic_load_set_lhs): Likewise.
2590 (gimple_omp_atomic_load_lhs): Likewise.
2591 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2592 (gimple_omp_atomic_load_set_rhs): Likewise.
2593 (gimple_omp_atomic_load_rhs): Likewise.
2594 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2595 (gimple_omp_continue_control_def): Likewise.
2596 (gimple_omp_continue_control_def_ptr): Likewise.
2597 (gimple_omp_continue_set_control_def): Likewise.
2598 (gimple_omp_continue_control_use): Likewise.
2599 (gimple_omp_continue_control_use_ptr): Likewise.
2600 (gimple_omp_continue_set_control_use): Likewise.
2601 (gimple_transaction_body_ptr): Likewise.
2602 (gimple_transaction_label): Likewise.
2603 (gimple_transaction_label_ptr): Likewise.
2604 (gimple_transaction_set_body): Likewise.
2605 (gimple_transaction_set_label): Likewise.
2607 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2608 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2609 * ipa-ref.c (ipa_record_reference): Likewise.
2610 * ipa-reference.c (analyze_function): Likewise.
2611 (ipa_reference_write_optimization_summary): Likewise.
2612 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2613 (address_taken_from_non_vtable_p): Likewise.
2614 (comdat_can_be_unshared_p_1): Likewise.
2615 * lto-cgraph.c (lto_output_ref): Likewise.
2616 (add_references): Likewise.
2617 (compute_ltrans_boundary): Likewise.
2618 (output_symtab): Likewise.
2619 (input_ref): Likewise.
2620 (input_cgraph_1): Likewise.
2621 (output_cgraph_opt_summary): Likewise.
2622 * lto-streamer-out.c (lto_output): Likewise.
2623 (output_symbol_p): Likewise.
2624 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2625 (lsei_start_function_in_partition): Likewise.
2626 (lsei_next_variable_in_partition): Likewise.
2627 (lsei_start_variable_in_partition): Likewise.
2628 * symtab.c (insert_to_assembler_name_hash): Likewise.
2629 (unlink_from_assembler_name_hash): Likewise.
2630 (symtab_unregister_node): Likewise.
2631 (symtab_remove_node): Likewise.
2632 (dump_symtab_node): Likewise.
2633 (verify_symtab_base): Likewise.
2634 (verify_symtab_node): Likewise.
2635 (symtab_make_decl_local): Likewise.
2636 (symtab_alias_ultimate_target): Likewise.
2637 (symtab_resolve_alias): Likewise.
2638 (symtab_get_symbol_partitioning_class): Likewise.
2639 * tree-phinodes.c (allocate_phi_node): Likewise.
2640 (reserve_phi_args_for_new_edge): Likewise.
2641 (remove_phi_args): Likewise.
2642 * varpool.c (varpool_node_for_asm): Likewise.
2643 (varpool_remove_unreferenced_decls): Likewise.
2645 2014-04-23 Jeff Law <law@redhat.com>
2647 PR tree-optimization/60902
2648 * tree-ssa-threadedge.c
2649 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2650 invalidate outputs from statements that do not produce useful
2651 outputs for threading.
2653 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2655 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2656 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2657 machine descriptions for Stack Smashing Protector.
2659 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2661 * aarch64.md (<optab>_rol<mode>3): New pattern.
2662 (<optab>_rolsi3_uxtw): Likewise.
2663 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2665 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2667 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2668 (arm_cortex_a12_tune): Likewise.
2670 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2672 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2674 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2676 * config/arm/arm.md (arm_rev16si2): New pattern.
2677 (arm_rev16si2_alt): Likewise.
2678 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2680 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2682 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2683 (rev16<mode>2_alt): Likewise.
2684 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2685 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2686 (aarch_rev16_shleft_mask_imm_p): Likewise.
2687 (aarch_rev16_p_1): Likewise.
2688 (aarch_rev16_p): Likewise.
2689 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2690 (aarch_rev16_shright_mask_imm_p): Likewise.
2691 (aarch_rev16_shleft_mask_imm_p): Likewise.
2693 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2695 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2696 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2698 (cortex_a53_extra_costs): Likewise.
2699 (cortex_a57_extra_costs): Likewise.
2700 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2701 (cortexa7_extra_costs): Likewise.
2702 (cortexa8_extra_costs): Likewise.
2703 (cortexa12_extra_costs): Likewise.
2704 (cortexa15_extra_costs): Likewise.
2705 (v7m_extra_costs): Likewise.
2706 (arm_new_rtx_costs): Handle BSWAP.
2708 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2710 * config/arm/arm.c (cortexa8_extra_costs): New table.
2711 (arm_cortex_a8_tune): New tuning struct.
2712 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2714 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2716 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2718 2014-04-23 Richard Biener <rguenther@suse.de>
2720 * Makefile.in (OBJS): Remove loop-unswitch.o.
2721 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2722 * passes.def (pass_rtl_unswitch): Likewise.
2723 * loop-init.c (gate_rtl_unswitch): Likewise.
2724 (rtl_unswitch): Likewise.
2725 (pass_data_rtl_unswitch): Likewise.
2726 (pass_rtl_unswitch): Likewise.
2727 (make_pass_rtl_unswitch): Likewise.
2728 * rtl.h (reversed_condition): Likewise.
2729 (compare_and_jump_seq): Likewise.
2730 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2732 * loop-unroll.c (compare_and_jump_seq): Likewise.
2734 2014-04-23 Richard Biener <rguenther@suse.de>
2736 PR tree-optimization/60903
2737 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2738 commented code block.
2739 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2740 loop flags to newly created BBs and edges.
2742 2014-04-23 Nick Clifton <nickc@redhat.com>
2744 * config/msp430/msp430.c (msp430_handle_option): Move function
2746 (msp430_option_override): Simplify mcu and mcpu option handling.
2747 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2748 support for -mhwmult command line option.
2749 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2750 -mhwmult command line option.
2751 (msp430_hwmult_enabled): Delete.
2752 (msp43o_output_labelref): Add support for -mhwmult command line option.
2753 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2754 (umulsidi3): Likewise.
2755 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2756 (mcpu, mlarge, msmall): Likewise.
2757 (mhwmult): New option.
2758 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2760 (msp430_is_f5_mcu): Remove prototype.
2761 (msp430_use_f5_series_hwmult): Add prototype.
2762 * config/msp430/msp430-opts.h: New file.
2763 * common/config/msp430: New directory.
2764 * common/config/msp430/msp430-common.c: New file.
2765 * config.gcc (msp430): Remove target_has_targetm_common.
2766 * doc/invoke.texi: Document -mhwmult command line option.
2768 2014-04-23 Nick Clifton <nickc@redhat.com>
2770 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2771 default-manifest.o if it can be found in the search path.
2772 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2774 2014-04-23 Terry Guo <terry.guo@arm.com>
2776 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2778 2014-04-23 Richard Biener <rguenther@suse.de>
2781 * tree-inline.c (declare_return_variable): Use mark_addressable.
2783 2014-04-23 Richard Biener <rguenther@suse.de>
2786 * loop-init.c (loop_optimizer_init): Make sure to apply
2787 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2789 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2792 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2794 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2795 if flag_sanitize_undefined_trap_on_error.
2796 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2797 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2798 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2799 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2800 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2801 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2802 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2803 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2804 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2805 * ubsan.c (ubsan_instrument_unreachable): Return
2806 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2807 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2808 if flag_sanitize_undefined_trap_on_error and
2809 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2810 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2811 instrument_bool_enum_load): Emit __builtin_trap () if
2812 flag_sanitize_undefined_trap_on_error and
2813 __builtin_handle_*_abort () if !flag_sanitize_recover.
2814 * doc/invoke.texi (-fsanitize-recover,
2815 -fsanitize-undefined-trap-on-error): Document.
2817 2014-04-22 Christian Bruel <christian.bruel@st.com>
2819 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2820 Force immediates to SImode.
2822 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2824 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2826 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2827 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2828 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2829 (nios2_fpu_insn): Add entry for round.
2830 (N2FPU_NO_ERRNO_P): Define.
2831 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2833 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2835 2014-04-22 Richard Henderson <rth@redhat.com>
2837 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2838 (add<GPI>3_compare0): Remove leading * from name.
2839 (add<GPI>3_carryin): Likewise.
2840 (sub<GPI>3_compare0): Likewise.
2841 (sub<GPI>3_carryin): Likewise.
2842 (<su_optab>mulditi3): New expander.
2843 (multi3): New expander.
2844 (madd<GPI>): Remove leading * from name.
2846 2014-04-22 Martin Jambor <mjambor@suse.cz>
2848 * cgraphclones.c (cgraph_function_versioning): Copy
2849 ipa_transforms_to_apply instead of asserting it is empty.
2851 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2854 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2855 on count_exp to get mode.
2857 2014-04-22 Andrew Pinski <apinski@cavium.com>
2859 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2860 Handle TLS for ILP32.
2861 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2862 (tlsie_small_<mode>): this and handle PTR.
2863 (tlsie_small_sidi): New pattern.
2864 (tlsle_small): Change to an expand to handle ILP32.
2865 (tlsle_small_<mode>): New pattern.
2866 (tlsdesc_small): Rename to ...
2867 (tlsdesc_small_<mode>): this and handle PTR.
2869 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2871 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2873 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2875 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2876 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2877 (aarch64_types_signed_poly_qualifiers): Likewise.
2878 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2879 (aarch64_types_poly_signed_qualifiers): Likewise.
2880 (TYPES_REINTERP_SS): Type macro added.
2881 (TYPES_REINTERP_SU): Likewise.
2882 (TYPES_REINTERP_SP): Likewise.
2883 (TYPES_REINTERP_US): Likewise.
2884 (TYPES_REINTERP_PS): Likewise.
2885 (aarch64_fold_builtin): New expression folding added.
2886 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2887 Declarations removed.
2888 (REINTERP_SS): Declarations added.
2889 (REINTERP_US): Likewise.
2890 (REINTERP_PS): Likewise.
2891 (REINTERP_SU): Likewise.
2892 (REINTERP_SP): Likewise.
2893 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2894 (vreinterpretq_p8_f64): Likewise.
2895 (vreinterpret_p16_f64): Likewise.
2896 (vreinterpretq_p16_f64): Likewise.
2897 (vreinterpret_f32_f64): Likewise.
2898 (vreinterpretq_f32_f64): Likewise.
2899 (vreinterpret_f64_f32): Likewise.
2900 (vreinterpret_f64_p8): Likewise.
2901 (vreinterpret_f64_p16): Likewise.
2902 (vreinterpret_f64_s8): Likewise.
2903 (vreinterpret_f64_s16): Likewise.
2904 (vreinterpret_f64_s32): Likewise.
2905 (vreinterpret_f64_s64): Likewise.
2906 (vreinterpret_f64_u8): Likewise.
2907 (vreinterpret_f64_u16): Likewise.
2908 (vreinterpret_f64_u32): Likewise.
2909 (vreinterpret_f64_u64): Likewise.
2910 (vreinterpretq_f64_f32): Likewise.
2911 (vreinterpretq_f64_p8): Likewise.
2912 (vreinterpretq_f64_p16): Likewise.
2913 (vreinterpretq_f64_s8): Likewise.
2914 (vreinterpretq_f64_s16): Likewise.
2915 (vreinterpretq_f64_s32): Likewise.
2916 (vreinterpretq_f64_s64): Likewise.
2917 (vreinterpretq_f64_u8): Likewise.
2918 (vreinterpretq_f64_u16): Likewise.
2919 (vreinterpretq_f64_u32): Likewise.
2920 (vreinterpretq_f64_u64): Likewise.
2921 (vreinterpret_s64_f64): Likewise.
2922 (vreinterpretq_s64_f64): Likewise.
2923 (vreinterpret_u64_f64): Likewise.
2924 (vreinterpretq_u64_f64): Likewise.
2925 (vreinterpret_s8_f64): Likewise.
2926 (vreinterpretq_s8_f64): Likewise.
2927 (vreinterpret_s16_f64): Likewise.
2928 (vreinterpretq_s16_f64): Likewise.
2929 (vreinterpret_s32_f64): Likewise.
2930 (vreinterpretq_s32_f64): Likewise.
2931 (vreinterpret_u8_f64): Likewise.
2932 (vreinterpretq_u8_f64): Likewise.
2933 (vreinterpret_u16_f64): Likewise.
2934 (vreinterpretq_u16_f64): Likewise.
2935 (vreinterpret_u32_f64): Likewise.
2936 (vreinterpretq_u32_f64): Likewise.
2938 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2940 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2941 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2942 (vreinterpret_p8_s8): Likewise.
2943 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2944 (vreinterpret_p8_s16): Likewise.
2945 (vreinterpret_p8_s32): Likewise.
2946 (vreinterpret_p8_s64): Likewise.
2947 (vreinterpret_p8_f32): Likewise.
2948 (vreinterpret_p8_u8): Likewise.
2949 (vreinterpret_p8_u16): Likewise.
2950 (vreinterpret_p8_u32): Likewise.
2951 (vreinterpret_p8_u64): Likewise.
2952 (vreinterpret_p8_p16): Likewise.
2953 (vreinterpretq_p8_s8): Likewise.
2954 (vreinterpretq_p8_s16): Likewise.
2955 (vreinterpretq_p8_s32): Likewise.
2956 (vreinterpretq_p8_s64): Likewise.
2957 (vreinterpretq_p8_f32): Likewise.
2958 (vreinterpretq_p8_u8): Likewise.
2959 (vreinterpretq_p8_u16): Likewise.
2960 (vreinterpretq_p8_u32): Likewise.
2961 (vreinterpretq_p8_u64): Likewise.
2962 (vreinterpretq_p8_p16): Likewise.
2963 (vreinterpret_p16_s8): Likewise.
2964 (vreinterpret_p16_s16): Likewise.
2965 (vreinterpret_p16_s32): Likewise.
2966 (vreinterpret_p16_s64): Likewise.
2967 (vreinterpret_p16_f32): Likewise.
2968 (vreinterpret_p16_u8): Likewise.
2969 (vreinterpret_p16_u16): Likewise.
2970 (vreinterpret_p16_u32): Likewise.
2971 (vreinterpret_p16_u64): Likewise.
2972 (vreinterpret_p16_p8): Likewise.
2973 (vreinterpretq_p16_s8): Likewise.
2974 (vreinterpretq_p16_s16): Likewise.
2975 (vreinterpretq_p16_s32): Likewise.
2976 (vreinterpretq_p16_s64): Likewise.
2977 (vreinterpretq_p16_f32): Likewise.
2978 (vreinterpretq_p16_u8): Likewise.
2979 (vreinterpretq_p16_u16): Likewise.
2980 (vreinterpretq_p16_u32): Likewise.
2981 (vreinterpretq_p16_u64): Likewise.
2982 (vreinterpretq_p16_p8): Likewise.
2983 (vreinterpret_f32_s8): Likewise.
2984 (vreinterpret_f32_s16): Likewise.
2985 (vreinterpret_f32_s32): Likewise.
2986 (vreinterpret_f32_s64): Likewise.
2987 (vreinterpret_f32_u8): Likewise.
2988 (vreinterpret_f32_u16): Likewise.
2989 (vreinterpret_f32_u32): Likewise.
2990 (vreinterpret_f32_u64): Likewise.
2991 (vreinterpret_f32_p8): Likewise.
2992 (vreinterpret_f32_p16): Likewise.
2993 (vreinterpretq_f32_s8): Likewise.
2994 (vreinterpretq_f32_s16): Likewise.
2995 (vreinterpretq_f32_s32): Likewise.
2996 (vreinterpretq_f32_s64): Likewise.
2997 (vreinterpretq_f32_u8): Likewise.
2998 (vreinterpretq_f32_u16): Likewise.
2999 (vreinterpretq_f32_u32): Likewise.
3000 (vreinterpretq_f32_u64): Likewise.
3001 (vreinterpretq_f32_p8): Likewise.
3002 (vreinterpretq_f32_p16): Likewise.
3003 (vreinterpret_s64_s8): Likewise.
3004 (vreinterpret_s64_s16): Likewise.
3005 (vreinterpret_s64_s32): Likewise.
3006 (vreinterpret_s64_f32): Likewise.
3007 (vreinterpret_s64_u8): Likewise.
3008 (vreinterpret_s64_u16): Likewise.
3009 (vreinterpret_s64_u32): Likewise.
3010 (vreinterpret_s64_u64): Likewise.
3011 (vreinterpret_s64_p8): Likewise.
3012 (vreinterpret_s64_p16): Likewise.
3013 (vreinterpretq_s64_s8): Likewise.
3014 (vreinterpretq_s64_s16): Likewise.
3015 (vreinterpretq_s64_s32): Likewise.
3016 (vreinterpretq_s64_f32): Likewise.
3017 (vreinterpretq_s64_u8): Likewise.
3018 (vreinterpretq_s64_u16): Likewise.
3019 (vreinterpretq_s64_u32): Likewise.
3020 (vreinterpretq_s64_u64): Likewise.
3021 (vreinterpretq_s64_p8): Likewise.
3022 (vreinterpretq_s64_p16): Likewise.
3023 (vreinterpret_u64_s8): Likewise.
3024 (vreinterpret_u64_s16): Likewise.
3025 (vreinterpret_u64_s32): Likewise.
3026 (vreinterpret_u64_s64): Likewise.
3027 (vreinterpret_u64_f32): Likewise.
3028 (vreinterpret_u64_u8): Likewise.
3029 (vreinterpret_u64_u16): Likewise.
3030 (vreinterpret_u64_u32): Likewise.
3031 (vreinterpret_u64_p8): Likewise.
3032 (vreinterpret_u64_p16): Likewise.
3033 (vreinterpretq_u64_s8): Likewise.
3034 (vreinterpretq_u64_s16): Likewise.
3035 (vreinterpretq_u64_s32): Likewise.
3036 (vreinterpretq_u64_s64): Likewise.
3037 (vreinterpretq_u64_f32): Likewise.
3038 (vreinterpretq_u64_u8): Likewise.
3039 (vreinterpretq_u64_u16): Likewise.
3040 (vreinterpretq_u64_u32): Likewise.
3041 (vreinterpretq_u64_p8): Likewise.
3042 (vreinterpretq_u64_p16): Likewise.
3043 (vreinterpret_s8_s16): Likewise.
3044 (vreinterpret_s8_s32): Likewise.
3045 (vreinterpret_s8_s64): Likewise.
3046 (vreinterpret_s8_f32): Likewise.
3047 (vreinterpret_s8_u8): Likewise.
3048 (vreinterpret_s8_u16): Likewise.
3049 (vreinterpret_s8_u32): Likewise.
3050 (vreinterpret_s8_u64): Likewise.
3051 (vreinterpret_s8_p8): Likewise.
3052 (vreinterpret_s8_p16): Likewise.
3053 (vreinterpretq_s8_s16): Likewise.
3054 (vreinterpretq_s8_s32): Likewise.
3055 (vreinterpretq_s8_s64): Likewise.
3056 (vreinterpretq_s8_f32): Likewise.
3057 (vreinterpretq_s8_u8): Likewise.
3058 (vreinterpretq_s8_u16): Likewise.
3059 (vreinterpretq_s8_u32): Likewise.
3060 (vreinterpretq_s8_u64): Likewise.
3061 (vreinterpretq_s8_p8): Likewise.
3062 (vreinterpretq_s8_p16): Likewise.
3063 (vreinterpret_s16_s8): Likewise.
3064 (vreinterpret_s16_s32): Likewise.
3065 (vreinterpret_s16_s64): Likewise.
3066 (vreinterpret_s16_f32): Likewise.
3067 (vreinterpret_s16_u8): Likewise.
3068 (vreinterpret_s16_u16): Likewise.
3069 (vreinterpret_s16_u32): Likewise.
3070 (vreinterpret_s16_u64): Likewise.
3071 (vreinterpret_s16_p8): Likewise.
3072 (vreinterpret_s16_p16): Likewise.
3073 (vreinterpretq_s16_s8): Likewise.
3074 (vreinterpretq_s16_s32): Likewise.
3075 (vreinterpretq_s16_s64): Likewise.
3076 (vreinterpretq_s16_f32): Likewise.
3077 (vreinterpretq_s16_u8): Likewise.
3078 (vreinterpretq_s16_u16): Likewise.
3079 (vreinterpretq_s16_u32): Likewise.
3080 (vreinterpretq_s16_u64): Likewise.
3081 (vreinterpretq_s16_p8): Likewise.
3082 (vreinterpretq_s16_p16): Likewise.
3083 (vreinterpret_s32_s8): Likewise.
3084 (vreinterpret_s32_s16): Likewise.
3085 (vreinterpret_s32_s64): Likewise.
3086 (vreinterpret_s32_f32): Likewise.
3087 (vreinterpret_s32_u8): Likewise.
3088 (vreinterpret_s32_u16): Likewise.
3089 (vreinterpret_s32_u32): Likewise.
3090 (vreinterpret_s32_u64): Likewise.
3091 (vreinterpret_s32_p8): Likewise.
3092 (vreinterpret_s32_p16): Likewise.
3093 (vreinterpretq_s32_s8): Likewise.
3094 (vreinterpretq_s32_s16): Likewise.
3095 (vreinterpretq_s32_s64): Likewise.
3096 (vreinterpretq_s32_f32): Likewise.
3097 (vreinterpretq_s32_u8): Likewise.
3098 (vreinterpretq_s32_u16): Likewise.
3099 (vreinterpretq_s32_u32): Likewise.
3100 (vreinterpretq_s32_u64): Likewise.
3101 (vreinterpretq_s32_p8): Likewise.
3102 (vreinterpretq_s32_p16): Likewise.
3103 (vreinterpret_u8_s8): Likewise.
3104 (vreinterpret_u8_s16): Likewise.
3105 (vreinterpret_u8_s32): Likewise.
3106 (vreinterpret_u8_s64): Likewise.
3107 (vreinterpret_u8_f32): Likewise.
3108 (vreinterpret_u8_u16): Likewise.
3109 (vreinterpret_u8_u32): Likewise.
3110 (vreinterpret_u8_u64): Likewise.
3111 (vreinterpret_u8_p8): Likewise.
3112 (vreinterpret_u8_p16): Likewise.
3113 (vreinterpretq_u8_s8): Likewise.
3114 (vreinterpretq_u8_s16): Likewise.
3115 (vreinterpretq_u8_s32): Likewise.
3116 (vreinterpretq_u8_s64): Likewise.
3117 (vreinterpretq_u8_f32): Likewise.
3118 (vreinterpretq_u8_u16): Likewise.
3119 (vreinterpretq_u8_u32): Likewise.
3120 (vreinterpretq_u8_u64): Likewise.
3121 (vreinterpretq_u8_p8): Likewise.
3122 (vreinterpretq_u8_p16): Likewise.
3123 (vreinterpret_u16_s8): Likewise.
3124 (vreinterpret_u16_s16): Likewise.
3125 (vreinterpret_u16_s32): Likewise.
3126 (vreinterpret_u16_s64): Likewise.
3127 (vreinterpret_u16_f32): Likewise.
3128 (vreinterpret_u16_u8): Likewise.
3129 (vreinterpret_u16_u32): Likewise.
3130 (vreinterpret_u16_u64): Likewise.
3131 (vreinterpret_u16_p8): Likewise.
3132 (vreinterpret_u16_p16): Likewise.
3133 (vreinterpretq_u16_s8): Likewise.
3134 (vreinterpretq_u16_s16): Likewise.
3135 (vreinterpretq_u16_s32): Likewise.
3136 (vreinterpretq_u16_s64): Likewise.
3137 (vreinterpretq_u16_f32): Likewise.
3138 (vreinterpretq_u16_u8): Likewise.
3139 (vreinterpretq_u16_u32): Likewise.
3140 (vreinterpretq_u16_u64): Likewise.
3141 (vreinterpretq_u16_p8): Likewise.
3142 (vreinterpretq_u16_p16): Likewise.
3143 (vreinterpret_u32_s8): Likewise.
3144 (vreinterpret_u32_s16): Likewise.
3145 (vreinterpret_u32_s32): Likewise.
3146 (vreinterpret_u32_s64): Likewise.
3147 (vreinterpret_u32_f32): Likewise.
3148 (vreinterpret_u32_u8): Likewise.
3149 (vreinterpret_u32_u16): Likewise.
3150 (vreinterpret_u32_u64): Likewise.
3151 (vreinterpret_u32_p8): Likewise.
3152 (vreinterpret_u32_p16): Likewise.
3153 (vreinterpretq_u32_s8): Likewise.
3154 (vreinterpretq_u32_s16): Likewise.
3155 (vreinterpretq_u32_s32): Likewise.
3156 (vreinterpretq_u32_s64): Likewise.
3157 (vreinterpretq_u32_f32): Likewise.
3158 (vreinterpretq_u32_u8): Likewise.
3159 (vreinterpretq_u32_u16): Likewise.
3160 (vreinterpretq_u32_u64): Likewise.
3161 (vreinterpretq_u32_p8): Likewise.
3162 (vreinterpretq_u32_p16): Likewise.
3164 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3166 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3168 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3170 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3171 (vqnegd_s64): Likewise.
3172 (vqabs_s64): Likewise.
3173 (vqabsd_s64): Likewise.
3175 2014-04-22 Richard Henderson <rth@redhat.com>
3177 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3178 computation to the top of the loop.
3180 2014-04-22 Renlin <renlin.li@arm.com>
3181 Jiong Wang <jiong.wang@arm.com>
3183 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3184 * config/aarch64/aarch64.c (aarch64_layout_frame)
3185 (aarch64_initial_elimination_offset): Likewise.
3187 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3189 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3192 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3194 * machmode.h (bitwise_mode_for_mode): Declare.
3195 * stor-layout.h (bitwise_type_for_mode): Likewise.
3196 * stor-layout.c (bitwise_mode_for_mode): New function.
3197 (bitwise_type_for_mode): Likewise.
3198 * builtins.c (fold_builtin_memory_op): Use it instead of
3199 int_mode_for_mode and build_nonstandard_integer_type.
3201 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3203 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3204 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3205 (*-*-solaris2*): Simplify.
3206 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3207 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3208 *-*-solaris2.9* handling.
3210 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3212 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3213 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3215 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3216 * configure: Regenerate.
3218 * config/i386/sol2-9.h: Remove.
3220 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3221 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3222 Remove Solaris 9 references.
3224 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3226 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3227 (floatuns<GPI:mode><GPF:mode>2): Remove.
3228 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3229 and floatuns conversions.
3230 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3231 and floatuns conversions.
3232 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3233 (w1,w2): New mode attributes for inequal width conversions.
3235 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3237 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3238 the output asm format.
3240 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3242 * config/aarch64/aarch64-simd.md
3243 (aarch64_cm<optab>di): Always split.
3244 (*aarch64_cm<optab>di): New.
3245 (aarch64_cmtstdi): Always split.
3246 (*aarch64_cmtstdi): New.
3248 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3250 PR tree-optimization/60823
3251 * omp-low.c (ipa_simd_modify_function_body): Go through
3252 all SSA_NAMEs and for those refering to vector arguments
3253 which are going to be replaced adjust SSA_NAME_VAR and,
3254 if it is a default definition, change it into a non-default
3255 definition assigned at the beginning of function from new_decl.
3256 (ipa_simd_modify_stmt_ops): Rewritten.
3257 * tree-dfa.c (set_ssa_default_def): When removing default def,
3258 check for NULL loc instead of NULL *loc.
3260 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3262 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3263 restrictions on core registers for DImode values in Thumb2.
3265 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3267 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3268 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3270 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3272 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3273 (*iordi_notzesidi_di): Likewise.
3274 (*iordi_notsesidi_di): Likewise.
3276 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3278 * config/arm/arm-protos.h (tune_params): New struct members.
3279 * config/arm/arm.c: Initialise tune_params per processor.
3280 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3281 for speed, based on new tune_params.
3283 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3285 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3286 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3287 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3288 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3289 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3290 (vrnda_f64): Likewise.
3291 (vrndi_f64): Likewise.
3292 (vrndm_f64): Likewise.
3293 (vrndn_f64): Likewise.
3294 (vrndp_f64): Likewise.
3295 (vrndx_f64): Likewise.
3297 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3299 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3300 GET_MODE_SIZE argument is enum machine_mode.
3302 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3305 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3306 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3308 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3311 * asan.c (asan_emit_stack_protection): Force the base to align to
3312 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3313 appropriate bits if STRICT_ALIGNMENT.
3314 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3316 (expand_used_vars): Leave a space in the stack frame for alignment
3317 if STRICT_ALIGNMENT.
3319 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3321 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3323 (gimple_store_p): Likewise.
3324 (gimple_assign_load_p): Likewise.
3325 (gimple_assign_cast_p): Likewise.
3326 (gimple_clobber_p): Likewise.
3328 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3329 rather than a gimple.
3330 (gimple_assign_cast_p): Likewise.
3332 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3335 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3336 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3338 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3339 more debug information for E500 if -mdebug=reg.
3341 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3344 * config/i386/i386.c (ix86_expand_builtin)
3345 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3346 register for target RTX.
3347 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3349 2014-04-18 Cong Hou <congh@google.com>
3351 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3352 the widen-mult pattern by handling two operands with different sizes,
3353 and operands whose size is smaller than half of the result type.
3355 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3357 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3358 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3359 (do_estimate_edge_time): Compute it.
3360 * ipa-inline.c (want_inline_small_function_p): Bypass
3361 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3363 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3365 * ipa-inline.c (spec_rem): New static variable.
3366 (dump_overall_stats): New function.
3367 (dump_inline_stats): New function.
3369 2014-04-18 Richard Henderson <rth@redhat.com>
3371 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3372 to GET_MODE_SIZE, not a reg_class_t.
3374 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3376 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3377 (vsx_xxmrglw_<mode>): Likewise.
3379 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3382 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3383 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3384 (rs6000_init_hard_regno_mode_ok): Likewise.
3386 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3388 * ipa-inline.c (inline_small_functions): Account only non-cold
3390 * doc/invoke.texi (inline-unit-growth): Update documentation.
3392 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3394 * config/rs6000/rs6000.md (addti3, subti3): New.
3396 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3399 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3400 comment. Check optimize_insn_for_size_p instead of
3401 optimize_insn_for_speed_p.
3403 2014-04-17 Martin Jambor <mjambor@suse.cz>
3405 * gimple-iterator.c (gsi_start_edge): New function.
3406 * gimple-iterator.h (gsi_start_edge): Declare.
3407 * tree-sra.c (single_non_eh_succ): New function.
3408 (disqualify_ops_if_throwing_stmt): Renamed to
3409 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3410 having one non-EH successor BB.
3411 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3412 generate loads into replacements.
3413 (sra_modify_assign): Likewise and and also use the simple path for
3415 (sra_modify_function_body): Commit statements on edges.
3417 2014-04-17 Richard Biener <rguenther@suse.de>
3420 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3421 comparison results and add clarifying comment.
3423 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3425 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3426 (blank_mode): Initialize it.
3427 (emit_mode_size_inline, emit_mode_nunits_inline,
3428 emit_mode_inner_inline): New functions.
3429 (emit_insn_modes_h): Call them and surround their output with
3430 #if GCC_VERSION >= 4001 ... #endif.
3431 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3432 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3433 mode_* arrays if the argument is __builtin_constant_p.
3434 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3435 is enum machine_mode.
3437 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3439 * passes.c (opt_pass::execute): Adjust.
3440 (pass_manager::execute_pass_mode_switching): Likewise.
3441 (early_local_passes::execute): Likewise.
3442 (execute_one_pass): Pass cfun to the pass's execute method.
3443 * tree-pass.h (opt_pass::execute): Add function * argument.
3444 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3445 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3446 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3447 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3448 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3449 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3450 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3451 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3452 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3453 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3454 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3455 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3456 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3457 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3458 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3459 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3460 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3461 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3462 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3463 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3464 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3465 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3466 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3467 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3468 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3469 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3470 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3471 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3472 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3475 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3477 * passes.c (opt_pass::gate): Take function * argument.
3478 (gate_all_early_local_passes): Merge into
3479 (early_local_passes::gate): this.
3480 (gate_all_early_optimizations): Merge into
3481 (all_early_optimizations::gate): this.
3482 (gate_all_optimizations): Mege into
3483 (all_optimizations::gate): this.
3484 (gate_all_optimizations_g): Merge into
3485 (all_optimizations_g::gate): this.
3486 (gate_rest_of_compilation): Mege into
3487 (rest_of_compilation::gate): this.
3488 (gate_postreload): Merge into
3489 (postreload::gate): this.
3490 (dump_one_pass): Pass cfun to the pass's gate method.
3491 (execute_ipa_summary_passes): Likewise.
3492 (execute_one_pass): Likewise.
3493 (ipa_write_summaries_2): Likewise.
3494 (ipa_write_optimization_summaries_1): Likewise.
3495 (ipa_read_summaries_1): Likewise.
3496 (ipa_read_optimization_summaries_1): Likewise.
3497 (execute_ipa_stmt_fixups): Likewise.
3498 * tree-pass.h (opt_pass::gate): Add function * argument.
3499 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3500 combine-stack-adj.c, combine.c, compare-elim.c,
3501 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3502 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3503 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3504 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3505 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3506 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3507 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3508 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3509 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3510 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3511 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3512 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3513 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3514 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3515 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3516 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3517 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3518 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3519 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3520 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3521 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3522 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3523 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3524 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3525 var-tracking.c, vtable-verify.c, web.c: Adjust.
3527 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3529 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3530 * configure: Regenerate.
3532 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3534 * passes.c (dump_one_pass): don't check pass->has_gate.
3535 (execute_ipa_summary_passes): Likewise.
3536 (execute_one_pass): Likewise.
3537 (ipa_write_summaries_2): Likewise.
3538 (ipa_write_optimization_summaries_1): Likewise.
3539 (ipa_read_optimization_summaries_1): Likewise.
3540 (execute_ipa_stmt_fixups): Likewise.
3541 * tree-pass.h (pass_data::has_gate): Remove.
3542 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3543 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3544 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3545 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3546 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3547 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3548 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3549 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3550 gimple-low.c, gimple-ssa-isolate-paths.c,
3551 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3552 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3553 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3554 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3555 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3556 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3557 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3558 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3559 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3560 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3561 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3562 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3563 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3564 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3565 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3566 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3567 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3568 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3569 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3570 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3571 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3572 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3575 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3577 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3579 * passes.c (pass_manager::register_dump_files_1): Merge into
3580 (pass_manager::register_dump_files): this, and remove its handling of
3581 properties since the pass always has the properties anyway.
3582 (pass_manager::pass_manager): Adjust.
3584 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3586 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3587 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3588 dealing with properties.
3589 (pass_manager::register_dump_files): Adjust.
3591 2014-03-20 Mark Wielaard <mjw@redhat.com>
3593 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3594 then represent the bound as normal constant value.
3596 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3599 Forward port from 4.8 branch
3600 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3602 * config/i386/bmiintrin.h (_blsi_u32): New.
3606 (_blsmsk_u32): Ditto.
3607 (_blsmsk_u64): Ditto.
3608 (_tzcnt_u32): Ditto.
3609 (_tzcnt_u64): Ditto.
3611 2014-04-17 Kito Cheng <kito@0xlab.org>
3613 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3615 2014-04-17 Richard Biener <rguenther@suse.de>
3618 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3619 boolean results for comparisons.
3621 2014-04-17 Richard Biener <rguenther@suse.de>
3623 PR tree-optimization/60836
3624 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3625 initial PHI args to be gimple values.
3627 2014-04-17 Richard Biener <rguenther@suse.de>
3629 PR tree-optimization/60841
3630 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3631 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3632 of stmts to SLP build.
3633 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3634 (vect_analyze_slp): Likewise.
3635 (vect_analyze_slp_instance): Likewise.
3636 (vect_build_slp_tree): Limit overall SLP tree growth.
3637 * tree-vectorizer.h (vect_analyze_data_refs,
3638 vect_analyze_slp): Adjust prototypes.
3640 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3642 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3645 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3647 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3648 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3649 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3650 for TARGET_SLOW_PSHUFB
3652 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3654 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3655 * config/i386/i386.c (intel_cost): Ditto.
3657 2014-04-17 Joey Ye <joey.ye@arm.com>
3659 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3661 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3663 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3666 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3668 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3669 (type_all_derivations_known_p): New predicate.
3670 (type_all_ctors_visible_p): New predicate.
3671 (type_possibly_instantiated_p): New predicate.
3672 (get_odr_type): Compute all_derivations_known.
3673 (dump_odr_type): Dump the flag.
3674 (maybe_record_type): Cleanup.
3675 (record_target_from_binfo): Add bases_to_consider array;
3676 record bases for types w/o instances and skip CXX destructor.
3677 (possible_polymorphic_call_targets_1): Add bases_to_consider
3678 and consider_construction parameters; check if type may have instance.
3679 (get_polymorphic_call_info): Set maybe_in_construction to true
3680 when we know nothing.
3681 (record_targets_from_bases): Skip CXX destructors; they are
3682 never called for types in construction.
3683 (possible_polymorphic_call_targets): Do not record target when
3684 type may not have instance.
3686 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3689 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3690 external aliases alive, too.
3692 2014-04-16 Andrew Pinski <apinski@cavium.com>
3694 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3697 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3699 * final.c (compute_alignments): Do not apply loop alignment to a block
3700 falling through to the exit.
3702 2014-04-16 Catherine Moore <clm@codesourcery.com>
3704 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3705 Adjust constraints for microMIPS store patterns.
3707 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3709 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3711 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3713 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3714 (append_use): Run at -O0.
3715 (append_vdef): Likewise.
3716 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3717 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3719 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3721 PR tree-optimization/60844
3722 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3723 (propagate_op_to_single_use, remove_visited_stmt_chain,
3724 linearize_expr, repropagate_negates, reassociate_bb): Use it
3725 instead of gsi_remove.
3727 2014-04-16 Martin Jambor <mjambor@suse.cz>
3729 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3730 ipa_transforms_to_apply.
3731 (cgraph_function_versioning): Assert that old_node has empty
3732 ipa_transforms_to_apply.
3733 * trans-mem.c (ipa_tm_create_version): Likewise.
3734 * tree-inline.c (tree_function_versioning): Do not duplicate
3735 ipa_transforms_to_apply.
3737 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3740 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3742 Pass necessary as flags on 64-bit Solaris/x86.
3743 Use lowercase relocs for x86_64-*-*.
3744 * configure: Regenerate.
3746 2014-04-15 Jan Hubicka <jh@suse.cz>
3748 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3749 (maybe_record_node, likely_target_p): Use it.
3751 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3754 Revert following patch
3756 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3759 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3760 software floating point or no floating point registers, do not
3761 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3762 in GPRs that occurs after we tested for GPRs that would never be
3765 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3766 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3767 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3768 specifically allow DDmode, since that does not use the SPE SIMD
3771 2014-03-21 Mark Wielaard <mjw@redhat.com>
3773 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3774 as unsigned or int depending on type and value used.
3776 2014-04-15 Richard Biener <rguenther@suse.de>
3778 PR rtl-optimization/56965
3779 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3780 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3782 * alias.c (true_dependence_1): Do not call
3783 nonoverlapping_component_refs_p.
3784 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3785 nonoverlapping_component_refs_p.
3786 (indirect_refs_may_alias_p): Likewise.
3788 2014-04-15 Teresa Johnson <tejohnson@google.com>
3790 * cfg.c (dump_bb_info): Fix flags check.
3791 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3793 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3795 PR rtl-optimization/60663
3796 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3799 2014-04-15 Richard Biener <rguenther@suse.de>
3801 * lto-streamer.h (LTO_major_version): Bump to 4.
3803 2014-04-15 Richard Biener <rguenther@suse.de>
3805 * common.opt (lto_partition_model): New enum.
3806 (flto-partition=): Merge separate options with a single with argument,
3807 add -flto-partition=one support.
3808 * flag-types.h (enum lto_partition_model): Declare.
3809 * opts.c (finish_options): Remove duplicate -flto-partition=
3811 * lto-wrapper.c (run_gcc): Adjust.
3813 2014-04-15 Richard Biener <rguenther@suse.de>
3815 * alias.c (ncr_compar): New function.
3816 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3818 2014-04-15 Richard Biener <rguenther@suse.de>
3820 * alias.c (record_component_aliases): Do not walk BINFOs.
3822 2014-04-15 Richard Biener <rguenther@suse.de>
3824 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3825 Add struct function argument and adjust.
3826 (find_func_aliases_for_call): Likewise.
3827 (find_func_aliases): Likewise.
3828 (find_func_clobbers): Likewise.
3829 (intra_create_variable_infos): Likewise.
3830 (compute_points_to_sets): Likewise.
3831 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3833 2014-04-15 Richard Biener <rguenther@suse.de>
3835 * tree.c (iterative_hash_expr): Use enum tree_code_class
3836 to store TREE_CODE_CLASS.
3837 (tree_block): Likewise.
3838 (tree_set_block): Likewise.
3839 * tree.h (fold_build_pointer_plus_loc): Use
3840 convert_to_ptrofftype_loc.
3842 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3845 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3848 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3850 * cfgloop.h (struct loop): Move force_vectorize down.
3851 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3852 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3853 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3854 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3855 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3856 * tree-core.h (enum annot_expr_kind): Add new kind values.
3857 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3858 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3860 * tree.def (ANNOTATE_EXPR): Tweak comment.
3862 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3864 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3867 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3869 * tree.h (TYPE_IDENTIFIER): Declare.
3870 * tree.c (subrange_type_for_debug_p): Use it.
3871 * godump.c (go_format_type): Likewise.
3872 * dwarf2out.c (is_cxx_auto, modified_type_die,
3873 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3874 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3876 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3879 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3881 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3883 * config/i386/i386.c (examine_argument): Return bool. Return true if
3884 parameter should be passed in memory.
3885 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3886 (construct_container): Update calls to examine_argument.
3887 (function_arg_advance_64): Ditto.
3888 (return_in_memory_32): Merge with ix86_return_in_memory.
3889 (return_in_memory_64): Ditto.
3890 (return_in_memory_ms_64): Ditto.
3892 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3894 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3895 * coverage.c (coverage_compute_profile_id): Handle externally visible
3898 2014-04-14 Martin Jambor <mjambor@suse.cz>
3900 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3901 DECL_DISREGARD_INLINE_LIMITS functions.
3903 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3906 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3908 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3911 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3912 optimize_insn_for_speed_p instead of
3913 optimize_function_for_speed_p.
3915 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3917 * doc/invoke.texi (free): Document AArch64.
3919 2014-04-14 Richard Biener <rguenther@suse.de>
3921 PR tree-optimization/60042
3922 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3923 (insert_into_preds_of_block): Do not prevent PHI insertion
3924 for REFERENCE exprs here ...
3925 (eliminate_dom_walker::before_dom_children): ... but prevent
3926 their use here under similar conditions when applied to the
3927 IL after PRE optimizations.
3929 2014-04-14 Richard Biener <rguenther@suse.de>
3931 * passes.def: Move early points-to after early SRA.
3933 2014-04-14 Richard Biener <rguenther@suse.de>
3935 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3936 check for which sign-changes we allow when forwarding
3937 a converted value into a switch.
3939 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3941 * stor-layout.c (place_field): Finalize non-constant offset for the
3944 2014-04-14 Richard Biener <rguenther@suse.de>
3946 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3948 (expand_switch_using_bit_tests_p): Likewise.
3949 (process_switch): Compute and pass on speed_p based on the
3951 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3952 optimize_bb_for_speed_p.
3954 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3956 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3957 * function.h (struct function): Rename has_force_vect_loops into
3958 has_force_vectorize_loops.
3959 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3960 (input_struct_function_base): Likewise.
3961 * lto-streamer-out.c (output_cfg): Likewise.
3962 (output_struct_function_base): Likewise.
3963 * omp-low.c (expand_omp_simd): Likewise.
3964 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3965 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3966 (version_loop_for_if_conversion): Likewise.
3967 (tree_if_conversion): Likewise.
3968 (main_tree_if_conversion): Likewise.
3969 (gate_tree_if_conversion): Likewise.
3970 * tree-inline.c (copy_loops): Likewise.
3971 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3972 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3973 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3974 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3975 * tree-vectorizer.c (vectorize_loops): Likewise.
3976 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3978 2014-04-14 Richard Biener <rguenther@suse.de>
3981 * lto-streamer-out.c (wrap_refs): New function.
3982 (lto_output): Wrap symbol references in global initializes in
3983 type-preserving MEM_REFs.
3985 2014-04-14 Christian Bruel <christian.bruel@st.com>
3987 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3989 2014-04-14 Christian Bruel <christian.bruel@st.com>
3991 * config/sh/sh.md (setmemqi): New expand pattern.
3992 * config/sh/sh.h (CLEAR_RATIO): Define.
3993 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3994 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3996 2014-04-14 Richard Biener <rguenther@suse.de>
3999 * fold-const.c (negate_expr_p): Don't negate directional rounding
4001 (fold_negate_expr): Likewise.
4003 2014-04-14 Richard Biener <rguenther@suse.de>
4005 PR tree-optimization/59817
4006 PR tree-optimization/60453
4007 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4008 recursion to catch all CHRECs in the scalar evolution and restrict
4009 the predicate for the remains appropriately.
4011 2014-04-12 Catherine Moore <clm@codesourcery.com>
4013 * config/mips/constraints.md: Add new register constraint "kb".
4014 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4015 (*movhi_internal): Likewise.
4016 (*movqi_internal): Likewise.
4017 * config/mips/mips.h (M16_STORE_REGS): New register class.
4018 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4019 (REG_CLASS_CONTENTS): Likewise.
4020 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4022 2014-04-11 Tobias Burnus <burnus@net-b.de>
4025 * doc/invoke.texi (-Wformat-signedness): Document it.
4026 (Wformat=2): Mention that this enables -Wformat-signedness.
4028 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4030 * common/config/epiphany/epiphany-common.c
4031 (epiphany_option_optimization_table): Enable section anchors by
4032 default at -O1 or higher.
4033 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4034 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4035 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4036 carries no extra cost.
4037 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4038 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4039 * config/epiphany/predicates.md (memclob_operand): New predicate.
4040 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4041 Use memclob_operand predicate and X constraint for operand 3.
4043 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4045 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4046 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4049 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4051 PR rtl-optimization/60651
4052 * mode-switching.c (optimize_mode_switching): Make sure to emit
4053 sets of a lower numbered entity before sets of a higher numbered
4054 entity to a mode of the same or lower priority.
4055 When creating a seginfo for a basic block that starts with a code
4056 label, move the insertion point past the code label.
4057 (new_seginfo): Document and enforce requirement that
4058 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4059 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4060 * doc/tm.texi: Regenerate.
4062 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4065 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4067 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4069 * BASE-VER: Set to 4.10.0.
4071 2014-04-11 Tobias Burnus <burnus@net-b.de>
4074 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4075 * doc/gcc.texi (Service): Update description in the @menu
4076 * doc/invoke.texi (Option Summary): Remove misplaced and
4079 2014-04-11 Steve Ellcey <sellcey@mips.com>
4080 Jakub Jelinek <jakub@redhat.com>
4083 * expr.c (convert_move): Use emit_store_flag_force instead of
4084 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4087 2014-04-11 Richard Biener <rguenther@suse.de>
4090 * varasm.c (assemble_alias): Avoid endless error reporting
4091 recursion by setting TREE_ASM_WRITTEN.
4093 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4095 * config/s390/s390.md: Add a splitter for NOT rtx.
4097 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4099 PR rtl-optimization/60663
4100 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4102 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4103 Jakub Jelinek <jakub@redhat.com>
4106 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4107 flag from decl_node to node.
4109 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4112 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4113 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4114 ameliorating the cases where it can be.
4116 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4119 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4121 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4122 (loadsync_<mode>): Change mode.
4123 (load_quadpti, store_quadpti): New.
4124 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4125 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4126 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4128 2014-04-09 Cong Hou <congh@google.com>
4131 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4134 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4136 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4137 instead of vnor to exploit possible fusion opportunity in the
4139 (altivec_expand_vec_perm_const_le): Likewise.
4141 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4143 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4144 (loadsync_<mode>): Change mode.
4145 (load_quadpti, store_quadpti): New.
4146 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4147 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4149 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4152 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4153 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4154 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4156 2014-04-08 Richard Biener <rguenther@suse.de>
4159 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4160 a 64bit widest int print double-int similar to on HWI64 hosts.
4162 2014-04-08 Richard Biener <rguenther@suse.de>
4164 PR tree-optimization/60785
4165 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4166 default defs properly.
4168 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4170 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4171 (Weffc++): Likewise.
4173 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4175 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4176 set completep to false rather than true.
4178 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4181 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4182 ARM_TARGET2_DWARF_FORMAT.
4184 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4187 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4188 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4191 2014-04-07 Richard Biener <rguenther@suse.de>
4193 PR tree-optimization/60766
4194 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4195 (may_eliminate_iv): Convert cand_value_at result to desired type.
4197 2014-04-07 Jason Merrill <jason@redhat.com>
4200 * common.opt (-fno-gnu-unique): Add.
4201 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4203 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4205 * haifa-sched.c: Fix outdated function reference and minor
4206 grammar errors in introductory comment.
4208 2014-04-07 Richard Biener <rguenther@suse.de>
4211 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4213 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4215 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4218 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4219 size accounting for thunks.
4220 (pa_asm_output_mi_thunk): Use final_start_function() and
4221 final_end_function() to output function start and end directives.
4223 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4225 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4226 device specific ISA/ feature information. Remove short_sp and
4227 errata_skip ds. Add avr_device_specific_features enum to have device
4229 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4230 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4231 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4232 updated device specific info.
4233 * config/avr/avr-mcus.def: Merge device specific details to
4234 dev_attribute field.
4235 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4237 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4238 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4239 assembler if RMW isa supported by current device.
4240 * config/avr/genmultilib.awk: Update as device info structure changed.
4241 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4243 2014-04-04 Cong Hou <congh@google.com>
4245 PR tree-optimization/60656
4246 * tree-vect-stmts.c (supportable_widening_operation):
4247 Fix a bug that elements in a vector with vect_used_by_reduction
4248 property are incorrectly reordered when the operation on it is not
4249 consistant with the one in reduction operation.
4251 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4253 PR rtl-optimization/60155
4254 * gcse.c (record_set_data): New function.
4255 (single_set_gcse): New function.
4256 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4257 (hoist_code): Likewise.
4258 (get_pressure_class_and_nregs): Likewise.
4260 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4262 * explow.c (probe_stack_range): Emit a final optimization blockage.
4264 2014-04-04 Anthony Green <green@moxielogic.com>
4266 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4269 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4272 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4273 flags are set only during streaming.
4274 * ipa.c (process_references, walk_polymorphic_call_targets,
4275 symtab_remove_unreachable_nodes): Drop bodies of always inline
4276 after early inlining.
4277 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4279 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4280 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4283 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4286 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4289 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4291 (cortex_a53_fdivd): Likewise.
4293 2014-04-04 Martin Jambor <mjambor@suse.cz>
4296 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4298 * cgraph.c (clone_of_p): Also return true if thunks match.
4299 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4300 cgraph_function_or_thunk_node and an obsolete comment.
4301 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4303 (build_function_decl_skip_args): Likewise.
4304 (set_new_clone_decl_and_node_flags): New function.
4305 (duplicate_thunk_for_node): Likewise.
4306 (redirect_edge_duplicating_thunks): Likewise.
4307 (cgraph_clone_node): New parameter args_to_skip, pass it to
4308 redirect_edge_duplicating_thunks which is called instead of
4309 cgraph_redirect_edge_callee.
4310 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4311 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4313 2014-04-04 Jeff Law <law@redhat.com>
4316 * config/arm/predicates.md (const_int_I_operand): New predicate.
4317 (const_int_M_operand): Similarly.
4318 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4320 (insv_t2, extv_reg, extzv_t2): Likewise.
4321 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4322 (pop_multiple_with_writeback_and_return): Likewise.
4323 (vfp_pop_multiple_with_writeback): Likewise
4325 2014-04-04 Richard Biener <rguenther@suse.de>
4328 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4329 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4331 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4332 * gimplify.c (gimple_add_tmp_var_fn): New function.
4333 * gimple-expr.h (create_tmp_reg_fn): Declare.
4334 * gimple-expr.c (create_tmp_reg_fn): New function.
4335 * gimple-low.c (record_vars_into): Don't change cfun.
4336 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4337 code generation without cfun.
4339 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4342 * Makefile.in (install-driver): Fix shell scripting.
4344 2014-04-03 Cong Hou <congh@google.com>
4346 PR tree-optimization/60505
4347 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4348 threshold of number of iterations below which no vectorization
4350 * tree-vect-loop.c (new_loop_vec_info):
4351 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4352 * tree-vect-loop.c (vect_analyze_loop_operations):
4353 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4354 * tree-vect-loop.c (vect_transform_loop):
4355 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4356 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4357 of iterations of the loop and see if we should build the epilogue.
4359 2014-04-03 Richard Biener <rguenther@suse.de>
4361 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4362 (streamer_tree_cache_create): Adjust.
4363 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4364 to allow optional nodes array.
4365 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4366 (streamer_tree_cache_append): Likewise.
4367 (streamer_tree_cache_create): Create nodes array optionally
4368 as specified by parameter.
4369 * lto-streamer-out.c (create_output_block): Avoid maintaining
4370 the node array in the writer cache.
4371 (DFS_write_tree): Remove assertion.
4372 (produce_asm_for_decls): Free the out decl state hash table early.
4373 * lto-streamer-in.c (lto_data_in_create): Adjust for
4374 streamer_tree_cache_create prototype change.
4376 2014-04-03 Richard Biener <rguenther@suse.de>
4378 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4379 set TREE_CHAIN to NULL_TREE.
4381 2014-04-03 Richard Biener <rguenther@suse.de>
4383 PR tree-optimization/60740
4384 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4385 over all GIMPLE_COND operands.
4387 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4389 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4390 (Weffc++): Remove Scott's numbering, merge lists and reference
4393 2014-04-03 Nick Clifton <nickc@redhat.com>
4395 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4398 2014-04-03 Martin Jambor <mjambor@suse.cz>
4400 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4401 mention gcc_unreachable before failing.
4402 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4405 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4408 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4409 inconsistent code and instead mark the context inconsistent.
4410 (possible_polymorphic_call_targets): For inconsistent contexts
4411 return empty complete list.
4413 2014-04-02 Anthony Green <green@moxielogic.com>
4415 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4416 (extendqisi2, extendhisi2): Define.
4417 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4418 (WCHAR_TYPE): Change to unsigned int.
4420 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4422 PR tree-optimization/60733
4423 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4424 insertion point for PHI candidates to be the end of the feeding
4425 block for the PHI argument.
4427 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4429 PR rtl-optimization/60650
4430 * lra-constraints.c (process_alt_operands): Decrease reject for
4431 earlyclobber matching.
4433 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4435 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4437 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4439 * config/spu/spu.c (pad_bb): Do not crash when the last
4440 insn is CODE_FOR_blockage.
4442 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4444 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4445 lies outside the target mode.
4447 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4450 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4451 software floating point or no floating point registers, do not
4452 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4453 in GPRs that occurs after we tested for GPRs that would never be
4456 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4457 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4458 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4459 specifically allow DDmode, since that does not use the SPE SIMD
4462 2014-04-02 Richard Biener <rguenther@suse.de>
4465 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4466 MODE_INTs. Properly use negv_optab.
4467 (expand_abs): Likewise.
4469 2014-04-02 Richard Biener <rguenther@suse.de>
4472 * Makefile.in (install-driver): Guard extra installs with special
4475 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4477 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4480 2014-04-01 Richard Henderson <rth@redhat.com>
4483 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4484 alternative enabled before register allocation.
4486 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4488 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4489 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4491 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4492 (nios2_got_address): Update nios2_large_got_address call site.
4493 (nios2_delegitimize_address): New function.
4494 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4495 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4496 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4498 2014-04-01 Martin Husemann <martin@duskware.de>
4500 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4503 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4505 PR rtl-optimization/60604
4506 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4507 check from register_operand.
4508 (register_operand): Redefine in terms of general_operand.
4509 (nonmemory_operand): Use register_operand for the non-constant cases.
4511 2014-04-01 Richard Biener <rguenther@suse.de>
4513 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4515 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4517 * doc/invoke.texi (mapp-regs): Clarify.
4519 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4521 * config/i386/avx512fintrin.h (__v32hi): Define type.
4522 (__v64qi): Likewise.
4523 (_mm512_set1_epi8): Define.
4524 (_mm512_set1_epi16): Define.
4525 (_mm512_set4_epi32): Define.
4526 (_mm512_set4_epi64): Define.
4527 (_mm512_set4_pd): Define.
4528 (_mm512_set4_ps): Define.
4529 (_mm512_setr4_epi64): Define.
4530 (_mm512_setr4_epi32): Define.
4531 (_mm512_setr4_pd): Define.
4532 (_mm512_setr4_ps): Define.
4533 (_mm512_setzero_epi32): Define.
4535 2014-03-31 Martin Jambor <mjambor@suse.cz>
4538 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4539 callsite_arguments_match_p. Updated all callers. Also check types of
4540 corresponding formal parameters and actual arguments.
4541 (not_all_callers_have_enough_arguments_p) Renamed to
4542 some_callers_have_mismatched_arguments_p.
4544 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4546 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4548 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4551 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4554 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4556 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4557 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4559 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4560 Use FMAMODE_NOVF512 mode iterator.
4561 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4562 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4563 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4564 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4566 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4567 Use VF_128_256 mode iterator.
4568 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4571 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4573 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4574 static chain if needed.
4576 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4579 * lra-constraints.c (index_part_to_reg): New.
4580 (process_address): Use it.
4582 2014-03-27 Jeff Law <law@redhat.com>
4583 Jakub Jelinek <jakub@redhat.com>
4586 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4587 gen_rtx_{PLUS,MULT} to build up the address expression.
4589 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4590 creating non-canonical RTL.
4592 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4595 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4596 functions; reorganize to make cheap checks first.
4597 (inline_small_functions): Do not estimate growth when dumping;
4599 * ipa-inline.h (inline_summary): Add min_size.
4600 (growth_likely_positive): New function.
4601 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4602 (set_cond_stmt_execution_predicate): Cleanup.
4603 (estimate_edge_size_and_time): Compute min_size.
4604 (estimate_calls_size_and_time): Likewise.
4605 (estimate_node_size_and_time): Likewise.
4606 (inline_update_overall_summary): Update min_size.
4607 (do_estimate_edge_time): Likewise.
4608 (do_estimate_edge_size): Update.
4609 (do_estimate_edge_hints): Update.
4610 (growth_likely_positive): New function.
4612 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4615 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4616 also if addr has VOIDmode.
4618 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4620 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4621 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4623 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4624 instructions as well as AdvancedSIMD loads.
4626 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4628 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4629 Use crypto_aese type.
4630 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4631 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4632 crypto_aese, crypto_aesmc. Move to types.md.
4633 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4635 * config/arm/iterators.md (crypto_type): Likewise.
4637 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4639 * cgraph.c: Include expr.h and tree-dfa.h.
4640 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4643 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4646 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4647 regs from checking multi-reg pseudos.
4649 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4651 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4653 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4655 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4656 if it would clobber the stack pointer, even temporarily.
4658 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4660 * mode-switching.c: Make small adjustments to the top comment.
4662 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4664 * config/rs6000/constraints.md (wD constraint): New constraint to
4665 match the constant integer to get the top DImode/DFmode out of a
4666 vector in a VSX register.
4668 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4669 match the constant integer to get the top DImode/DFmode out of a
4670 vector in a VSX register.
4672 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4675 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4678 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4679 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4681 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4682 Optimize vec_extract of 64-bit values, where the value being
4683 extracted is in the top word, where we can use scalar
4684 instructions. Add direct move and store support. Combine the big
4685 endian/little endian vector select load support into a single insn.
4686 (vsx_extract_<mode>_internal1): Likewise.
4687 (vsx_extract_<mode>_internal2): Likewise.
4688 (vsx_extract_<mode>_load): Likewise.
4689 (vsx_extract_<mode>_store): Likewise.
4690 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4691 combined into vsx_extract_<mode>_load.
4692 (vsx_extract_<mode>_one_le): Likewise.
4694 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4695 define the top 64-bit vector element.
4697 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4700 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4701 Document vec_vbpermq builtin.
4704 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4705 enable use of xxsldwi and xxpermdi builtin functions.
4706 (vec_xxpermdi): Likewise.
4708 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4709 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4711 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4713 PR rtl-optimization/60650
4714 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4716 (find_spills_for): New.
4717 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4718 Spill all pseudos on the second iteration.
4720 2014-03-27 Marek Polacek <polacek@redhat.com>
4723 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4726 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4728 * config/s390/s390.c (s390_can_use_return_insn): Check for
4729 call-saved FPRs on 31 bit.
4731 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4734 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4735 if they need regimplification, just drop them instead of
4736 calling gimple_regimplify_operands on them.
4738 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4741 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4742 (aarch64_frame_pointer_required): Adjust logic.
4743 (aarch64_can_eliminate): Adjust logic.
4744 (aarch64_override_options_after_change): Adjust logic.
4746 2014-03-27 Dehao Chen <dehao@google.com>
4748 * ipa-inline.c (early_inliner): Update node's inline info.
4750 2014-03-26 Dehao Chen <dehao@google.com>
4752 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4753 compiler inserted conditional jumps for NAN float check.
4755 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4757 * ubsan.h (ubsan_create_data): Change second argument's type
4758 to const location_t *.
4759 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4761 (ubsan_create_data): Change second argument to const location_t *PLOC.
4762 Create Loc field whenever PLOC is non-NULL.
4763 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4764 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4768 * real.c (real_to_integer2): Change type of low to UHWI.
4770 2014-03-26 Tobias Burnus <burnus@net-b.de>
4772 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4773 (CILK_SELF_SPECS): New define.
4774 (driver_self_specs): Use it.
4776 2014-03-26 Richard Biener <rguenther@suse.de>
4778 * tree-pretty-print.c (percent_K_format): Implement special
4779 case for LTO and its stripped down BLOCK tree.
4781 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4784 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4786 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4787 one range is range_int_cst_p, but not both, at least optimize
4788 addition/subtraction of 0 and multiplication by 0 or 1.
4789 * gimple-fold.c (gimple_fold_call): Fold
4790 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4791 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4792 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4794 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4796 PR rtl-optimization/60452
4797 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4798 <case REG>: Return 1 for invalid offsets from the frame pointer.
4800 2014-03-26 Marek Polacek <polacek@redhat.com>
4803 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4806 2014-03-26 Marek Polacek <polacek@redhat.com>
4809 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4812 2014-03-26 Marek Polacek <polacek@redhat.com>
4815 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4816 multiplication in unsigned type.
4818 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4820 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4822 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4824 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4826 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4829 * cif-code.def (UNREACHABLE) New code.
4830 * ipa-inline.c (inline_small_functions): Skip edges to
4831 __builtlin_unreachable.
4832 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4833 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4834 predicate to __bulitin_unreachable.
4835 (set_cond_stmt_execution_predicate): Fix issue when
4836 invert_tree_comparison returns ERROR_MARK.
4837 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4838 propagate to inline clones.
4839 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4841 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4842 * cgraphclones.c (cgraph_clone_node): If call destination is already
4843 ureachable, do not redirect it back.
4844 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4847 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4849 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4850 Do not modify inline clones.
4852 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4854 * config/i386/i386.md (general_sext_operand): New mode attr.
4855 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4856 don't generate (sign_extend (const_int)).
4857 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4858 operands[2]. Use We constraint instead of <i> and
4859 <general_sext_operand> predicate instead of <general_operand>.
4860 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4861 * config/i386/constraints.md (We): New constraint.
4862 * config/i386/predicates.md (x86_64_sext_operand,
4863 sext_operand): New predicates.
4865 2014-03-25 Martin Jambor <mjambor@suse.cz>
4868 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4869 inconsistent devirtualizations to __builtin_unreachable.
4871 2014-03-25 Marek Polacek <polacek@redhat.com>
4874 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4876 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4878 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4879 order of elements for big-endian.
4881 2014-03-25 Richard Biener <rguenther@suse.de>
4884 * gimplify-me.c (gimple_regimplify_operands): Update the
4887 2014-03-25 Martin Jambor <mjambor@suse.cz>
4890 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4891 (lto_output_varpool_node): Likewise.
4892 (input_overwrite_node): Likewise.
4893 (input_varpool_node): Likewise.
4895 2014-03-25 Richard Biener <rguenther@suse.de>
4897 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4898 (run_gcc): Likewise.
4900 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4902 * combine.c (simplify_compare_const): Add MODE argument.
4903 Handle mode_width 0 as very large mode_width.
4904 (try_combine, simplify_comparison): Adjust callers.
4906 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4907 type to avoid signed integer overflow.
4908 * explow.c (plus_constant): Likewise.
4910 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4912 * doc/generic.texi: Correct typos.
4914 2014-03-24 Tobias Burnus <burnus@net-b.de>
4916 * doc/invoke.texi (-flto): Expand section about
4917 using static libraries with LTO.
4919 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4921 PR rtl-optimization/60501
4922 * optabs.def (addptr3_optab): New optab.
4923 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4924 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4925 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4927 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4929 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4931 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4933 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4936 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4937 (_mm256_undefined_ps): Define.
4938 (_mm256_undefined_pd): Define.
4939 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4940 (_mm_undefined_pd): Define.
4941 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4942 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4943 (_mm512_undefined_ps): Define.
4944 (_mm512_undefined_pd): Define.
4945 Use _mm*_undefined_*.
4946 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4948 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4950 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4951 (lshr_simd): DI mode added.
4952 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4953 (aarch64_ushr_simddi): Likewise.
4954 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4955 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4956 (vshrd_n_u64): Likewise.
4958 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4960 * Makefile.in (s-macro_list): Depend on cc1.
4962 2014-03-23 Teresa Johnson <tejohnson@google.com>
4964 * ipa-utils.c (ipa_print_order): Use specified dump file.
4966 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4968 PR rtl-optimization/60601
4969 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4971 * gcc.c (eval_spec_function): Initialize save_growing_value.
4973 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4976 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4977 code == MINUS_EXPR, never swap op0 with op1.
4979 * toplev.c (init_local_tick): Avoid signed integer multiplication
4981 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4982 shift by first operand's bitsize.
4984 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4987 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4989 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4990 TARGET_ISA_64BIT_P(x).
4992 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4994 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4995 pattern for vector nor instead of subtract from splat(-1).
4996 (altivec_expand_vec_perm_const_le): Likewise.
4998 2014-03-21 Richard Henderson <rth@twiddle.net>
5001 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5002 related insns after epilogue_completed.
5004 2014-03-21 Martin Jambor <mjambor@suse.cz>
5007 * cgraph.h (symtab_node): New flag body_removed.
5008 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5009 when removing bodies.
5010 * symtab.c (dump_symtab_base): Dump body_removed flag.
5011 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5012 had their bodies removed.
5014 2014-03-21 Martin Jambor <mjambor@suse.cz>
5017 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5020 2014-03-21 Richard Biener <rguenther@suse.de>
5022 PR tree-optimization/60577
5023 * tree-core.h (struct tree_base): Document nothrow_flag use
5024 in VAR_DECL_NONALIASED.
5025 * tree.h (VAR_DECL_NONALIASED): New.
5026 (may_be_aliased): Adjust.
5027 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5029 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5031 * expr.c (expand_expr_real_1): Remove outdated comment.
5033 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5036 * ira.c (adjust_cleared_regs): Call copy_rtx on
5037 *reg_equiv[REGNO (loc)].src_p before passing it to
5038 simplify_replace_fn_rtx.
5041 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5042 into CONST, put pic register as first operand of PLUS. Use
5043 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5045 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5047 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5049 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5051 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5052 around for store forwarding issue in the FPU on the UT699.
5053 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5054 loads and operations if -mfix-ut699 is specified.
5055 (divtf3_hq): Tweak attribute.
5056 (sqrttf2_hq): Likewise.
5058 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5060 * calls.c (store_one_arg): Remove incorrect const qualification on the
5061 type of the temporary.
5062 * cfgexpand.c (expand_return): Likewise.
5063 * expr.c (expand_constructor): Likewise.
5064 (expand_expr_real_1): Likewise.
5066 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5068 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5071 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5074 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5076 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5078 * config/arm/aarch-common-protos.h
5079 (alu_cost_table): Fix spelling of "extend".
5080 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5082 2014-03-19 Richard Biener <rguenther@suse.de>
5085 * tree-core.h (tree_type_common): Re-order pointer members
5086 to reduce recursion depth during GC walks.
5088 2014-03-19 Marek Polacek <polacek@redhat.com>
5091 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5092 before accessing it.
5094 2014-03-19 Richard Biener <rguenther@suse.de>
5097 * lto-streamer-in.c (input_function): In WPA stage do not drop
5100 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5102 PR tree-optimization/60559
5103 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5104 with build_zero_cst assignment.
5106 2014-03-18 Kai Tietz <ktietz@redhat.com>
5108 PR rtl-optimization/56356
5109 * sdbout.c (sdbout_parms): Verify that parms'
5110 incoming argument is valid.
5111 (sdbout_reg_parms): Likewise.
5113 2014-03-18 Richard Henderson <rth@redhat.com>
5116 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5117 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5118 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5120 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5122 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5123 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5124 Italicize plugin event names in description. Explain that
5125 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5126 Remind that no GCC functions should be called after PLUGIN_FINISH.
5127 Explain what pragmas with expansion are.
5129 2014-03-18 Martin Liska <mliska@suse.cz>
5131 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5132 gimple call statement is update.
5133 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5134 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5136 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5139 * ubsan.c (ubsan_instrument_unreachable): Call
5140 initialize_sanitizer_builtins.
5141 (ubsan_pass): Likewise.
5144 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5145 varpool_finalize_decl instead of rest_of_decl_compilation.
5147 2014-03-18 Richard Biener <rguenther@suse.de>
5149 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5150 by using bitmap_and_compl instead of bitmap_and_compl_into.
5151 (df_rd_transfer_function): Likewise.
5153 2014-03-18 Richard Biener <rguenther@suse.de>
5155 * doc/lto.texi (fresolution): Fix typo.
5157 2014-03-18 Richard Biener <rguenther@suse.de>
5159 * doc/invoke.texi (flto): Update for changes in 4.9.
5161 2014-03-18 Richard Biener <rguenther@suse.de>
5163 * doc/loop.texi: Remove section on the removed lambda framework.
5164 Update loop docs with recent changes in preserving loop structure.
5166 2014-03-18 Richard Biener <rguenther@suse.de>
5168 * doc/lto.texi (-fresolution): Document.
5170 2014-03-18 Richard Biener <rguenther@suse.de>
5172 * doc/contrib.texi: Adjust my name.
5174 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5177 * internal-fn.c: Include diagnostic-core.h.
5178 (expand_BUILTIN_EXPECT): New function.
5179 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5180 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5181 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5182 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5184 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5185 Revert 3 argument __builtin_expect code.
5186 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5187 * gimple-fold.c (gimple_fold_call): Likewise.
5188 * tree.h (fold_builtin_expect): New prototype.
5189 * builtins.c (build_builtin_expect_predicate): Add predictor
5190 argument, if non-NULL, create 3 argument __builtin_expect.
5191 (fold_builtin_expect): No longer static. Add ARG2 argument,
5192 pass it through to build_builtin_expect_predicate.
5193 (fold_builtin_2): Adjust caller.
5194 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5195 * internal-fn.def (BUILTIN_EXPECT): New.
5197 2014-03-18 Tobias Burnus <burnus@net-b.de>
5200 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5201 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5202 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5204 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5207 * predict.c (combine_predictions_for_bb): Fix up formatting.
5208 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5209 fill what it points to if non-NULL.
5210 (tree_predict_by_opcode): Adjust caller, use the predictor.
5211 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5213 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5215 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5216 proper constant for the store mode.
5218 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5220 * symtab.c (change_decl_assembler_name): Fix transparent alias
5223 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5225 * config/aarch64/aarch64.c: Correct the comments about the
5226 aarch64 stack layout.
5228 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5230 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5231 check for GF_OMP_FOR_KIND_FOR.
5233 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5235 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5236 ymm and zmm register names.
5238 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5241 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5242 note creation for the 2010-08-31 changes.
5244 2014-03-17 Marek Polacek <polacek@redhat.com>
5247 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5248 as -fno-tree-loop-vectorize.
5249 (expand_omp_simd): Likewise.
5251 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5253 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5254 (eligible_for_call_delay): New prototype.
5255 * config/sparc/sparc.c (tls_call_delay): Rename into...
5256 (eligible_for_call_delay): ...this. Return false if the instruction
5257 cannot be put in the delay slot of a branch.
5258 (eligible_for_restore_insn): Simplify.
5259 (eligible_for_return_delay): Return false if the instruction cannot be
5260 put in the delay slot of a branch and simplify.
5261 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5262 put in the delay slot of a branch.
5263 * config/sparc/sparc.md (fix_ut699): New attribute.
5264 (tls_call_delay): Delete.
5265 (in_call_delay): Reimplement.
5266 (eligible_for_sibcall_delay): Rename into...
5267 (in_sibcall_delay): ...this.
5268 (eligible_for_return_delay): Rename into...
5269 (in_return_delay): ...this.
5270 (in_branch_delay): Reimplement.
5271 (in_uncond_branch_delay): Delete.
5272 (in_annul_branch_delay): Delete.
5274 2014-03-14 Richard Henderson <rth@redhat.com>
5277 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5278 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5279 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5280 (floathi splitters): Remove.
5281 (float<SWI48x>xf2): New pattern.
5282 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5283 code that tried to handle DImode for 32-bit, but which was excluded
5284 by the pattern's condition. Drop allocation of stack temporary.
5285 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5286 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5287 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5288 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5289 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5290 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5291 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5292 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5293 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5294 (*float<SWI48x><X87MODEF>2_i387): Remove.
5295 (all float _with_temp splitters): Remove.
5296 (*float<SWI48x><MODEF>2_i387): New pattern.
5297 (*float<SWI48><MODEF>2_sse): New pattern.
5298 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5299 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5301 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5302 Marek Polacek <polacek@redhat.com>
5305 * common.opt (dump_base_name_prefixed): New Variable.
5306 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5307 if x_dump_base_name_prefixed is already set, set it at the end.
5309 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5311 PR rtl-optimization/60508
5312 * lra-constraints.c (get_reload_reg): Add new parameter
5314 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5315 Pass the new parameter values.
5317 2014-03-14 Richard Biener <rguenther@suse.de>
5319 * common.opt: Revert unintented changes from r205065.
5322 2014-03-14 Richard Biener <rguenther@suse.de>
5325 * cfghooks.c (split_block): Properly adjust all loops the
5326 block was a latch of.
5328 2014-03-14 Martin Jambor <mjambor@suse.cz>
5331 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5334 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5337 * config/avr/avr.c (avr_set_current_function): Pass function name
5338 through default_strip_name_encoding before sanity checking instead
5339 of skipping the first char of the assembler name.
5341 2014-03-13 Richard Henderson <rth@redhat.com>
5344 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5345 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5346 * config/i386/i386-protos.h: Likewise.
5347 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5348 in the expander instead of a splitter.
5349 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5350 any possibility of requiring a memory.
5351 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5352 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5353 (fp branch splitters): Update for ix86_split_fp_branch.
5354 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5355 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5356 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5357 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5358 (*fop_<MODEF>_3_i387): Likewise.
5359 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5360 (splitters for the fop_* register patterns): Remove.
5361 (fscalexf4_i387): Rename from *fscalexf4_i387.
5362 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5364 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5366 PR tree-optimization/59779
5367 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5368 type for bitsize and maxsize instead of HOST_WIDE_INT.
5370 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5372 PR rtl-optimization/57320
5373 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5374 the CFG after thread_prologue_and_epilogue_insns.
5376 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5378 PR rtl-optimization/57189
5379 * lra-constraints.c (process_alt_operands): Disfavor spilling
5382 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5384 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5386 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5388 PR tree-optimization/59025
5390 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5391 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5393 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5396 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5397 calls of avr_out_plus_1.
5399 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5401 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5402 BB's single pred and update the father loop's latch info later.
5404 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5406 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5410 (VEC_base): Likewise.
5411 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5412 registers, we need to swap double words in little endian mode.
5414 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5415 to be a container mode for 128-bit integer operations added in ISA
5416 2.07. Unlike TImode and PTImode, the preferred register set is
5417 the Altivec/VMX registers for the 128-bit operations.
5419 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5421 (rs6000_split_128bit_ok_p): Likewise.
5423 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5424 macros for creating ISA 2.07 normal and overloaded builtin
5425 functions with 3 arguments.
5426 (BU_P8V_OVERLOAD_3): Likewise.
5427 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5428 for use as overloaded functions.
5429 (VPERM_1TI_UNS): Likewise.
5430 (VSEL_1TI): Likewise.
5431 (VSEL_1TI_UNS): Likewise.
5432 (ST_INTERNAL_1ti): Likewise.
5433 (LD_INTERNAL_1ti): Likewise.
5434 (XXSEL_1TI): Likewise.
5435 (XXSEL_1TI_UNS): Likewise.
5436 (VPERM_1TI): Likewise.
5437 (VPERM_1TI_UNS): Likewise.
5438 (XXPERMDI_1TI): Likewise.
5439 (SET_1TI): Likewise.
5440 (LXVD2X_V1TI): Likewise.
5441 (STXVD2X_V1TI): Likewise.
5442 (VEC_INIT_V1TI): Likewise.
5443 (VEC_SET_V1TI): Likewise.
5444 (VEC_EXT_V1TI): Likewise.
5445 (EQV_V1TI): Likewise.
5446 (NAND_V1TI): Likewise.
5447 (ORC_V1TI): Likewise.
5448 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5449 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5451 (VADDUQM): Likewise.
5452 (VSUBCUQ): Likewise.
5453 (VADDEUQM): Likewise.
5454 (VADDECUQ): Likewise.
5455 (VSUBEUQM): Likewise.
5456 (VSUBECUQ): Likewise.
5458 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5459 __int128_t and __uint128_t types.
5460 (__uint128_type): Likewise.
5461 (altivec_categorize_keyword): Add support for vector __int128_t,
5462 vector __uint128_t, vector __int128, and vector unsigned __int128
5463 as a container type for TImode operations that need to be done in
5464 VSX/Altivec registers.
5465 (rs6000_macro_to_expand): Likewise.
5466 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5467 to support 128-bit integer instructions vaddcuq, vadduqm,
5468 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5469 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5471 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5472 for V1TImode, and set up preferences to use VSX/Altivec registers.
5473 Setup VSX reload handlers.
5474 (rs6000_debug_reg_global): Likewise.
5475 (rs6000_init_hard_regno_mode_ok): Likewise.
5476 (rs6000_preferred_simd_mode): Likewise.
5477 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5478 (easy_altivec_constant): Likewise.
5479 (output_vec_const_move): Likewise.
5480 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5482 (rs6000_expand_vector_extract): Likewise.
5483 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5485 (rs6000_const_vec): Add support for V1TImode.
5486 (rs6000_emit_le_vsx_load): Swap double words when loading or
5487 storing TImode/V1TImode.
5488 (rs6000_emit_le_vsx_store): Likewise.
5489 (rs6000_emit_le_vsx_move): Likewise.
5490 (rs6000_emit_move): Add support for V1TImode.
5491 (altivec_expand_ld_builtin): Likewise.
5492 (altivec_expand_st_builtin): Likewise.
5493 (altivec_expand_vec_init_builtin): Likewise.
5494 (altivec_expand_builtin): Likewise.
5495 (rs6000_init_builtins): Add support for V1TImode type. Add
5496 support for ISA 2.07 128-bit integer builtins. Define type names
5497 for the VSX/Altivec vector types.
5498 (altivec_init_builtins): Add support for overloaded vector
5499 functions with V1TImode type.
5500 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5501 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5503 (rs6000_split_128bit_ok_p): Likewise.
5504 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5505 __int128_t and vector __uint128_t.
5507 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5508 and mode attributes.
5515 (VS_scalar): Likewise.
5516 (VS_double): Likewise.
5517 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5519 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5520 we support the ISA 2.07 128-bit integer arithmetic instructions.
5521 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5522 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5523 and TImode types for use with the builtin functions.
5524 (V1TI_type_node): Likewise.
5525 (unsigned_V1TI_type_node): Likewise.
5526 (intTI_type_internal_node): Likewise.
5527 (uintTI_type_internal_node): Likewise.
5529 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5530 128-bit builtin functions.
5531 (UNSPEC_VADDEUQM): Likewise.
5532 (UNSPEC_VADDECUQ): Likewise.
5533 (UNSPEC_VSUBCUQ): Likewise.
5534 (UNSPEC_VSUBEUQM): Likewise.
5535 (UNSPEC_VSUBECUQ): Likewise.
5536 (VM): Add V1TImode to vector mode iterators.
5538 (VI_unit): Likewise.
5539 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5540 (altivec_vaddcuq): Likewise.
5541 (altivec_vsubuqm): Likewise.
5542 (altivec_vsubcuq): Likewise.
5543 (altivec_vaddeuqm): Likewise.
5544 (altivec_vaddecuq): Likewise.
5545 (altivec_vsubeuqm): Likewise.
5546 (altivec_vsubecuq): Likewise.
5548 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5550 (BOOL_128): Likewise.
5551 (BOOL_REGS_OUTPUT): Likewise.
5552 (BOOL_REGS_OP1): Likewise.
5553 (BOOL_REGS_OP2): Likewise.
5554 (BOOL_REGS_UNARY): Likewise.
5555 (BOOL_REGS_AND_CR0): Likewise.
5557 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5558 128-bit integer builtin support.
5559 (vec_vadduqm): Likewise.
5560 (vec_vaddecuq): Likewise.
5561 (vec_vaddeuqm): Likewise.
5562 (vec_vsubecuq): Likewise.
5563 (vec_vsubeuqm): Likewise.
5564 (vec_vsubcuq): Likewise.
5565 (vec_vsubuqm): Likewise.
5567 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5568 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5569 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5570 128-bit integer add/subtract to ISA 2.07.
5572 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5574 * config/arc/arc.c (arc_predicate_delay_insns):
5575 Fix third argument passed to conditionalize_nonjump.
5577 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5579 * config/aarch64/aarch64-builtins.c
5580 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5581 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5582 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5583 instead of __builtin_lfloor.
5584 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5586 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5588 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5589 (tree_ssa_ifcombine_bb_1): New function.
5590 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5591 is an empty forwarder block to then_bb or vice versa and then_bb
5592 and else_bb are effectively swapped.
5594 2014-03-12 Christian Bruel <christian.bruel@st.com>
5597 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5598 REG_CFA_DEF_CFA note.
5599 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5600 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5602 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5604 PR tree-optimization/60454
5605 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5607 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5609 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5610 Do not define target_cpu_default2 to generic.
5611 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5612 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5613 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5615 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5616 Marc Glisse <marc.glisse@inria.fr>
5618 PR tree-optimization/60502
5619 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5620 instead of build_low_bits_mask.
5622 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5625 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5626 if there are multiple uses, but op doesn't live on E edge.
5627 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5628 clobber stmts before __builtin_unreachable.
5630 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5632 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5633 hard_frame_pointer_rtx.
5634 * cse.c (cse_insn): Remove volatile check.
5635 * cselib.c (cselib_process_insn): Likewise.
5636 * dse.c (scan_insn): Likewise.
5638 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5640 * config/arc/arc.c (conditionalize_nonjump): New function,
5642 (arc_ifcvt): ... this.
5643 (arc_predicate_delay_insns): Use it.
5645 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5647 * config/arc/predicates.md (extend_operand): During/after reload,
5648 allow const_int_operand.
5649 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5650 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5651 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5653 (umulsi3_highpart_i): Likewise.
5655 2014-03-11 Richard Biener <rguenther@suse.de>
5657 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5658 Add asserts to guard possible wrong-code bugs.
5660 2014-03-11 Richard Biener <rguenther@suse.de>
5662 PR tree-optimization/60429
5663 PR tree-optimization/60485
5664 * tree-ssa-structalias.c (set_union_with_increment): Properly
5665 take into account all fields that overlap the shifted vars.
5666 (do_sd_constraint): Likewise.
5667 (do_ds_constraint): Likewise.
5668 (get_constraint_for_ptr_offset): Likewise.
5670 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5672 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5673 (nios2_compute_frame_layout):
5674 Add calculation of cfun->machine->fp_save_offset.
5675 (nios2_expand_prologue): Correct setting of frame pointer register
5677 (nios2_expand_epilogue): Update recovery of stack pointer from
5678 frame pointer accordingly.
5679 (nios2_initial_elimination_offset): Update calculation of offset
5680 for eliminating to HARD_FRAME_POINTER_REGNUM.
5682 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5685 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5686 cgraph_get_create_node on VAR_DECLs.
5688 2014-03-10 Richard Biener <rguenther@suse.de>
5691 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5693 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5695 * config/vms/vms.opt (vms_float_format): New variable.
5697 2014-03-08 Tobias Burnus <burnus@net-b.de>
5699 * doc/invoke.texi (-fcilkplus): Update implementation status.
5701 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5702 Richard Biener <rguenther@suse.de>
5704 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5705 consistently accross all TUs.
5706 (run_gcc): Enable -fshort-double automatically at link at link-time
5707 and disallow override.
5709 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5712 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5713 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5714 if they can't be used.
5716 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5718 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5719 for Solaris 11/x86 ld.
5720 * configure: Regenerate.
5722 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5724 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5725 (LIB_TLS_SPEC): Save as ld_tls_libs.
5726 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5727 (HAVE_AS_IX86_TLSLDM): New test.
5728 * configure, config.in: Regenerate.
5729 * config/i386/i386.c (legitimize_tls_address): Fall back to
5730 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5731 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5732 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5733 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5735 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5737 * common.opt (fira-loop-pressure): Mark as optimization.
5739 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5741 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5742 an OpenMP mappable type.
5744 2014-03-06 Matthias Klose <doko@ubuntu.com>
5746 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5747 MULTILIB_OSDIRNAMES is not defined.
5749 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5750 Meador Inge <meadori@codesourcery.com>
5753 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5754 (arm_legitimize_address): Call legitimize_tls_address for any
5755 arm_tls_referenced_p expression, handle constant addend. Call it
5756 before testing for !TARGET_ARM.
5757 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5759 2014-03-06 Richard Biener <rguenther@suse.de>
5765 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5767 * tree-streamer.c (record_common_node): Assert we don't record
5768 nodes with type double.
5769 (preload_common_node): Skip type double, complex double and double
5770 pointer since it is now frontend dependent due to fshort-double option.
5772 2014-03-06 Richard Biener <rguenther@suse.de>
5774 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5775 or -fno-lto is specified and the linker has full plugin support.
5776 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5777 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5778 * lto-wrapper.c (merge_and_complain): Merge compile-time
5779 optimization levels.
5780 (run_gcc): And pass it through to the link options.
5782 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5786 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5788 * cselib.c (remove_useless_values): Skip to avoid quadratic
5789 behavior if the condition moved from...
5790 (cselib_process_insn): ... here holds.
5792 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5795 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5796 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5799 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5800 (TM_H): Add x86-tune.def.
5802 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5804 * config/aarch64/aarch64.c (generic_tunings):
5805 Use cortexa57_extra_costs.
5807 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5810 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5811 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5812 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5815 2014-03-04 Heiher <r@hev.cc>
5817 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5818 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5820 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5822 * config/i386/predicates.md (const2356_operand): Change to ...
5823 (const2367_operand): ... this.
5824 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5826 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5827 (*avx512pf_scatterpf<mode>sf): Ditto.
5828 (avx512pf_scatterpf<mode>df): Ditto.
5829 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5830 (*avx512pf_scatterpf<mode>df): Ditto.
5831 * config/i386/i386.c (ix86_expand_builtin): Update
5832 incorrect hint operand error message.
5834 2014-03-04 Richard Biener <rguenther@suse.de>
5836 * lto-section-in.c (lto_get_section_data): Fix const cast.
5838 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5840 * tree-streamer.c (record_common_node): Assert we don't record
5841 nodes with type double.
5842 (preload_common_node): Skip type double, complex double and double
5843 pointer since it is now frontend dependent due to fshort-double option.
5845 2014-03-04 Richard Biener <rguenther@suse.de>
5848 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5849 (lto_input_toplevel_asms): Likewise.
5850 * lto-section-in.c (lto_get_section_data): Instead do it here
5853 2014-03-04 Richard Biener <rguenther@suse.de>
5855 PR tree-optimization/60382
5856 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5857 dead PHIs a reduction.
5859 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5861 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5863 (_mm_prefetch): Move out of GCC target("sse") pragma.
5864 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5865 GCC target("prfchw") pragma.
5866 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5868 * config/i386/i386.c (ix86_option_override_internal): Enable
5869 -mprfchw with -mprefetchwt1.
5871 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5873 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5876 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5878 * opts.h (CL_PCH_IGNORE): Define.
5879 * targhooks.c (option_affects_pch_p):
5880 Return false for options that have CL_PCH_IGNORE set.
5881 * opt-functions.awk: Process PchIgnore.
5882 * doc/options.texi: Document PchIgnore.
5884 * config/arc/arc.opt (misize): Add PchIgnore property.
5886 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5888 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5889 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5890 constraint on constants to permit them being loaded into
5891 GENERAL_REGS or BASE_REGS.
5893 2014-03-03 Nick Clifton <nickc@redhat.com>
5895 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5896 anti-cacnonical alternatives.
5897 (negandhi3_real): New pattern.
5898 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5900 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5902 * config/avr/avr-mcus.def: Remove atxmega16x1.
5903 * config/avr/avr-tables.opt: Regenerate.
5904 * config/avr/t-multilib: Regenerate.
5905 * doc/avr-mmcu.texi: Regenerate.
5907 2014-03-03 Tobias Grosser <tobias@grosser.es>
5908 Mircea Namolaru <mircea.namolaru@inria.fr>
5910 PR tree-optimization/58028
5911 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5914 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5916 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5917 not handled by recognizers.
5919 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5922 * function.c (expand_function_end): Don't emit
5923 clobber_return_register sequence if clobber_after is a BARRIER.
5924 * cfgexpand.c (construct_exit_block): Append instructions before
5925 return_label to prev_bb.
5927 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5929 * config/rs6000/constraints.md: Document reserved use of "wc".
5931 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5934 * ipa.c (function_and_variable_visibility): When dissolving comdat
5935 group, also set all symbols to local.
5937 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5942 2013-12-14 Jan Hubicka <jh@suse.cz>
5944 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5946 2014-03-02 Jon Beniston <jon@beniston.com>
5952 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5953 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5954 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5955 (simple_return, *simple_return): New patterns
5956 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5957 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5959 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5961 * dwarf2out.c (gen_subprogram_die): Tidy.
5963 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5966 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5967 (*mov_t_msb_neg_negc): ... this new insn.
5969 2014-02-28 Jason Merrill <jason@redhat.com>
5972 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5975 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5978 * dwarf2out.c (decltype_auto_die): New static.
5979 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5980 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5981 (is_cxx_auto): Likewise.
5983 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5985 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5986 we are not using general regs only.
5988 2014-02-28 Richard Biener <rguenther@suse.de>
5991 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5992 previous fix and only allow to remove trivial pre-headers
5993 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5994 (remove_forwarder_block): Properly update the latch of a loop.
5996 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5999 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6000 (cselib_preserved_hash_table): New.
6001 (preserve_constants_and_equivs): Move preserved vals to it.
6002 (cselib_find_slot): Look it up first.
6003 (cselib_init): Initialize it.
6004 (cselib_finish): Release it.
6005 (dump_cselib_table): Dump it.
6007 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6010 * cselib.c (remove_useless_values): Skip to avoid quadratic
6011 behavior if the condition moved from...
6012 (cselib_process_insn): ... here holds.
6014 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6017 * var-tracking.c (vt_initialize): Apply the same condition to
6018 preserve the CFA base value.
6020 2014-02-28 Joey Ye <joey.ye@arm.com>
6023 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6024 if reload in progress or completed.
6026 2014-02-28 Tobias Burnus <burnus@net-b.de>
6029 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6032 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6034 * doc/tm.texi.in (Condition Code Status): Update documention for
6035 relative locations of cc0-setter and cc0-user.
6037 2014-02-27 Jeff Law <law@redhat.com>
6039 PR rtl-optimization/52714
6040 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6041 into two independent simple sets, if I3 is a jump, ensure the
6042 pattern we place into I3 is a (set (pc) ...).
6044 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6045 Jeff Law <law@redhat.com>
6047 PR rtl-optimization/49847
6048 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6049 are in different blocks.
6050 * doc/tm.texi (Condition Code Status): Update documention for
6051 relative locations of cc0-setter and cc0-user.
6053 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6056 * lra.c (lra_emit_add): Check SUBREG too.
6058 2014-02-27 Andreas Schwab <schwab@suse.de>
6060 * config/m68k/m68k.c (m68k_option_override): Disable
6061 -flive-range-shrinkage for classic m68k.
6062 (m68k_override_options_after_change): Likewise.
6064 2014-02-27 Marek Polacek <polacek@redhat.com>
6067 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6068 -Wmaybe-uninitialized.
6070 2014-02-27 Alan Modra <amodra@gmail.com>
6073 * reload1.c (emit_input_reload_insns): When reload_override_in,
6074 set old to rl->in_reg when rl->in_reg is a subreg.
6076 2014-02-26 Richard Biener <rguenther@suse.de>
6079 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6081 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6083 * common/config/i386/predicates.md (const1256_operand): Remove.
6084 (const2356_operand): New.
6085 (const_1_to_2_operand): Remove.
6086 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6087 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6088 (*avx512pf_gatherpf<mode>sf): Ditto.
6089 (avx512pf_gatherpf<mode>df): Ditto.
6090 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6091 (*avx512pf_gatherpf<mode>df): Ditto.
6092 (avx512pf_scatterpf<mode>sf): Ditto.
6093 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6094 (*avx512pf_scatterpf<mode>sf): Ditto.
6095 (avx512pf_scatterpf<mode>df): Ditto.
6096 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6097 (*avx512pf_scatterpf<mode>df): Ditto.
6098 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6100 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6102 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6103 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6104 (_mm512_mask_testn_epi64_mask): Move to ...
6105 * config/i386/avx512cdintrin.h: Here.
6106 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6107 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6108 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6109 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6110 TARGET_AVX512F from TARGET_AVX512CD.
6112 2014-02-26 Richard Biener <rguenther@suse.de>
6115 * ipa.c (walk_polymorphic_call_targets): Properly guard
6116 call to inline_update_overall_summary.
6118 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6121 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6122 and latches only if requested. Fix latch if it is removed.
6123 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6124 LOOPS_HAVE_PREHEADERS.
6126 2014-02-25 Andrew Pinski <apinski@cavium.com>
6128 * builtins.c (expand_builtin_thread_pointer): Create a new target
6129 when the target is NULL.
6131 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6133 PR rtl-optimization/60317
6134 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6135 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6136 * lra-assigns.c: Include params.h.
6137 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6138 other reload pseudos considerations.
6140 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6142 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6143 to use canonical form for nor<mode>3.
6145 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6148 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6151 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6153 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6154 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6155 (ix86_handle_option): Handle OPT_mprefetchwt1.
6156 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6157 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6159 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6160 OPTION_MASK_ISA_PREFETCHWT1.
6161 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6162 (PTA_PREFETCHWT1): New.
6163 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6164 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6165 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6166 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6167 (*prefetch_avx512pf_<mode>_: Change into ...
6168 (*prefetch_prefetchwt1_<mode>: This.
6169 * config/i386/i386.opt (mprefetchwt1): New.
6170 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6171 (_mm_prefetch): Handle intent to write.
6172 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6174 2014-02-25 Richard Biener <rguenther@suse.de>
6177 * emit-rtl.c (mem_attrs_htab): Remove.
6178 (mem_attrs_htab_hash): Likewise.
6179 (mem_attrs_htab_eq): Likewise.
6180 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6181 (init_emit_once): Do not allocate mem_attrs_htab.
6183 2014-02-25 Richard Biener <rguenther@suse.de>
6186 * lto-opts.c (lto_write_options): Output non-explicit conservative
6187 -fwrapv, -fno-trapv and -fno-strict-overflow.
6188 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6189 (run_gcc): And pass them through.
6191 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6193 * sel-sched.c (calculate_new_fences): New parameter ptime.
6194 Calculate it as a maximum over all fence cycles.
6195 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6196 Print the final schedule timing when sched_verbose.
6198 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6200 PR rtl-optimization/60292
6201 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6202 bit fot the fence instruction.
6204 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6206 * calls.h: Fix typo in comment.
6208 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6210 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6211 adjusting offsetable addresses.
6213 2014-02-24 Guozhi Wei <carrot@google.com>
6215 * sparseset.h (sparseset_pop): Fix the wrong index.
6217 2014-02-24 Walter Lee <walt@tilera.com>
6219 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6220 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6222 * common/config/tilegx/tilegx-common.c
6223 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6224 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6226 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6227 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6228 (tilegx_gimplify_va_arg_expr): Handle big endian.
6229 (tilegx_expand_unaligned_load): Ditto.
6230 (tilegx_expand_unaligned_store): Ditto.
6231 (TARGET_RETURN_IN_MSB): New.
6232 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6233 (TARGET_ENDIAN_DEFAULT): New.
6234 (TARGET_BIG_ENDIAN): Handle big endian.
6235 (BYTES_BIG_ENDIAN): Ditto.
6236 (WORDS_BIG_ENDIAN): Ditto.
6237 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6240 * config/tilegx/tilegx.md (extv): Handle big endian.
6242 (insn_st<n>): Ditto.
6243 (insn_st<n>_add<bitsuffix>): Ditto.
6244 (insn_stnt<n>): Ditto.
6245 (insn_stnt<n>_add<bitsuffix>):Ditto.
6246 (vec_interleave_highv8qi): Handle big endian.
6247 (vec_interleave_highv8qi_be): New.
6248 (vec_interleave_highv8qi_le): New.
6249 (insn_v1int_h): Handle big endian.
6250 (vec_interleave_lowv8qi): Handle big endian.
6251 (vec_interleave_lowv8qi_be): New.
6252 (vec_interleave_lowv8qi_le): New.
6253 (insn_v1int_l): Handle big endian.
6254 (vec_interleave_highv4hi): Handle big endian.
6255 (vec_interleave_highv4hi_be): New.
6256 (vec_interleave_highv4hi_le): New.
6257 (insn_v2int_h): Handle big endian.
6258 (vec_interleave_lowv4hi): Handle big endian.
6259 (vec_interleave_lowv4hi_be): New.
6260 (vec_interleave_lowv4hi_le): New.
6261 (insn_v2int_l): Handle big endian.
6262 (vec_interleave_highv2si): Handle big endian.
6263 (vec_interleave_highv2si_be): New.
6264 (vec_interleave_highv2si_le): New.
6265 (insn_v4int_h): Handle big endian.
6266 (vec_interleave_lowv2si): Handle big endian.
6267 (vec_interleave_lowv2si_be): New.
6268 (vec_interleave_lowv2si_le): New.
6269 (insn_v4int_l): Handle big endian.
6270 * config/tilegx/tilegx.opt (mbig-endian): New option.
6271 (mlittle-endian): New option.
6272 * doc/install.texi: Document tilegxbe-linux.
6273 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6275 2014-02-24 Martin Jambor <mjambor@suse.cz>
6278 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6279 there are no parameter descriptors.
6281 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6283 PR rtl-optimization/60268
6284 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6285 initialization to ...
6286 (sched_rgn_init): ... here.
6287 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6289 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6291 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6294 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6296 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6299 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6301 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6302 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6304 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6306 * config/microblaze/predicates.md: Add cmp_op predicate.
6307 * config/microblaze/microblaze.md: Add branch_compare instruction
6308 which uses cmp_op predicate and emits cmp insn before branch.
6309 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6310 to microblaze_expand_conditional_branch and consolidate logic.
6311 (microblaze_expand_conditional_branch): emit branch_compare
6312 insn instead of handling cmp op separate from branch insn.
6314 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6316 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6319 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6321 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6322 define_insn with define_expand and new define_insn
6323 *altivec_lve<VI_char>x_internal.
6324 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6325 and new define_insn *altivec_stve<VI_char>x_internal.
6326 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6328 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6330 (altivec_expand_stvex_be): New function.
6332 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6334 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6335 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6336 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6337 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6339 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6342 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6343 instead of emit_move_insn.
6345 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6347 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6349 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6350 gen_altivec_vsumsws.
6352 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6354 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6355 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6356 (altivec_lvxl_<mode>): New define_expand incorporating
6357 -maltivec=be semantics where needed.
6358 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6359 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6360 semantics where needed.
6361 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6362 (altivec_stvx_<mode>): New define_expand incorporating
6363 -maltivec=be semantics where needed.
6364 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6365 VM2 iterator instead of V4SI.
6366 (altivec_stvxl_<mode>): New define_expand incorporating
6367 -maltivec=be semantics where needed.
6368 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6369 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6370 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6371 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6372 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6373 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6374 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6375 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6376 ALTIVEC_BUILTIN_STVXL.
6377 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6378 (altivec_expand_stvx_be): Likewise.
6379 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6380 (altivec_expand_lvx_be): Likewise.
6381 (altivec_expand_stvx_be): Likewise.
6382 (altivec_expand_builtin): Add cases for
6383 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6384 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6385 (altivec_init_builtins): Add definitions for
6386 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6387 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6389 2014-02-21 Catherine Moore <clm@codesourcery.com>
6391 * doc/invoke.texi (mvirt, mno-virt): Document.
6392 * config/mips/mips.opt (mvirt): New option.
6393 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6395 2014-02-21 Richard Biener <rguenther@suse.de>
6397 PR tree-optimization/60276
6398 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6399 (STMT_VINFO_MIN_NEG_DIST): New macro.
6400 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6401 STMT_VINFO_MIN_NEG_DIST.
6402 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6403 made for negative dependence distances still hold.
6405 2014-02-21 Richard Biener <rguenther@suse.de>
6408 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6409 DECL_INITIAL for globals not in the current function context.
6411 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6413 PR tree-optimization/56490
6414 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6415 * tree-ssa-uninit.c: Include params.h.
6416 (compute_control_dep_chain): Add num_calls argument, return false
6417 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6418 num_calls to recursive call.
6419 (find_predicates): Change dep_chain into normal array,
6420 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6421 variable and adjust compute_control_dep_chain caller.
6422 (find_def_preds): Likewise.
6424 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6426 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6427 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6429 2014-02-21 Nick Clifton <nickc@redhat.com>
6431 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6432 (pushhi1): Likewise.
6433 (popqi1): Add mode to pre_dec.
6436 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6438 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6439 mode for mask of V8SFmode permutation.
6441 2014-02-20 Richard Henderson <rth@redhat.com>
6444 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6445 a new pseudo for OLDVAL.
6447 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6450 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6451 gen_reg_rtx if d->testing_p.
6452 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6453 if d->testing_p and we will certainly return true.
6454 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6457 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6459 * emit-rtl.c (gen_reg_rtx): Assert that
6460 crtl->emit.regno_pointer_align_length is non-zero.
6462 2014-02-20 Richard Henderson <rth@redhat.com>
6465 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6466 on failure the store back into EXPECT.
6468 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6469 Sandra Loosemore <sandra@codesourcery.com>
6471 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6472 * config/nios2/nios2.c (nios2_function_profiler): Add
6473 -fPIC (flag_pic == 2) support.
6474 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6475 (nios2_large_offset_p): New function.
6476 (nios2_unspec_reloc_p): Move up position, update to use
6477 nios2_large_offset_p.
6478 (nios2_unspec_address): Remove function.
6479 (nios2_unspec_offset): New function.
6480 (nios2_large_got_address): New function.
6481 (nios2_got_address): Add large offset support.
6482 (nios2_legitimize_tls_address): Update usage of removed and new
6484 (nios2_symbol_binds_local_p): New function.
6485 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6486 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6487 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6488 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6489 processing for (const (unspec ...)).
6490 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6492 2014-02-20 Richard Biener <rguenther@suse.de>
6494 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6495 doing the substitution.
6496 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6498 2014-02-20 Martin Jambor <mjambor@suse.cz>
6501 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6502 info when checking whether lattices are bottom.
6504 2014-02-20 Richard Biener <rguenther@suse.de>
6507 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6510 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6513 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6514 parameter specifying the scaling.
6515 (inline_call): Update.
6516 (want_inline_recursively): Guard division by zero.
6517 (recursive_inlining): Update.
6518 * ipa-inline.h (clone_inlined_nodes): Update.
6520 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6523 * config/i386/i386.c (classify_argument): Pass structures of size
6524 64 bytes or less in register.
6526 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6527 Kirill Yukhin <kirill.yukhin@intel.com>
6529 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6530 (_mm_rcp28_round_ss): Ditto.
6531 (_mm_rsqrt28_round_sd): Ditto.
6532 (_mm_rsqrt28_round_ss): Ditto.
6533 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6534 (_mm_rcp14_round_ss): Ditto.
6535 (_mm_rsqrt14_round_sd): Ditto.
6536 (_mm_rsqrt14_round_ss): Ditto.
6537 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6538 the first input operand, get rid of match_dup.
6539 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6541 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6543 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6544 operand as the first input operand, set type attribute.
6545 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6547 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6548 operand as the first input operand, set type attribute.
6550 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6552 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6555 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6558 * config/i386/i386.c (construct_container): Remove TFmode check
6559 for X86_64_INTEGER_CLASS.
6561 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6564 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6565 only when -Wpsabi is enabled.
6567 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6570 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6571 passing arrays in registers are the same as for structs, so remove the
6572 special case for them.
6574 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6576 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6577 destination type, extract only the valid bits if the source type is not
6578 integral and has a different mode.
6580 2014-02-19 Richard Biener <rguenther@suse.de>
6583 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6586 2014-02-19 Richard Biener <rguenther@suse.de>
6589 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6590 (ipa_modify_call_arguments): Emit an argument load explicitely and
6591 preserve virtual SSA form there and for the replacement call.
6592 Do not update SSA form nor free dominance info.
6594 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6596 * ipa.c (function_and_variable_visibility): Also clear WEAK
6597 flag when disolving COMDAT_GROUP.
6599 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6601 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6602 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6604 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6605 do more sanity checks.
6606 (detect_type_change): Return true when giving up early.
6607 (compute_complex_assign_jump_func): Fix type parameter of
6608 ipa_set_ancestor_jf.
6609 (compute_complex_ancestor_jump_func): Likewise.
6610 (update_jump_functions_after_inlining): Fix updating of
6612 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6614 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6616 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6617 inline clones when edge disappears.
6619 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6622 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6623 Split 64-bit moves into 2 patterns. Do not allow the use of
6624 direct move for TDmode in little endian, since the decimal value
6625 has little endian bytes within a word, but the 64-bit pieces are
6626 ordered in a big endian fashion, and normal subreg's of TDmode are
6628 (mov<mode>_64bit_dm): Likewise.
6629 (movtd_64bit_nodm): Likewise.
6631 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6633 PR tree-optimization/60174
6634 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6635 statement of an SSA_NAME that occurs in an abnormal PHI node.
6637 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6640 * final.c (SEEN_BB): Remove.
6641 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6642 (final_scan_insn): Don't force_source_line on second
6643 NOTE_INSN_BASIC_BLOCK.
6645 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6648 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6649 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6650 (type_natural_mode): Warn ABI change when %zmm register is not
6651 available for AVX512F vector value passing.
6653 2014-02-18 Kai Tietz <ktietz@redhat.com>
6656 * config/i386/i386.c (ix86_expand_prologue): Use value in
6657 rax register as displacement when restoring %r10 or %rax.
6658 Fix wrong offset when restoring both registers.
6660 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6662 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6663 assertion with conditional return.
6665 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6666 Uros Bizjak <ubizjak@gmail.com>
6669 * config/i386/driver-i386.c (host_detect_local_cpu): If
6670 YMM state is not saved by the OS, also clear has_f16c. Move
6671 CPUID 0x80000001 handling before YMM state saving checking.
6673 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6675 PR rtl-optimization/58960
6676 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6677 factored out from ...
6678 (sched_init): ... here.
6679 (free_global_sched_pressure_data): New, factored out from ...
6680 (sched_finish): ... here.
6681 * sched-int.h (free_global_sched_pressure_data): Declare.
6682 * sched-rgn.c (nr_regions_initial): New static global.
6683 (haifa_find_rgns): Initialize it.
6684 (schedule_region): Disable sched-pressure for the newly
6687 2014-02-17 Richard Biener <rguenther@suse.de>
6689 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6690 release SSA defs of pattern stmts.
6692 2014-02-17 Richard Biener <rguenther@suse.de>
6694 * tree-inline.c (expand_call_inline): Release the virtual
6695 operand defined by the call we are about to inline.
6697 2014-02-17 Richard Biener <rguenther@suse.de>
6699 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6701 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6702 Ilya Tocar <ilya.tocar@intel.com>
6704 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6705 arguments order in builtin.
6706 (_mm512_permutexvar_epi64): Ditto.
6707 (_mm512_mask_permutexvar_epi64): Ditto
6708 (_mm512_maskz_permutexvar_epi32): Ditto
6709 (_mm512_permutexvar_epi32): Ditto
6710 (_mm512_mask_permutexvar_epi32): Ditto
6712 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6714 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6715 (p8_vmrgow): Likewise.
6717 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6719 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6722 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6725 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6726 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6727 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6728 using direct move instructions on ISA 2.07. Also adjust
6729 instruction length for 64-bit.
6730 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6731 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6733 2014-02-15 Alan Modra <amodra@gmail.com>
6737 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6738 find_replacement on parts of insn rtl that might be reloaded.
6740 2014-02-15 Richard Biener <rguenther@suse.de>
6742 PR tree-optimization/60183
6743 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6744 (tree_ssa_phiprop): Calculate and free post-dominators.
6746 2014-02-14 Jeff Law <law@redhat.com>
6748 PR rtl-optimization/60131
6749 * ree.c (get_extended_src_reg): New function.
6750 (combine_reaching_defs): Use it rather than assuming location of REG.
6751 (find_and_remove_re): Verify first operand of extension is
6752 a REG before adding the insns to the copy list.
6754 2014-02-14 Roland McGrath <mcgrathr@google.com>
6756 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6757 * configure: Regenerated.
6758 * config.in: Regenerated.
6759 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6760 instead of ASM_SHORT.
6762 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6763 Richard Earnshaw <rearnsha@arm.com>
6765 PR rtl-optimization/59535
6766 * lra-constraints.c (process_alt_operands): Encourage alternative
6767 when unassigned pseudo class is superset of the alternative class.
6768 (inherit_reload_reg): Don't inherit when optimizing for code size.
6769 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6770 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6771 modes not less than 4 for Thumb1.
6773 2014-02-14 Kyle McMartin <kyle@redhat.com>
6776 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6778 2014-02-14 Richard Biener <rguenther@suse.de>
6780 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6781 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6782 Do not strip INDIRECT_REFs.
6784 2014-02-14 Richard Biener <rguenther@suse.de>
6787 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6788 DECL_FUNCTION_SPECIFIC_TARGET.
6789 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6790 * tree-streamer-out.c (pack_ts_target_option): Remove.
6791 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6792 (write_ts_function_decl_tree_pointers): Do not stream
6793 DECL_FUNCTION_SPECIFIC_TARGET.
6794 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6795 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6796 (lto_input_ts_function_decl_tree_pointers): Do not stream
6797 DECL_FUNCTION_SPECIFIC_TARGET.
6799 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6801 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6802 (get_initial_def_for_induction, vectorizable_induction): Ignore
6803 debug stmts when looking for exit_phi.
6804 (vectorizable_live_operation): Fix up condition.
6806 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6808 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6809 nreverse() because it changes the content of original tree list.
6811 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6813 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6814 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6816 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6818 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6819 GNU coding standards.
6821 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6824 * dwarf2out.c (gen_subprogram_die): Don't call
6825 add_calling_convention_attribute if subr_die is old_die.
6827 2014-02-13 Sharad Singhai <singhai@google.com>
6829 * doc/optinfo.texi: Fix order of nodes.
6831 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6833 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6834 operands[2], not operands[3].
6836 2014-02-13 Richard Biener <rguenther@suse.de>
6839 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6840 mention the possibility of an in-tree build.
6841 (CLooG): Update recommended version to 0.18.1, mention the
6842 possibility of an in-tree build and clarify that the ISL
6843 bundled with CLooG does not work.
6845 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6848 * expr.c (compress_float_constant): If x is a hard register,
6849 extend into a pseudo and then move to x.
6851 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6853 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6854 caused by bad second argument to warning_at() with -mhotpatch and
6855 nested functions (e.g. with gfortran).
6857 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6859 * opts.c (option_name): Remove "enabled by default" rider.
6861 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6863 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6865 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6866 Uros Bizjak <ubizjak@gmail.com>
6869 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6870 * configure: Regenerated.
6872 2014-02-12 Richard Biener <rguenther@suse.de>
6874 * vec.c (vec_prefix::calculate_allocation): Move as
6875 inline variant to vec.h.
6876 (vec_prefix::calculate_allocation_1): New out-of-line version.
6877 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6878 (vec_prefix::m_has_auto_buf): Rename to ...
6879 (vec_prefix::m_using_auto_storage): ... this.
6880 (vec_prefix::calculate_allocation): Inline the easy cases
6881 and dispatch to calculate_allocation_1 which doesn't need the
6883 (va_heap::reserve): Use gcc_checking_assert.
6884 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6885 m_using_auto_storage.
6886 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6888 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6889 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6890 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6892 2014-02-12 Richard Biener <rguenther@suse.de>
6894 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6895 when we found a dependence.
6897 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6899 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6901 (maybe_fold_stmt): ... into this new function.
6902 * omp-low.c (lower_omp): Update comment.
6904 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6907 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6910 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6912 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6913 identifiers in comments.
6914 (cortexa53_extra_costs): Likewise.
6915 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6916 (cortexa7_extra_costs): Likewise.
6917 (cortexa12_extra_costs): Likewise.
6918 (cortexa15_extra_costs): Likewise.
6919 (v7m_extra_costs): Likewise.
6921 2014-02-12 Richard Biener <rguenther@suse.de>
6924 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6925 of posix_memalign being successful.
6926 (lower_stmt): Restrict lowering of posix_memalign to when
6927 -ftree-bit-ccp is enabled.
6929 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6931 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6933 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6935 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6937 PR rtl-optimization/60116
6938 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6939 other_insn once the combination has been validated.
6941 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6944 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6946 * ipa-devirt.c: Include demangle.h
6947 (odr_violation_reported): New static variable.
6948 (add_type_duplicate): Update odr_violations.
6949 (maybe_record_node): Add completep parameter; update it.
6950 (record_target_from_binfo): Add COMPLETEP parameter;
6951 update it as needed.
6952 (possible_polymorphic_call_targets_1): Likewise.
6953 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6954 rename FINAL to COMPLETE.
6955 (record_targets_from_bases): Sanity check we found the binfo;
6956 fix COMPLETEP updating.
6957 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6958 parameter, fix computing of COMPLETEP.
6959 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6960 at LTO time do demangling.
6961 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6962 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6964 (gimple_get_virt_method_for_binfo): Likewise.
6965 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6966 gimple_get_virt_method_for_vtable): Update prototypes.
6968 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6971 * genautomata.c (add_presence_absence): Fix typo with
6972 {final_}presence_list.
6974 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6977 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6978 for VSX/Altivec vectors that land in GPR registers.
6980 2014-02-11 Richard Henderson <rth@redhat.com>
6981 Jakub Jelinek <jakub@redhat.com>
6984 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6985 around drhs if type conversion to lacc->type is not useless.
6987 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6989 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6991 (cortex-a57.cortex-a53): Likewise.
6992 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6994 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6996 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6999 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7001 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7002 add_options_for_arm_vfp3.
7004 2014-02-11 Jeff Law <law@redhat.com>
7007 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7008 object with an undesirable mode.
7010 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7013 * config/i386/sol2-9.h: New file.
7014 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7015 *-*-solaris2.9*): Use it.
7017 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7019 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7020 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7022 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7024 * config/microblaze/microblaze.c: Extend mcpu version format
7026 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7028 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7030 2014-02-10 Richard Henderson <rth@redhat.com>
7033 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7034 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7035 ms-abi vs -mno-accumulate-outgoing-args.
7036 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7037 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7040 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7043 * cfgexpand.c (expand_asm_operands): Attach source location to
7044 ASM_INPUT rtx objects.
7045 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7047 2014-02-10 Nick Clifton <nickc@redhat.com>
7049 * config/mn10300/mn10300.c (popcount): New function.
7050 (mn10300_expand_prologue): Include saved registers in stack usage
7053 2014-02-10 Jeff Law <law@redhat.com>
7056 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7057 when changing the SET_DEST of a prior insn to avoid an input reload.
7059 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7061 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7062 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7063 -mcall-openbsd, or -mcall-linux.
7064 (CC1_ENDIAN_BIG_SPEC): Remove.
7065 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7066 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7067 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7068 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7069 and %cc1_endian_default.
7070 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7072 2014-02-10 Richard Biener <rguenther@suse.de>
7074 PR tree-optimization/60115
7075 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7076 MEM_REF handling. Properly verify that the accesses are not
7077 out of the objects bound.
7079 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7081 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7084 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7086 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7087 proper constants and fix formatting.
7088 (possible_polymorphic_call_targets): Fix formatting.
7090 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7091 Ilya Tocar <ilya.tocar@intel.com>
7093 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7094 (_mm512_loadu_epi32): Renamed into...
7095 (_mm512_loadu_si512): This.
7096 (_mm512_storeu_epi32): Renamed into...
7097 (_mm512_storeu_si512): This.
7098 (_mm512_maskz_ceil_ps): Removed.
7099 (_mm512_maskz_ceil_pd): Ditto.
7100 (_mm512_maskz_floor_ps): Ditto.
7101 (_mm512_maskz_floor_pd): Ditto.
7102 (_mm512_floor_round_ps): Ditto.
7103 (_mm512_floor_round_pd): Ditto.
7104 (_mm512_ceil_round_ps): Ditto.
7105 (_mm512_ceil_round_pd): Ditto.
7106 (_mm512_mask_floor_round_ps): Ditto.
7107 (_mm512_mask_floor_round_pd): Ditto.
7108 (_mm512_mask_ceil_round_ps): Ditto.
7109 (_mm512_mask_ceil_round_pd): Ditto.
7110 (_mm512_maskz_floor_round_ps): Ditto.
7111 (_mm512_maskz_floor_round_pd): Ditto.
7112 (_mm512_maskz_ceil_round_ps): Ditto.
7113 (_mm512_maskz_ceil_round_pd): Ditto.
7114 (_mm512_expand_pd): Ditto.
7115 (_mm512_expand_ps): Ditto.
7116 * config/i386/i386.c (ix86_builtins): Remove
7117 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7118 (bdesc_args): Ditto.
7119 * config/i386/predicates.md (const1256_operand): New.
7120 (const_1_to_2_operand): Ditto.
7121 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7122 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7123 (*avx512pf_gatherpf<mode>sf): Ditto.
7124 (avx512pf_gatherpf<mode>df): Ditto.
7125 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7126 (*avx512pf_gatherpf<mode>df): Ditto.
7127 (avx512pf_scatterpf<mode>sf): Ditto.
7128 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7129 (*avx512pf_scatterpf<mode>sf): Ditto.
7130 (avx512pf_scatterpf<mode>df): Ditto.
7131 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7132 (*avx512pf_scatterpf<mode>df): Ditto.
7133 (avx512f_expand<mode>): Removed.
7134 (<shift_insn><mode>3<mask_name>): Change predicate type.
7136 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7138 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7139 not at the end of datarefs vector use ordered_remove to avoid
7140 reordering datarefs vector.
7143 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7144 mark local addressable non-static vars as GOVD_PRIVATE
7145 instead of GOVD_LOCAL.
7146 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7147 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7151 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7152 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7153 assume_aligned or alloc_align attributes.
7154 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7155 arguments. Handle also assume_aligned and alloc_align attributes.
7156 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7157 calls to functions with assume_aligned or alloc_align attributes.
7158 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7160 2014-02-08 Terry Guo <terry.guo@arm.com>
7162 * doc/invoke.texi: Document ARM -march=armv7e-m.
7164 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7166 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7167 flag on __cilkrts_rethrow builtin.
7170 * ipa-cp.c (determine_versionability): Fail at -O0
7171 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7172 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7175 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7178 * tree-inline.c (copy_forbidden): Fail for
7179 __attribute__((optimize (0))) functions.
7181 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7183 * varpool.c: Include pointer-set.h.
7184 (varpool_remove_unreferenced_decls): Variables in other partitions
7185 will not be output; be however careful to not lose information
7188 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7190 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7191 lookup in the vtable constructor.
7193 2014-02-07 Jeff Law <law@redhat.com>
7196 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7197 define_insn_and_split.
7199 * ipa-inline.c (inline_small_functions): Fix typos.
7201 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7203 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7204 (s390_can_use_return_insn): Declare.
7205 * config/s390/s390.h (EPILOGUE_USES): Define.
7206 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7208 (s390_chunkify_start): Handle return JUMP_LABELs.
7209 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7210 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7211 (s390_can_use_return_insn): New functions.
7212 (s390_fix_long_loop_prediction): Handle conditional returns.
7213 (TARGET_SET_UP_BY_PROLOGUE): Define.
7214 * config/s390/s390.md (ANY_RETURN): New code iterator.
7215 (*creturn, *csimple_return, return, simple_return): New patterns.
7217 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7219 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7220 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7221 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7222 REG_CFA_RESTORE list when deciding not to restore a register.
7224 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7226 * config/s390/s390.c: Include tree-pass.h and context.h.
7227 (s390_early_mach): New function, split out from...
7228 (s390_emit_prologue): ...here.
7229 (pass_data_s390_early_mach): New pass structure.
7230 (pass_s390_early_mach): New class.
7231 (s390_option_override): Create and register early_mach pass.
7232 Move to end of file.
7234 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7236 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7237 to match for the exit block.
7239 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7241 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7242 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7243 Reject misaligned operands.
7245 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7247 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7249 2014-02-07 Richard Biener <rguenther@suse.de>
7252 * gimple-low.c (lower_builtin_posix_memalign): New function.
7253 (lower_stmt): Call it to lower posix_memalign in a way
7254 to make alignment info accessible.
7256 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7259 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7260 __builtin_setjmp_receiver.
7262 2014-02-07 Richard Biener <rguenther@suse.de>
7265 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7266 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7267 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7268 Handle BUILT_IN_POSIX_MEMALIGN.
7269 (find_func_clobbers): Likewise.
7270 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7271 (call_may_clobber_ref_p_1): Likewise.
7273 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7276 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7279 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7282 * lto-cgraph.c (lto_output_node): Use
7283 symtab_get_symbol_partitioning_class.
7284 (lto_output_varpool_node): likewise.
7285 (symtab_get_symbol_partitioning_class): Move here from
7287 * cgraph.h (symbol_partitioning_class): Likewise.
7288 (symtab_get_symbol_partitioning_class): Declare.
7290 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7292 * ggc.h (ggc_internal_cleared_alloc): New macro.
7293 * vec.h (vec_safe_copy): Handle memory stats.
7294 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7295 * target-globals.c (save_target_globals): Likewise.
7297 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7300 * expr.c (emit_move_resolve_push): Export; be bit more selective
7301 on when to clear alias set.
7302 * expr.h (emit_move_resolve_push): Declare.
7303 * function.h (struct function): Add tail_call_marked.
7304 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7305 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7306 * config/i386/i386.md (TImode move expander): De not call
7308 (FP push expanders): Preserve memory attributes.
7309 * config/i386/sse.md (push<mode>1): Remove.
7310 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7311 (ix86_expand_push): Remove.
7312 * config/i386/mmx.md (push<mode>1): Remove.
7314 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7316 PR rtl-optimization/60030
7317 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7318 lopart with paradoxical subreg before shifting it up by hprec.
7320 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7322 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7323 Remove extra newline at end of file.
7324 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7325 (arm_issue_rate): Handle cortexa57.
7326 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7327 (cortex-a57.cortex-a53): Likewise.
7329 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7332 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7333 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7335 (arm_expand_prologue): Adjust all callers.
7336 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7337 info, registers also at the lowest numbered registers side. Use
7338 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7342 * var-tracking.c (adjust_mems): Before adding a SET to
7343 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7345 2014-02-06 Alan Modra <amodra@gmail.com>
7348 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7349 change SDmode to DDmode when lra_in_progress.
7351 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7354 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7355 free_data_ref on the dr first, and before goto again also set dr
7356 to the next dr. For simd_lane_access, free old datarefs[i] before
7357 overwriting it. For get_vectype_for_scalar_type failure, don't
7358 free_data_ref if simd_lane_access.
7360 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7363 * tree.h (opts_for_fn): New inline function.
7364 (opt_for_fn): Define.
7365 * config/i386/i386.c (ix86_function_regparm): Use
7366 opt_for_fn (decl, optimize) instead of optimize.
7368 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7370 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7371 for SYMBOL_REF in large memory model.
7373 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7375 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7377 (cortex-a57): Likewise.
7378 (cortex-a57.cortex-a53): Likewise.
7380 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7381 Kugan Vivekanandarajah <kuganv@linaro.org>
7383 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7385 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7387 2014-02-06 Richard Biener <rguenther@suse.de>
7389 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7390 set_loop_copy and initialize_original_copy_tables.
7392 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7394 * config/aarch64/aarch64-simd.md
7395 (aarch64_ashr_simddi): Change QI to SI.
7397 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7398 Jakub Jelinek <jakub@redhat.com>
7401 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7404 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7406 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7407 CODE_FOR_altivec_vpku[hw]um to
7408 CODE_FOR_altivec_vpku[hw]um_direct.
7409 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7410 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7411 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7412 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7414 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7416 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7417 generation for -maltivec=be.
7418 (altivec_vsumsws): Simplify redundant test.
7420 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7422 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7423 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7424 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7425 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7426 gen_altivec_vpkuwum.
7427 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7429 (altivec_vpks<VI_char>ss): Likewise.
7430 (altivec_vpks<VI_char>us): Likewise.
7431 (altivec_vpku<VI_char>us): Likewise.
7432 (altivec_vpku<VI_char>um): Likewise.
7433 (altivec_vpku<VI_char>um_direct): New (copy of
7434 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7436 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7437 target is little endian and -maltivec=be is not specified.
7438 (*altivec_vupkhs<VU_char>_direct): New (copy of
7439 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7440 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7441 target is little endian and -maltivec=be is not specified.
7442 (*altivec_vupkls<VU_char>_direct): New (copy of
7443 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7444 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7445 little endian and -maltivec=be is not specified.
7446 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7447 little endian and -maltivec=be is not specified.
7449 2014-02-05 Richard Henderson <rth@redhat.com>
7452 * combine-stack-adj.c: Revert r206943.
7453 * sched-int.h (struct deps_desc): Add last_args_size.
7454 * sched-deps.c (init_deps): Initialize it.
7455 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7456 contain REG_ARGS_SIZE notes.
7458 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7460 * lto-cgraph.c (asm_nodes_output): Make global.
7461 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7462 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7463 (driver_handle_option): Handle OPT_fwpa.
7465 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7468 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7469 a comment typo and formatting issue. If odr_hash hasn't been
7470 created, return vNULL and set *completep to false.
7473 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7474 bb with no successors.
7476 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7479 * doc/invoke.texi (-march): Clarify documentation for ARM.
7483 2014-02-05 Richard Biener <rguenther@suse.de>
7485 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7486 when not vectorizing because of too many alias checks.
7487 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7488 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7490 2014-02-05 Nick Clifton <nickc@redhat.com>
7492 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7493 accept extended registers in any mode when compiling for the MN10300.
7495 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7497 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7498 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7499 sanitization attributes.
7500 (can_inline_edge_p): Likewise.
7501 (sanitize_attrs_match_for_inline_p): New function.
7503 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7505 * ipa-prop.c (detect_type_change): Shor circuit testing of
7506 type changes on THIS pointer.
7508 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7511 * config/pa/pa.c (legitimize_tls_address): Return original address
7512 if not passed a SYMBOL_REF rtx.
7513 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7515 (pa_emit_move_sequence): Simplify TLS source operands.
7516 (pa_legitimate_constant_p): Reject all TLS constants.
7517 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7518 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7520 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7522 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7523 groups when we know they are controlled by LTO.
7524 * varasm.c (default_binds_local_p_1): If object is in other partition,
7525 it will be resolved locally.
7527 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7529 * config/host-linux.c (linux_gt_pch_use_address): Don't
7530 use SSIZE_MAX because it is not always defined.
7532 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7535 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7536 threshold for pseudo splitting.
7537 (update_ebb_live_info): Process call argument hard registers and
7538 hard registers from insn definition too.
7539 (max_small_class_regs_num): New constant.
7540 (inherit_in_ebb): Update live hard regs through EBBs. Update
7541 reloads_num only for small register classes. Don't split for
7544 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7547 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7550 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7552 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7555 2014-02-04 Marek Polacek <polacek@redhat.com>
7557 * gdbinit.in (pel): Define.
7559 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7561 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7564 2014-02-04 Richard Biener <rguenther@suse.de>
7567 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7568 in function context local.
7569 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7570 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7571 similar to LTO_imported_decl_ref.
7573 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7575 PR tree-optimization/60002
7576 * cgraphclones.c (build_function_decl_skip_args): Clear
7579 PR tree-optimization/60023
7580 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7581 false to gsi_replace.
7582 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7583 has been in some EH region and vec_stmt could throw, add
7584 vec_stmt into the same EH region.
7585 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7586 has no lhs, ignore it.
7587 * internal-fn.c (expand_MASK_LOAD): Likewise.
7590 * tree-inline.c (copy_forbidden): Fail for
7591 __attribute__((optimize (0))) functions.
7594 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7595 is set, copy one less argument.
7596 (expand_simd_clones): Don't subtract clone_info->inbranch
7597 from simd_clone_struct_alloc argument.
7599 PR rtl-optimization/57915
7600 * recog.c (simplify_while_replacing): If all unary/binary/relational
7601 operation arguments are constant, attempt to simplify those.
7604 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7605 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7607 2014-02-04 Richard Biener <rguenther@suse.de>
7609 PR tree-optimization/60012
7610 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7611 TBAA disambiguation to all DDRs.
7613 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7616 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7617 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7619 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7622 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7624 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7626 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7627 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7629 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7632 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7633 to figure out targets of polymorphic calls with known decl.
7634 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7635 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7636 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7637 (get_polymorphic_call_info): ... here.
7638 (get_polymorphic_call_info_from_invariant): New function.
7640 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7642 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7643 lookup via vtable pointer; check for type consistency
7644 and turn inconsitent facts into UNREACHABLE.
7645 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7646 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7647 type inconsistent querries; return UNREACHABLE instead.
7649 2014-02-03 Richard Henderson <rth@twiddle.net>
7652 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7653 already processed this node.
7654 (normalize_one_pred_1): Pass along mark_set.
7655 (normalize_one_pred): Create and destroy a pointer_set_t.
7656 (normalize_one_pred_chain): Likewise.
7658 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7660 PR gcov-profile/58602
7661 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7663 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7666 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7667 -fno-devirtualize; try to devirtualize by the knowledge of
7668 virtual table pointer given by aggregate propagation.
7669 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7670 (ipa_print_node_jump_functions): Dump also offset that
7671 is relevant for polymorphic calls.
7672 (determine_known_aggregate_parts): Add arg_type parameter; use it
7673 instead of determining the type from pointer type.
7674 (ipa_compute_jump_functions_for_edge): Update call of
7675 determine_known_aggregate_parts.
7676 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7677 (gimple_get_virt_method_for_binfo): ... here; simplify using
7678 vtable_pointer_value_to_vtable.
7679 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7680 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7681 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7682 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7684 (vtable_pointer_value_to_binfo): ... here.
7685 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7687 2014-02-03 Teresa Johnson <tejohnson@google.com>
7689 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7690 redef of outer loop index variable.
7692 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7696 * doc/extend.texi (Function Attributes): Typo.
7698 2014-02-03 Cong Hou <congh@google.com>
7700 PR tree-optimization/60000
7701 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7702 if the vectorized statement is a store. A store statement can only
7703 appear at the end of pattern statements.
7705 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7707 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7708 (ix86_option_override_internal): Default long double to 64-bit for
7709 32-bit Bionic and to 128-bit for 64-bit Bionic.
7711 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7712 TARGET_LONG_DOUBLE_128 is true.
7713 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7715 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7716 (mlong-double-64): Negate -mlong-double-128.
7717 (mlong-double-128): New option.
7719 * config/i386/i386-c.c (ix86_target_macros): Define
7720 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7722 * doc/invoke.texi: Document -mlong-double-128.
7724 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7726 PR rtl-optimization/60024
7727 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7729 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7731 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7733 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7735 PR rtl-optimization/57662
7736 * sel-sched.c (code_motion_path_driver): Do not mark already not
7737 existing blocks in the visiting bitmap.
7739 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7741 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7742 on the insn being emitted.
7744 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7745 Will Deacon <will.deacon@arm.com>
7747 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7749 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7751 * config/arm/arm-tables.opt: Regenerate.
7753 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7755 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7756 for vector types other than V16QImode.
7757 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7758 define_expand, and call altivec_expand_vec_perm_le when producing
7759 code with little endian element order.
7760 (*altivec_vperm_<mode>_internal): New insn having previous
7761 behavior of altivec_vperm_<mode>.
7762 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7763 altivec_expand_vec_perm_le when producing code with little endian
7765 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7766 behavior of altivec_vperm_<mode>_uns.
7768 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7770 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7771 (altivec_vsumsws): Add handling for -maltivec=be with a little
7773 (altivec_vsumsws_direct): New.
7774 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7775 gen_altivec_vsumsws.
7777 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7779 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7780 vtable_pointer_value_to_binfo): New functions.
7781 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7782 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7784 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7786 * config/nios2/nios2.md (load_got_register): Initialize GOT
7787 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7788 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7790 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7792 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7793 preserverd by passthrough, do not propagate the type.
7795 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7797 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7798 (mips_atomic_assign_expand_fenv): New function.
7799 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7801 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7803 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7804 (__builtin_mips_set_fcsr): Likewise.
7805 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7806 MIPS_USI_FTYPE_VOID.
7807 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7808 (mips16_expand_set_fcsr): Likewise.
7809 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7810 (mips16_set_fcsr_stub): Likewise.
7811 (mips16_get_fcsr_one_only_stub): New class.
7812 (mips16_set_fcsr_one_only_stub): Likewise.
7813 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7814 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7815 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7816 (hard_float): New availability predicate.
7817 (mips_builtins): Add get_fcsr and set_fcsr.
7818 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7819 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7820 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7821 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7822 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7825 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7827 * config/mips/mips.c (mips_one_only_stub): New class.
7828 (mips_need_mips16_rdhwr_p): Replace with...
7829 (mips16_rdhwr_stub): ...this new variable.
7830 (mips16_stub_call_address): New function.
7831 (mips16_rdhwr_one_only_stub): New class.
7832 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7833 (mips_output_mips16_rdhwr): Delete.
7834 (mips_finish_stub): New function.
7835 (mips_code_end): Use it to handle rdhwr stubs.
7837 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7840 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7841 when calculating size of integer atomic types.
7843 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7845 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7847 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7849 PR tree-optimization/60003
7850 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7851 * profile.c (branch_prob): Use gimple_call_builtin_p
7852 to check for BUILT_IN_SETJMP_RECEIVER.
7853 * tree-inline.c (copy_bb): Call notice_special_calls.
7855 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7858 * lra-constraints.c (process_alt_operands): Update reload_sum only
7861 2014-01-31 Richard Henderson <rth@redhat.com>
7864 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7865 until after else_eh is processed.
7867 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7869 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7870 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7871 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7872 in smmintrin.h, remove them.
7873 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7874 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7875 * config/i386/i386.md (ROUND_SAE): Fix value.
7876 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7877 (const48_operand): New.
7878 * config/i386/subst.md (round), (round_expand): Use
7879 const_4_or_8_to_11_operand.
7880 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7882 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7884 * config/i386/constraints.md (Yk): Swap meaning with k.
7885 * config/i386/i386.md (movhi_internal): Change Yk to k.
7886 (movqi_internal): Ditto.
7887 (*k<logic><mode>): Ditto.
7890 (kandn<mode>): Ditto.
7891 (*<code>hi_1): Ditto.
7892 (*<code>qi_1): Ditto.
7893 (kxnor<mode>): Ditto.
7894 (kortestzhi): Ditto.
7895 (kortestchi): Ditto.
7897 (*one_cmplhi2_1): Ditto.
7898 (*one_cmplqi2_1): Ditto.
7899 * config/i386/sse.md (): Change k to Yk.
7900 (avx512f_load<mode>_mask): Ditto.
7901 (avx512f_blendm<mode>): Ditto.
7902 (avx512f_store<mode>_mask): Ditto.
7903 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7904 (avx512f_storedqu<mode>_mask): Ditto.
7905 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7907 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7908 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7909 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7910 (avx512f_maskcmp<mode>3): Ditto.
7911 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7912 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7913 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7914 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7915 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7916 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7917 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7918 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7919 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7920 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7921 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7922 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7923 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7924 (vec_extract_lo_<mode>_maskm): Ditto.
7925 (vec_extract_hi_<mode>_maskm): Ditto.
7926 (avx512f_vternlog<mode>_mask): Ditto.
7927 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7928 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7929 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7930 (avx512f_<code>v8div16qi2_mask): Ditto.
7931 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7932 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7933 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7934 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7935 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7936 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7937 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7938 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7939 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7940 (avx512cd_maskb_vec_dupv8di): Ditto.
7941 (avx512cd_maskw_vec_dupv16si): Ditto.
7942 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7943 (avx512f_vpermi2var<mode>3_mask): Ditto.
7944 (avx512f_vpermi2var<mode>3_mask): Ditto.
7945 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7946 (*avx512f_gathersi<mode>): Ditto.
7947 (*avx512f_gathersi<mode>_2): Ditto.
7948 (*avx512f_gatherdi<mode>): Ditto.
7949 (*avx512f_gatherdi<mode>_2): Ditto.
7950 (*avx512f_scattersi<mode>): Ditto.
7951 (*avx512f_scatterdi<mode>): Ditto.
7952 (avx512f_compress<mode>_mask): Ditto.
7953 (avx512f_compressstore<mode>_mask): Ditto.
7954 (avx512f_expand<mode>_mask): Ditto.
7955 * config/i386/subst.md (mask): Change k to Yk.
7956 (mask_scalar_merge): Ditto.
7959 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7961 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7963 2014-01-31 Richard Biener <rguenther@suse.de>
7966 * builtins.c (fold_builtin_memory_op): Make sure to not
7967 use a floating-point mode or a boolean or enumeral type for
7970 2014-01-30 DJ Delorie <dj@redhat.com>
7972 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7973 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7974 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7975 whenever main() has an epilogue.
7977 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7979 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7980 unused variable "field".
7981 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7982 (vsx_mergeh_<mode>): Likewise.
7983 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7984 (altivec_vmrghh): Likewise.
7985 (altivec_vmrghw): Likewise.
7986 (altivec_vmrglb): Likewise.
7987 (altivec_vmrglh): Likewise.
7988 (altivec_vmrglw): Likewise.
7989 (altivec_vspltb): Add missing uses.
7990 (altivec_vsplth): Likewise.
7991 (altivec_vspltw): Likewise.
7992 (altivec_vspltsf): Likewise.
7994 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7997 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7998 frame related instructions.
8000 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8002 PR rtl-optimization/59959
8003 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8004 any reload of register whose subreg is invalid.
8006 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8008 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8009 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8010 Add missing return type - void.
8012 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8014 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8015 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8016 remove element index adjustment for endian (now handled in vsx.md
8018 (altivec_expand_vec_perm_const): Use
8019 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8020 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8021 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8022 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8023 define_expand and a new define_insn *altivec_vspltb_internal;
8024 adjust for -maltivec=be on a little endian target.
8025 (altivec_vspltb_direct): New.
8026 (altivec_vsplth): Divide into a define_expand and a new
8027 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8028 little endian target.
8029 (altivec_vsplth_direct): New.
8030 (altivec_vspltw): Divide into a define_expand and a new
8031 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8032 little endian target.
8033 (altivec_vspltw_direct): New.
8034 (altivec_vspltsf): Divide into a define_expand and a new
8035 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8036 a little endian target.
8038 2014-01-30 Richard Biener <rguenther@suse.de>
8040 PR tree-optimization/59993
8041 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8042 can propagate form the earlier stmt and avoid the transform
8043 when the intermediate result is needed.
8045 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8047 * README.Portability: Fix typo.
8049 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8051 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8052 comparison_operator with ordered_comparison_operator.
8054 2014-01-30 Nick Clifton <nickc@redhat.com>
8056 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8057 Rename to mn10300_store_multiple_regs.
8058 * config/mn10300/mn10300.c: Likewise.
8059 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8060 store_multiple_regs.
8061 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8062 Call mn10300_store_multiple_regs.
8064 2014-01-30 Nick Clifton <nickc@redhat.com>
8065 DJ Delorie <dj@redhat.com>
8067 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8068 %fp 2 to keep registers after it properly word-aligned.
8069 (rl78_alloc_physical_registers_umul): Handle the case where both
8070 input operands are the same.
8072 2014-01-30 Richard Biener <rguenther@suse.de>
8074 PR tree-optimization/59903
8075 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8078 2014-01-30 Jason Merrill <jason@redhat.com>
8081 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8084 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8086 2014-01-30 Richard Biener <rguenther@suse.de>
8088 PR tree-optimization/59951
8089 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8091 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8094 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8095 SFmode to DFmode case.
8097 2014-01-29 DJ Delorie <dj@redhat.com>
8099 * config/msp430/msp430.opt (-minrt): New.
8100 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8102 (ENDFILE_SPEC): Likewise.
8104 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8106 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8107 (estimate_function_body_sizes): Use it.
8109 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8112 * dwarf2out.c (is_cxx_auto): New.
8113 (is_base_type): Use it.
8114 (gen_type_die_with_usage): Likewise.
8116 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8118 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8119 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8120 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8121 -maltivec=be with LE targets.
8122 (vsx_mergeh_<mode>): Likewise.
8123 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8124 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8125 (altivec_vmrghb): Replace with define_expand and new
8126 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8127 (altivec_vmrghb_direct): New define_insn.
8128 (altivec_vmrghh): Replace with define_expand and new
8129 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8130 (altivec_vmrghh_direct): New define_insn.
8131 (altivec_vmrghw): Replace with define_expand and new
8132 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8133 (altivec_vmrghw_direct): New define_insn.
8134 (*altivec_vmrghsf): Adjust for endianness.
8135 (altivec_vmrglb): Replace with define_expand and new
8136 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8137 (altivec_vmrglb_direct): New define_insn.
8138 (altivec_vmrglh): Replace with define_expand and new
8139 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8140 (altivec_vmrglh_direct): New define_insn.
8141 (altivec_vmrglw): Replace with define_expand and new
8142 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8143 (altivec_vmrglw_direct): New define_insn.
8144 (*altivec_vmrglsf): Adjust for endianness.
8145 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8146 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8147 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8148 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8149 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8150 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8151 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8152 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8154 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8156 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8157 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8160 2014-01-29 Richard Biener <rguenther@suse.de>
8162 PR tree-optimization/58742
8163 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8164 associate_pointerplus_align.
8165 (associate_pointerplus_diff): New function.
8166 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8167 and associate_pointerplus_diff.
8169 2014-01-29 Richard Biener <rguenther@suse.de>
8171 * lto-streamer.h (LTO_major_version): Bump to 3.
8172 (LTO_minor_version): Reset to 0.
8174 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8176 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8177 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8178 (arm_file_start): Generate correct asm header for armv7ve.
8179 * config/arm/bpabi.h: Add multilib support for armv7ve.
8180 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8181 and cortex-a15 to armv7ve.
8182 * config/arm/t-aprofile: Add multilib support for armv7ve.
8183 * doc/invoke.texi: Document -march=armv7ve.
8185 2014-01-29 Richard Biener <rguenther@suse.de>
8187 PR tree-optimization/58742
8188 * tree-ssa-forwprop.c (associate_plusminus): Return true
8189 if we changed sth, defer EH cleanup to ...
8190 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8191 (simplify_mult): New function.
8193 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8196 PR tree-optimization/59920
8197 * tree.c (build_common_builtin_nodes): Remove
8198 __builtin_setjmp_dispatcher initialization.
8199 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8200 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8201 instead of gsi_after_labels + manually skipping debug stmts.
8202 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8203 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8204 * tree-inline.c (copy_edges_for_bb): Remove
8205 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8206 argument. Ignore computed_goto_p stmts. Don't call
8207 make_abnormal_goto_edges. If a call might need abnormal edges
8208 for non-local gotos, see if it already has an edge to
8209 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8210 with true argument, don't do anything then, otherwise add
8211 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8212 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8214 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8215 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8216 (lower_stmt): Don't set data->calls_builtin_setjmp.
8217 (lower_builtin_setjmp): Adjust comment.
8218 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8219 * tree-cfg.c (found_computed_goto): Remove.
8220 (factor_computed_gotos): Remove.
8221 (make_goto_expr_edges): Return bool, true for computed gotos.
8222 Don't call make_abnormal_goto_edges.
8223 (build_gimple_cfg): Don't set found_computed_goto, don't call
8224 factor_computed_gotos.
8225 (computed_goto_p): No longer static.
8226 (make_blocks): Don't set found_computed_goto.
8227 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8228 (make_edges): If make_goto_expr_edges returns true, push bb
8229 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8230 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8231 vector. Record mapping between bbs and OpenMP regions if there
8232 are any, adjust make_gimple_omp_edges caller. Call
8233 handle_abnormal_edges.
8234 (make_abnormal_goto_edges): Remove.
8235 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8236 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8237 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8238 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8239 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8240 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8241 filling *region also set *region_idx to (*region)->entry->index.
8244 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8245 For REGs set ORIGINAL_REGNO.
8247 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8249 * doc/md.texi: Mention that a target shouldn't implement
8250 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8253 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8255 PR tree-optimization/59594
8256 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8257 a copy of the datarefs vector rather than the vector itself.
8259 2014-01-28 Jason Merrill <jason@redhat.com>
8262 * dwarf2out.c (auto_die): New static.
8263 (gen_type_die_with_usage): Handle C++1y 'auto'.
8264 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8267 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8270 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8271 (SPEC_X32): Likewise.
8272 (SPEC_64): Likewise.
8273 * config/i386/i386.c (ix86_option_override_internal): Turn off
8274 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8276 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8277 * config/i386/i386.h (TARGET_16BIT): New macro.
8278 (TARGET_16BIT_P): Likewise.
8279 * config/i386/i386.opt: Add m16.
8280 * doc/invoke.texi: Document -m16.
8282 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8284 PR preprocessor/59935
8285 * input.c (location_get_source_line): Bail out on when line number
8286 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8288 2014-01-28 Richard Biener <rguenther@suse.de>
8290 PR tree-optimization/58742
8291 * tree-ssa-forwprop.c (associate_plusminus): Handle
8292 pointer subtraction of the form (T)(P + A) - (T)P.
8294 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8296 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8299 2014-01-28 Richard Biener <rguenther@suse.de>
8302 2014-01-28 Richard Biener <rguenther@suse.de>
8304 PR rtl-optimization/45364
8305 PR rtl-optimization/59890
8306 * var-tracking.c (local_get_addr_clear_given_value): Handle
8307 already cleared slot.
8308 (val_reset): Handle not allocated local_get_addr_cache.
8309 (vt_find_locations): Use post-order on the inverted CFG.
8311 2014-01-28 Richard Biener <rguenther@suse.de>
8313 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8315 2014-01-28 Richard Biener <rguenther@suse.de>
8317 PR rtl-optimization/45364
8318 PR rtl-optimization/59890
8319 * var-tracking.c (local_get_addr_clear_given_value): Handle
8320 already cleared slot.
8321 (val_reset): Handle not allocated local_get_addr_cache.
8322 (vt_find_locations): Use post-order on the inverted CFG.
8324 2014-01-28 Alan Modra <amodra@gmail.com>
8326 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8327 * configure.ac <recursive call for build != host>: Define
8328 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8329 and LD_FOR_BUILD too.
8330 * configure: Regenerate.
8332 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8334 * config/i386/i386.c (get_builtin_code_for_version): Separate
8335 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8336 Broadwell from Haswell.
8338 2014-01-27 Steve Ellcey <sellcey@mips.com>
8340 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8341 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8342 * config/mips/mips.c (mips_option_override): Change setting
8344 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8345 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8346 Change from Mask to Var.
8348 2014-01-27 Jeff Law <law@redhat.com>
8350 * ipa-inline.c (inline_small_functions): Fix typo.
8352 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8354 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8355 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8356 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8357 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8358 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8359 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8360 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8361 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8362 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8363 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8364 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8365 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8366 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8367 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8368 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8369 (_mm512_storeu_epi64): Ditto.
8370 (_mm512_cmpge_epi32_mask): Ditto.
8371 (_mm512_cmpge_epu32_mask): Ditto.
8372 (_mm512_cmpge_epi64_mask): Ditto.
8373 (_mm512_cmpge_epu64_mask): Ditto.
8374 (_mm512_cmple_epi32_mask): Ditto.
8375 (_mm512_cmple_epu32_mask): Ditto.
8376 (_mm512_cmple_epi64_mask): Ditto.
8377 (_mm512_cmple_epu64_mask): Ditto.
8378 (_mm512_cmplt_epi32_mask): Ditto.
8379 (_mm512_cmplt_epu32_mask): Ditto.
8380 (_mm512_cmplt_epi64_mask): Ditto.
8381 (_mm512_cmplt_epu64_mask): Ditto.
8382 (_mm512_cmpneq_epi32_mask): Ditto.
8383 (_mm512_cmpneq_epu32_mask): Ditto.
8384 (_mm512_cmpneq_epi64_mask): Ditto.
8385 (_mm512_cmpneq_epu64_mask): Ditto.
8386 (_mm512_expand_pd): Ditto.
8387 (_mm512_expand_ps): Ditto.
8388 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8389 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8390 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8391 * config/i386/i386.c (ix86_builtins): Add
8392 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8393 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8394 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8395 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8396 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8397 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8398 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8399 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8400 IX86_BUILTIN_PMOVUSQW512_MEM.
8401 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8402 __builtin_ia32_pmovsqd512mem_mask,
8403 __builtin_ia32_pmovqd512mem_mask,
8404 __builtin_ia32_pmovusqw512mem_mask,
8405 __builtin_ia32_pmovsqw512mem_mask,
8406 __builtin_ia32_pmovqw512mem_mask,
8407 __builtin_ia32_pmovusdw512mem_mask,
8408 __builtin_ia32_pmovsdw512mem_mask,
8409 __builtin_ia32_pmovdw512mem_mask,
8410 __builtin_ia32_pmovqb512mem_mask,
8411 __builtin_ia32_pmovusqb512mem_mask,
8412 __builtin_ia32_pmovsqb512mem_mask,
8413 __builtin_ia32_pmovusdb512mem_mask,
8414 __builtin_ia32_pmovsdb512mem_mask,
8415 __builtin_ia32_pmovdb512mem_mask.
8416 (bdesc_args): Add __builtin_ia32_expanddf512,
8417 __builtin_ia32_expandsf512.
8418 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8419 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8420 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8421 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8422 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8423 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8424 (avx512f_<code>v8div16qi2_mask_store): This.
8425 (avx512f_expand<mode>): New.
8427 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8429 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8431 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8432 (_mm512_prefetch_i32scatter_pd): Ditto.
8433 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8434 (_mm512_prefetch_i64scatter_pd): Ditto.
8435 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8436 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8437 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8438 (_mm512_prefetch_i32scatter_ps): Ditto.
8439 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8440 (_mm512_prefetch_i64scatter_ps): Ditto.
8441 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8442 * config/i386/i386-builtin-types.def: Define
8443 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8444 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8445 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8446 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8447 IX86_BUILTIN_SCATTERPFQPD.
8448 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8449 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8450 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8451 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8452 __builtin_ia32_scatterpfqps.
8453 (ix86_expand_builtin): Expand new built-ins.
8454 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8455 fix memory access data type.
8456 (*avx512pf_gatherpf<mode>_mask): Ditto.
8457 (*avx512pf_gatherpf<mode>): Ditto.
8458 (avx512pf_scatterpf<mode>): Ditto.
8459 (*avx512pf_scatterpf<mode>_mask): Ditto.
8460 (*avx512pf_scatterpf<mode>): Ditto.
8461 (GATHER_SCATTER_SF_MEM_MODE): New.
8462 (avx512pf_gatherpf<mode>df): Ditto.
8463 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8464 (*avx512pf_scatterpf<mode>df): Ditto.
8466 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8469 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8470 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8473 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8475 * common/config/arm/arm-common.c
8476 (arm_rewrite_mcpu): Handle multiple names.
8478 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8480 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8482 * gimple-builder.h (create_gimple_tmp): Delete.
8484 2014-01-27 Christian Bruel <christian.bruel@st.com>
8486 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8489 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8491 * config/pa/pa.md (call): Generate indirect long calls to non-local
8492 functions when outputing 32-bit code.
8493 (call_value): Likewise except for special call to buggy powf function.
8495 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8496 portable runtime and PIC indirect calls.
8497 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8498 and PIC call sequences. Use ldo instead of blr to set return register
8499 in PIC call sequence.
8501 2014-01-25 Walter Lee <walt@tilera.com>
8503 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8504 avoid clobbering a live register.
8506 2014-01-25 Walter Lee <walt@tilera.com>
8508 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8509 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8510 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8511 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8513 2014-01-25 Walter Lee <walt@tilera.com>
8515 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8516 arguments on even registers.
8517 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8519 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8520 (BIGGEST_ALIGNMENT): Ditto.
8521 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8523 2014-01-25 Walter Lee <walt@tilera.com>
8525 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8526 insns before bundling.
8527 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8529 2014-01-25 Walter Lee <walt@tilera.com>
8531 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8532 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8533 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8535 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8537 * config/mips/constraints.md (kl): Delete.
8538 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8539 define expands, using...
8540 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8541 instructions for MIPS16.
8542 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8543 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8545 2014-01-25 Walter Lee <walt@tilera.com>
8547 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8551 2014-01-25 Walter Lee <walt@tilera.com>
8553 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8554 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8556 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8558 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8561 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8563 * print-rtl.c (in_call_function_usage): New var.
8564 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8565 EXPR_LIST mode as mode and not as reg note name.
8568 * cfgloopmanip.c (copy_loop_info): If
8569 loop->warned_aggressive_loop_optimizations, make sure
8570 the flag is set in target loop too.
8572 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8574 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8576 * builtins.def: Likewise.
8577 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8578 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8579 * ira.c (ira_setup_eliminable_regset): Likewise.
8580 * omp-low.c (gate_expand_omp): Likewise.
8581 (execute_lower_omp): Likewise.
8582 (diagnose_sb_0): Likewise.
8583 (gate_diagnose_omp_blocks): Likewise.
8584 (simd_clone_clauses_extract): Likewise.
8587 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8589 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8590 correction for little endian...
8591 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8594 2014-01-24 Jeff Law <law@redhat.com>
8596 PR tree-optimization/59919
8597 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8598 for non-returning calls.
8600 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8602 * common/config/aarch64/aarch64-common.c
8603 (aarch64_rewrite_mcpu): Handle multiple names.
8604 * config/aarch64/aarch64.h
8605 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8607 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8609 * input.c (add_file_to_cache_tab): Handle the case where fopen
8612 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8615 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8616 from push operand if code of push isn't PRE_DEC.
8618 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8621 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8622 -mquad-memory-atomic. Update -mquad-memory documentation to say
8623 it is only used for non-atomic loads/stores.
8625 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8626 -mquad-memory or -mquad-memory-atomic switches.
8628 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8629 -mquad-memory-atomic to ISA 2.07 support.
8631 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8632 to separate support of normal quad word memory operations (ldq, stq)
8633 from the atomic quad word memory operations.
8635 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8636 support to separate non-atomic quad word operations from atomic
8637 quad word operations. Disable non-atomic quad word operations in
8638 little endian mode so that we don't have to swap words after the
8639 load and before the store.
8640 (quad_load_store_p): Add comment about atomic quad word support.
8641 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8642 options printed with -mdebug=reg.
8644 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8645 -mquad-memory-atomic as the test for whether we have quad word
8646 atomic instructions.
8647 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8648 or -mp8-vector are used, allow byte/half-word atomic operations.
8650 * config/rs6000/sync.md (load_lockedti): Insure that the address
8651 is a proper indexed or indirect address for the lqarx instruction.
8652 On little endian systems, swap the hi/lo registers after the lqarx
8654 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8655 insure the address is valid for the lqarx instruction.
8656 (store_conditionalti): Insure that the address is a proper indexed
8657 or indirect address for the stqcrx. instruction. On little endian
8658 systems, swap the hi/lo registers before doing the stqcrx.
8660 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8661 insure the address is valid for the stqcrx. instruction.
8663 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8664 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8665 type of quad memory support is available.
8667 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8670 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8671 there is a danger of looping.
8673 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8675 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8676 force flag_ira_loop_pressure if set via command line.
8678 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8680 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8681 (ashr_simd): New builtin handling DI mode.
8682 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8683 (aarch64_sshr_simddi): New match pattern.
8684 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8685 (vshrd_n_s64): Likewise.
8686 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8688 2014-01-23 Nick Clifton <nickc@redhat.com>
8690 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8691 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8692 favour of mcu specific scripts.
8693 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8696 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8697 Alex Velenko <Alex.Velenko@arm.com>
8699 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8700 (vaddv_s16): Likewise.
8701 (vaddv_s32): Likewise.
8702 (vaddv_u8): Likewise.
8703 (vaddv_u16): Likewise.
8704 (vaddv_u32): Likewise.
8705 (vaddvq_s8): Likewise.
8706 (vaddvq_s16): Likewise.
8707 (vaddvq_s32): Likewise.
8708 (vaddvq_s64): Likewise.
8709 (vaddvq_u8): Likewise.
8710 (vaddvq_u16): Likewise.
8711 (vaddvq_u32): Likewise.
8712 (vaddvq_u64): Likewise.
8713 (vaddv_f32): Likewise.
8714 (vaddvq_f32): Likewise.
8715 (vaddvq_f64): Likewise.
8716 (vmaxv_f32): Likewise.
8717 (vmaxv_s8): Likewise.
8718 (vmaxv_s16): Likewise.
8719 (vmaxv_s32): Likewise.
8720 (vmaxv_u8): Likewise.
8721 (vmaxv_u16): Likewise.
8722 (vmaxv_u32): Likewise.
8723 (vmaxvq_f32): Likewise.
8724 (vmaxvq_f64): Likewise.
8725 (vmaxvq_s8): Likewise.
8726 (vmaxvq_s16): Likewise.
8727 (vmaxvq_s32): Likewise.
8728 (vmaxvq_u8): Likewise.
8729 (vmaxvq_u16): Likewise.
8730 (vmaxvq_u32): Likewise.
8731 (vmaxnmv_f32): Likewise.
8732 (vmaxnmvq_f32): Likewise.
8733 (vmaxnmvq_f64): Likewise.
8734 (vminv_f32): Likewise.
8735 (vminv_s8): Likewise.
8736 (vminv_s16): Likewise.
8737 (vminv_s32): Likewise.
8738 (vminv_u8): Likewise.
8739 (vminv_u16): Likewise.
8740 (vminv_u32): Likewise.
8741 (vminvq_f32): Likewise.
8742 (vminvq_f64): Likewise.
8743 (vminvq_s8): Likewise.
8744 (vminvq_s16): Likewise.
8745 (vminvq_s32): Likewise.
8746 (vminvq_u8): Likewise.
8747 (vminvq_u16): Likewise.
8748 (vminvq_u32): Likewise.
8749 (vminnmv_f32): Likewise.
8750 (vminnmvq_f32): Likewise.
8751 (vminnmvq_f64): Likewise.
8753 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8755 * config/aarch64/aarch64-simd.md
8756 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8757 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8758 (*aarch64_mul3_elt<mode>): Likewise.
8759 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8760 (*aarch64_mul3_elt_to_64v2df): Likewise.
8761 (*aarch64_mla_elt<mode>): Likewise.
8762 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8763 (*aarch64_mls_elt<mode>): Likewise.
8764 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8765 (*aarch64_fma4_elt<mode>): Likewise.
8766 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8767 (*aarch64_fma4_elt_to_64v2df): Likewise.
8768 (*aarch64_fnma4_elt<mode>): Likewise.
8769 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8770 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8771 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8772 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8773 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8774 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8775 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8776 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8777 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8779 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8781 * config/aarch64/aarch64-simd.md
8782 (aarch64_be_checked_get_lane<mode>): New define_expand.
8783 * config/aarch64/aarch64-simd-builtins.def
8784 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8785 New builtin definition.
8786 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8787 Use new safe be builtin.
8789 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8791 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8793 (aarch64_be_st1<mode>): Likewise.
8794 (aarch_ld1<VALL:mode>): Define_expand modified.
8795 (aarch_st1<VALL:mode>): Likewise.
8796 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8797 (UNSPEC_ST1): Likewise.
8799 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8801 * config/microblaze/microblaze.md: Add trap insn and attribute
8803 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8805 PR preprocessor/58580
8806 * input.h (location_get_source_line): Take an additional line_size
8808 (void diagnostics_file_cache_fini): Declare new function.
8809 * input.c (struct fcache): New type.
8810 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8811 New static constants.
8812 (diagnostic_file_cache_init, total_lines_num)
8813 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8814 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8815 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8816 (get_next_line, read_next_line, goto_next_line, read_line_num):
8817 New static function definitions.
8818 (diagnostic_file_cache_fini): New function.
8819 (location_get_source_line): Take an additional output line_len
8820 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8822 * diagnostic.c (diagnostic_finish): Call
8823 diagnostic_file_cache_fini.
8824 (adjust_line): Take an additional input parameter for the length
8825 of the line, rather than calculating it with strlen.
8826 (diagnostic_show_locus): Adjust the use of
8827 location_get_source_line and adjust_line with respect to their new
8828 signature. While displaying a line now, do not stop at the first
8829 null byte. Rather, display the zero byte as a space and keep
8830 going until we reach the size of the line.
8831 * Makefile.in: Add vec.o to OBJS-libcommon
8833 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8834 Ilya Tocar <ilya.tocar@intel.com>
8836 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8837 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8838 (__builtin_ia32_kmov16): Ditto.
8839 * config/i386/i386.md (UNSPEC_KMOV): New.
8842 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8844 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8845 (_mm512_storeu_si512): Ditto.
8847 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8850 * rtl.h (get_referenced_operands): Declare.
8851 * recog.c (get_referenced_operands): New function.
8852 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8853 operands have been referenced when recording LO_SUM references.
8855 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8857 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8859 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8861 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8862 Enable for generic and recent AMD targets.
8864 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8866 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8867 ARG_SIZE note when adjustment was eliminated.
8869 2014-01-22 Jeff Law <law@redhat.com>
8871 PR tree-optimization/59597
8872 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8873 in file. Accept new argument REGISTERING and use it to modify
8874 dump output appropriately.
8875 (register_jump_thread): Corresponding changes.
8876 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8877 thread paths involving joiner blocks. Add code to dump cancelled
8878 jump threading paths.
8880 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8882 PR rtl-optimization/59477
8883 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8884 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8886 2014-01-22 Tom Tromey <tromey@redhat.com>
8888 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8890 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8892 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8894 PR rtl-optimization/59896
8895 * lra-constraints.c (process_alt_operands): Check unused note for
8896 matched operands of insn with no output reloads.
8898 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8900 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8901 (mips_move_from_gpr_cost): Likewise.
8903 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8905 PR rtl-optimization/59858
8906 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8907 ira_class_hard_regs_num.
8908 (process_alt_operands): Increase reject for dying matched operand.
8910 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8913 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8914 smaller than size, perform several stores or loads and stores
8915 at dst + count - size to store or copy all of size bytes, rather
8916 than just last modesize bytes.
8918 2014-01-20 DJ Delorie <dj@redhat.com>
8920 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8921 that CLOBBERs are REGs before propogating their values.
8923 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8926 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8927 (cgraph_inline_failed_type): New function.
8928 * cgraph.h (DEFCIFCODE): Add type.
8929 (cgraph_inline_failed_type_t): New enum.
8930 (cgraph_inline_failed_type): New prototype.
8931 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8932 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8933 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8934 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8935 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8936 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8937 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8938 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8939 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8940 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8941 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8942 OPTIMIZATION_MISMATCH.
8943 * tree-inline.c (expand_call_inline): Emit errors during
8944 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8946 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8949 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8950 mode attribute in insn output.
8952 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8954 * output.h (output_constant): Delete.
8955 * varasm.c (output_constant): Make private.
8957 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8959 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8961 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8964 * tree.h (fold_builtin_strcat): New prototype.
8965 * builtins.c (fold_builtin_strcat): No longer static. Add len
8966 argument, if non-NULL, don't call c_strlen. Optimize
8967 directly into __builtin_memcpy instead of __builtin_strcpy.
8968 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8969 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8971 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8973 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8974 for SImode_address_operand operands, having only a REG argument.
8976 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8978 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8979 loader name using mbig-endian.
8980 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8982 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8984 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8988 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8990 * config/aarch64/aarch64-protos.h
8991 (aarch64_cannot_change_mode_class_ptr): Declare.
8992 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8993 aarch64_cannot_change_mode_class_ptr): New.
8994 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8995 backend hook aarch64_cannot_change_mode_class.
8997 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8999 * common/config/aarch64/aarch64-common.c
9000 (aarch64_handle_option): Don't handle any option order logic here.
9001 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9002 selected_cpu, warn on architecture version mismatch.
9003 (aarch64_override_options): Fix parsing order for option strings.
9005 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9006 Iain Sandoe <iain@codesourcery.com>
9009 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9010 warning. Amend comment to reflect current functionality.
9012 2014-01-20 Richard Biener <rguenther@suse.de>
9015 * builtins.c (fold_builtin_strcat): Remove case better handled
9016 by tree-ssa-strlen.c.
9018 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9020 * config/aarch64/aarch64.opt
9021 (mcpu, march, mtune): Make case-insensitive.
9023 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9026 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9027 if operands[1] is a REG or ZERO_EXTEND of a REG.
9029 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9031 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9033 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9035 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9036 long non-pic millicode calls.
9038 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9040 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9042 2014-01-19 Kito Cheng <kito@0xlab.org>
9044 * builtins.c (expand_movstr): Check movstr expand done or fail.
9046 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9047 H.J. Lu <hongjiu.lu@intel.com>
9050 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9051 to DImode for zero-extended addresses.
9053 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9055 PR rtl-optimization/57763
9056 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9057 on the new indirect jump_insn and increment LABEL_NUSES (label).
9059 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9063 * config.gcc (x86_archs): New variable.
9064 (x86_64_archs): Likewise.
9065 (x86_cpus): Likewise.
9066 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9067 --with-arch/--with-cpu= options.
9068 Support --with-arch=/--with-cpu={nehalem,westmere,
9069 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9071 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9073 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9074 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9076 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9078 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9080 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9083 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9084 clear cpp_get_options (parse_in)->warn_unused_macros for
9085 ix86_target_macros_internal with cpp_define.
9087 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9089 * jump.c (delete_related_insns): Keep (use (insn))s.
9090 * reorg.c (redundant_insn): Check for barriers too.
9092 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9094 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9096 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9098 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9099 call to $$dyncall when TARGET_LONG_CALLS is true.
9101 2014-01-17 Jeff Law <law@redhat.com>
9103 * ree.c (combine_set_extension): Temporarily disable test for
9104 changing number of hard registers.
9106 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9109 * ipa-inline-analysis.c (inline_free_summary):
9110 Do not free summary of aliases.
9112 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9115 * gimplify.c (gimplify_expr): Use create_tmp_var
9116 instead of create_tmp_var_raw. If cond doesn't have
9117 integral type, don't add the IFN_ANNOTATE builtin at all.
9119 2014-01-17 Martin Jambor <mjambor@suse.cz>
9122 * ipa-cp.c (prev_edge_clone): New variable.
9123 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9124 Also resize prev_edge_clone vector.
9125 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9126 (ipcp_edge_removal_hook): New function.
9127 (ipcp_driver): Register ipcp_edge_removal_hook.
9129 2014-01-17 Andrew Pinski <apinski@cavium.com>
9130 Steve Ellcey <sellcey@mips.com>
9133 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9136 2014-01-17 Jeff Law <law@redhat.com>
9139 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9140 so that pass_ccp runs first.
9142 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9144 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9145 (ix86_adjust_cost): Use !TARGET_XXX.
9146 (do_reorder_for_imul): Likewise.
9147 (swap_top_of_ready_list): Likewise.
9148 (ix86_sched_reorder): Likewise.
9150 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9152 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9153 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9154 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9155 (intel_memset): New. Duplicate slm_memset.
9156 (intel_cost): New. Duplicate slm_cost.
9157 (m_INTEL): New macro.
9158 (processor_target_table): Add "intel".
9159 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9160 with PROCESSOR_INTEL for "intel".
9161 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9162 PROCESSOR_SILVERMONT.
9163 (ix86_issue_rate): Likewise.
9164 (ix86_adjust_cost): Likewise.
9165 (ia32_multipass_dfa_lookahead): Likewise.
9166 (swap_top_of_ready_list): Likewise.
9167 (ix86_sched_reorder): Likewise.
9168 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9169 instead of TARGET_OPT_AGU.
9170 * config/i386/i386.h (TARGET_INTEL): New.
9171 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9172 (processor_type): Add PROCESSOR_INTEL.
9173 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9174 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9176 2014-01-17 Marek Polacek <polacek@redhat.com>
9179 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9182 2014-01-17 Richard Biener <rguenther@suse.de>
9184 PR tree-optimization/46590
9185 * opts.c (default_options_table): Add entries for
9186 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9187 all enabled at -O1 but not for -Og.
9188 * common.opt (fbranch-count-reg): Remove Init(1).
9189 (fmove-loop-invariants): Likewise.
9190 (ftree-pta): Likewise.
9192 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9194 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9195 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9196 decls to at least the GCC 4.8 used alignments.
9199 * tree-nested.c (convert_nonlocal_reference_stmt,
9200 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9201 of GIMPLE_BIND stmts, adjust associated decls.
9203 2014-01-17 Richard Biener <rguenther@suse.de>
9205 PR tree-optimization/46590
9206 * vec.h (vec<>::bseach): New member function implementing
9207 binary search according to C89 bsearch.
9208 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9209 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9210 bitmap pointer again. Make accesses_in_loop a flat array.
9211 (mem_ref_obstack): New global.
9212 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9213 (mark_ref_stored): Likewise.
9214 (ref_indep_loop_p_2): Likewise.
9215 (set_ref_stored_in_loop): New helper function.
9216 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9217 (memref_free): Adjust.
9218 (record_mem_ref_loc): Simplify.
9219 (gather_mem_refs_stmt): Adjust.
9220 (sort_locs_in_loop_postorder_cmp): New function.
9221 (analyze_memory_references): Sort accesses_in_loop after
9222 loop postorder number.
9223 (find_ref_loc_in_loop_cmp): New function.
9224 (for_all_locs_in_loop): Find relevant cluster of locs in
9225 accesses_in_loop and iterate without recursion.
9226 (execute_sm): Avoid uninit warning.
9227 (struct ref_always_accessed): Simplify.
9228 (ref_always_accessed::operator ()): Likewise.
9229 (ref_always_accessed_p): Likewise.
9230 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9231 loop postorder numbers here.
9232 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9235 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9238 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9239 on decls for which assemble_alias has been called.
9241 2014-01-17 Nick Clifton <nickc@redhat.com>
9243 * config/msp430/msp430.opt: (mcpu): New option.
9244 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9245 (msp430_option_override): Parse target_cpu. If the MCU name
9246 matches a generic string, clear target_mcu.
9247 (msp430_attr): Allow numeric interrupt values up to 63.
9248 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9249 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9251 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9253 * config/msp430/msp430.md (popm): Use %J rather than %I.
9254 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9255 (addhi_cy_i): Use immediate_operand for operand 2.
9256 * doc/invoke.texi: Document -mcpu option.
9258 2014-01-17 Richard Biener <rguenther@suse.de>
9260 PR rtl-optimization/38518
9261 * df.h (df_analyze_loop): Declare.
9262 * df-core.c: Include cfgloop.h.
9263 (df_analyze_1): Split out main part of df_analyze.
9264 (df_analyze): Adjust.
9265 (loop_inverted_post_order_compute): New function.
9266 (loop_post_order_compute): Likewise.
9267 (df_analyze_loop): New function avoiding whole-function
9269 * loop-invariant.c (find_defs): Use df_analyze_loop.
9270 (find_invariants): Adjust.
9271 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9273 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9275 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9276 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9278 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9280 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9281 traversal when removing references.
9283 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9286 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9288 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9291 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9292 pushing a reload for an autoinc when we had previously reloaded an
9293 inner part of the address.
9295 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9297 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9299 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9300 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9301 when not giving up or versioning for alias only because of
9303 (vect_analyze_data_ref_dependences): Set to true.
9304 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9306 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9307 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9311 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9314 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9315 operand 0 predicate for gathers, use a new pseudo as subtarget.
9317 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9320 * lra-constraints.c (process_alt_operands): Add printing debug info.
9321 Check absence of input/output reloads for matched operands too.
9323 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9325 PR rtl-optimization/59835
9326 * ira.c (ira_init_register_move_cost): Increase cost for
9329 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9331 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9333 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9336 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9337 non-register objects. Use gen_(high/low)part more consistently.
9340 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9343 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9344 endian support, remove tests for WORDS_BIG_ENDIAN.
9345 (p8_mfvsrd_3_<mode>): Likewise.
9346 (reload_gpr_from_vsx<mode>): Likewise.
9347 (reload_gpr_from_vsxsf): Likewise.
9348 (p8_mfvsrd_4_disf): Likewise.
9350 2014-01-16 Richard Biener <rguenther@suse.de>
9352 PR rtl-optimization/46590
9353 * lcm.c (compute_antinout_edge): Use postorder iteration.
9354 (compute_laterin): Use inverted postorder iteration.
9356 2014-01-16 Nick Clifton <nickc@redhat.com>
9359 * varasm.c (output_constant): Return the number of bytes actually
9361 (output_constructor_array_range): Update the field size with the
9362 number of bytes emitted by output_constant.
9363 (output_constructor_regular_field): Likewise. Also do not
9364 complain if the total number of bytes emitted is now greater
9365 than the expected fieldpos.
9366 * output.h (output_constant): Update prototype and descriptive comment.
9368 2014-01-16 Marek Polacek <polacek@redhat.com>
9371 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9372 it is error_mark_node.
9374 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9376 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9377 VALID_AVX256_REG_OR_OI_MODE.
9379 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9381 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9382 current procedure should be profiled.
9384 2014-01-15 Andrew Pinski <apinski@cavium.com>
9386 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9387 of moving from/to the STACK_REG register class.
9389 2014-01-15 Richard Henderson <rth@redhat.com>
9392 * reginfo.c (global_regs_decl): Globalize.
9393 * rtl.h (global_regs_decl): Declare.
9394 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9395 reserved via global_regs.
9397 2014-01-15 Teresa Johnson <tejohnson@google.com>
9399 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9401 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9403 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9404 and vmulosh rather than call gen_vec_widen_smult_*.
9405 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9406 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9407 (vec_widen_smult_even_v16qi): Likewise.
9408 (vec_widen_umult_even_v8hi): Likewise.
9409 (vec_widen_smult_even_v8hi): Likewise.
9410 (vec_widen_umult_odd_v16qi): Likewise.
9411 (vec_widen_smult_odd_v16qi): Likewise.
9412 (vec_widen_umult_odd_v8hi): Likewise.
9413 (vec_widen_smult_odd_v8hi): Likewise.
9414 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9415 vmuloub rather than call gen_vec_widen_umult_*.
9416 (vec_widen_umult_lo_v16qi): Likewise.
9417 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9418 vmulosb rather than call gen_vec_widen_smult_*.
9419 (vec_widen_smult_lo_v16qi): Likewise.
9420 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9421 rather than call gen_vec_widen_umult_*.
9422 (vec_widen_umult_lo_v8hi): Likewise.
9423 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9424 rather than call gen_vec_widen_smult_*.
9425 (vec_widen_smult_lo_v8hi): Likewise.
9427 2014-01-15 Jeff Law <law@redhat.com>
9429 PR tree-optimization/59747
9430 * ree.c (find_and_remove_re): Properly handle case where a second
9431 eliminated extension requires widening a copy created for elimination
9432 of a prior extension.
9433 (combine_set_extension): Ensure that the number of hard regs needed
9434 for a destination register does not change when we widen it.
9436 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9438 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9439 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9440 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9441 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9442 (avr-*-rtems*): Likewise.
9443 (bfin*-rtems*): Likewise.
9444 (moxie-*-rtems*): Likewise.
9445 (h8300-*-rtems*): Likewise.
9446 (i[34567]86-*-rtems*): Likewise.
9447 (lm32-*-rtems*): Likewise.
9448 (m32r-*-rtems*): Likewise.
9449 (m68k-*-rtems*): Likewise.
9450 (microblaze*-*-rtems*): Likewise.
9451 (mips*-*-rtems*): Likewise.
9452 (powerpc-*-rtems*): Likewise.
9453 (sh-*-rtems*): Likewise.
9454 (sparc-*-rtems*): Likewise.
9455 (sparc64-*-rtems*): Likewise.
9456 (v850-*-rtems*): Likewise.
9457 (m32c-*-rtems*): Likewise.
9459 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9461 PR rtl-optimization/59511
9462 * ira.c (ira_init_register_move_cost): Use memory costs for some
9463 cases of register move cost calculations.
9464 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9465 instead of BB frequency.
9466 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9467 * lra-assigns.c (find_hard_regno_for): Ditto.
9469 2014-01-15 Richard Biener <rguenther@suse.de>
9471 PR tree-optimization/59822
9472 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9473 (vectorizable_load): Use it to hoist defs of uses of invariant
9474 loads out of the loop.
9476 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9477 Kugan Vivekanandarajah <kuganv@linaro.org>
9480 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9483 2014-01-15 Richard Biener <rguenther@suse.de>
9485 PR rtl-optimization/59802
9486 * lcm.c (compute_available): Use inverted postorder to seed
9487 the initial worklist.
9489 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9492 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9493 ADDR_REGS for invalid symrefs in non-PIC code.
9495 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9498 * builtins.c (determine_block_size): Initialize *probable_max_size
9499 even if len_rtx is CONST_INT.
9501 2014-01-14 Andrew Pinski <apinski@cavium.com>
9503 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9504 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9505 (cortexa53_tunings): Likewise.
9506 (aarch64_sched_issue_rate): New function.
9507 (TARGET_SCHED_ISSUE_RATE): Define.
9509 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9511 * ira-costs.c (find_costs_and_classes): Add missed
9512 ira_init_register_move_cost_if_necessary.
9514 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9517 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9519 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9522 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9523 to indicate if type is used for function return value. Warn ABI
9524 change if the vector mode isn't available for function return value.
9525 (ix86_function_arg_advance): Pass false to type_natural_mode.
9526 (ix86_function_arg): Likewise.
9527 (ix86_gimplify_va_arg): Likewise.
9528 (function_arg_32): Don't warn ABI change.
9529 (ix86_function_value): Pass true to type_natural_mode.
9530 (ix86_return_in_memory): Likewise.
9531 (ix86_struct_value_rtx): Removed.
9532 (TARGET_STRUCT_VALUE_RTX): Likewise.
9534 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9536 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9537 converting a conditional jump into a conditional return.
9539 2014-01-14 Richard Biener <rguenther@suse.de>
9541 PR tree-optimization/58921
9542 PR tree-optimization/59006
9543 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9544 hoisting invariant stmts.
9545 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9546 invariant loads on the preheader edge if possible.
9548 2014-01-14 Joey Ye <joey.ye@arm.com>
9550 * doc/plugin.texi (Building GCC plugins): Update to C++.
9552 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9554 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9555 (_mm_rcp28_round_ss): Ditto.
9556 (_mm_rsqrt28_round_sd): Ditto.
9557 (_mm_rsqrt28_round_ss): Ditto.
9558 (_mm_rcp28_sd): Ditto.
9559 (_mm_rcp28_ss): Ditto.
9560 (_mm_rsqrt28_sd): Ditto.
9561 (_mm_rsqrt28_ss): Ditto.
9562 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9563 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9564 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9565 (IX86_BUILTIN_RCP28SD): Ditto.
9566 (IX86_BUILTIN_RCP28SS): Ditto.
9567 (IX86_BUILTIN_RSQRT28SD): Ditto.
9568 (IX86_BUILTIN_RSQRT28SS): Ditto.
9569 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9570 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9571 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9572 (ix86_expand_special_args_builtin): Expand new FTYPE.
9573 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9574 (srcp14<mode>): Make insn unary.
9575 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9576 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9577 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9578 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9579 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9580 Fix rounding: make it SAE only.
9581 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9583 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9585 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9586 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9587 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9588 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9589 (round_saeonly_mask_scalar_operand4): Ditto.
9590 (round_saeonly_mask_scalar_op3): Ditto.
9591 (round_saeonly_mask_scalar_op4): Ditto.
9593 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9595 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9596 Implement -maltivec=be for vec_insert and vec_extract.
9598 2014-01-10 DJ Delorie <dj@redhat.com>
9600 * config/msp430/msp430.md (call_internal): Don't allow memory
9601 references with SP as the base register.
9602 (call_value_internal): Likewise.
9603 * config/msp430/constraints.md (Yc): New. For memory references
9604 that don't use SP as a base register.
9606 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9607 "an integer without a # prefix"
9608 * config/msp430/msp430.md (epilogue_helper): Use it.
9610 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9613 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9614 AVX512F gather builtins.
9615 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9616 on gather decls with INTEGER_TYPE masktype.
9617 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9618 directly into the builtin rather than hoisting it before loop.
9620 PR tree-optimization/59387
9621 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9622 (scev_const_prop): If folded_casts and type has undefined overflow,
9623 use force_gimple_operand instead of force_gimple_operand_gsi and
9624 for each added stmt if it is assign with
9625 arith_code_with_undefined_signed_overflow, call
9626 rewrite_to_defined_overflow.
9627 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9628 gimple-fold.h instead.
9629 (arith_code_with_undefined_signed_overflow,
9630 rewrite_to_defined_overflow): Moved to ...
9631 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9632 rewrite_to_defined_overflow): ... here. No longer static.
9633 Include gimplify-me.h.
9634 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9635 rewrite_to_defined_overflow): New prototypes.
9637 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9639 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9641 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9643 * builtins.c (get_object_alignment_2): Minor tweak.
9644 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9646 2014-01-13 Christian Bruel <christian.bruel@st.com>
9648 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9649 optimized non constant lengths.
9651 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9654 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9655 load as __atomic_load_N if possible.
9657 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9659 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9661 (rs6000_expand_builtin): Adjust call.
9663 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9666 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9667 * config/rs6000/rs6000.c: Include target-globals.h.
9668 (rs6000_set_current_function): Instead of doing target_reinit
9669 unconditionally, use save_target_globals_default_opts and
9670 restore_target_globals.
9672 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9674 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9675 (rs6000_expand_builtin): Handle mffs and mtfsf.
9676 (rs6000_init_builtins): Define mffs and mtfsf.
9677 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9678 (rs6000_mffs): New pattern.
9679 (rs6000_mtfsf): New pattern.
9681 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9683 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9684 Start narrowing with START. Apply candidate-use pair
9685 and check overall cost in narrowing.
9686 (iv_ca_prune): Pass new argument.
9688 2014-01-10 Jeff Law <law@redhat.com>
9691 * ree.c (combine_reaching_defs): Ensure the defining statement
9692 occurs before the extension when optimizing extensions with
9693 different source and destination hard registers.
9695 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9698 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9699 vtables into the type inheritance graph.
9701 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9703 PR rtl-optimization/59754
9704 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9705 modes in the REGNO != REGNO case.
9707 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9709 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9711 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9713 PR tree-optimization/59745
9714 * tree-predcom.c (tree_predictive_commoning_loop): Call
9715 free_affine_expand_cache if giving up because components is NULL.
9717 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9718 GC in payload of target_globals struct instead of allocating them on
9719 the heap and the larger structs separately using GC.
9720 * target-globals.h (struct target_globals): Make regs, hard_regs,
9721 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9722 of GTY((skip)) and change type to void *.
9723 (reset_target_globals): Cast loads from those fields to corresponding
9726 2014-01-10 Steve Ellcey <sellcey@mips.com>
9729 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9730 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9731 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9733 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9736 * aarch64-modes.def (CC_Zmode): New flags mode.
9737 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9738 represents an equality.
9739 (aarch64_get_condition_code): Handle CC_Zmode.
9740 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9742 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9744 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9745 extraction in good case.
9747 2014-01-10 Richard Biener <rguenther@suse.de>
9749 PR tree-optimization/59374
9750 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9751 checking after SLP discovery. Mark stmts not participating
9752 in any SLP instance properly.
9754 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9756 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9757 when handling a SET rtx.
9759 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9761 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9762 (cortex-a57): Likewise.
9763 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9765 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9767 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9768 non-iwmmxt builtins.
9770 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9774 * ipa-devirt.c record_target_from_binfo): Take as argument
9775 stack of binfos and lookup matching one for virtual inheritance.
9776 (possible_polymorphic_call_targets_1): Update.
9778 2014-01-10 Huacai Chen <chenhc@lemote.com>
9780 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9781 kernel strings for Loongson-2E/2F/3A.
9783 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9786 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9787 is_gimple_call before calling gimple_call_internal_p.
9789 2014-01-09 Steve Ellcey <sellcey@mips.com>
9791 * Makefile.in (TREE_FLOW_H): Remove.
9792 (TREE_SSA_H): Add file names from tree-flow.h.
9793 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9794 * tree.h: Remove tree-flow.h reference.
9795 * hash-table.h: Remove tree-flow.h reference.
9796 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9797 reference with tree-ssa-loop.h.
9799 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9801 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9802 default element-order behavior for -maltivec.
9803 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9804 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9805 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9806 when targeting big endian, at least for now.
9807 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9809 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9812 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9813 var satisfies use_register_for_decl, just take into account type
9814 alignment, rather than decl alignment.
9816 PR tree-optimization/59622
9817 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9818 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9819 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9820 Don't devirtualize for inplace at all. For targets.length () == 1,
9821 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9823 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9825 * config/i386/i386.md (cpu): Remove the unused btver1.
9827 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9829 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9831 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9834 * tree-core.h (struct target_globals): New forward declaration.
9835 (struct tree_target_option): Add globals field.
9836 * tree.h (TREE_TARGET_GLOBALS): Define.
9837 (prepare_target_option_nodes_for_pch): New prototype.
9838 * target-globals.h (struct target_globals): Define even if
9840 * tree.c (prepare_target_option_node_for_pch,
9841 prepare_target_option_nodes_for_pch): New functions.
9842 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9843 * config/i386/i386.c: Include target-globals.h.
9844 (ix86_set_current_function): Instead of doing target_reinit
9845 unconditionally, use save_target_globals_default_opts and
9846 restore_target_globals.
9848 2014-01-09 Richard Biener <rguenther@suse.de>
9850 PR tree-optimization/59715
9851 * tree-cfg.h (split_critical_edges): Declare.
9852 * tree-cfg.c (split_critical_edges): Export.
9853 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9855 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9857 * cfgexpand.c (expand_stack_vars): Optionally disable
9858 asan stack protection.
9859 (expand_used_vars): Likewise.
9860 (partition_stack_vars): Likewise.
9861 * asan.c (asan_emit_stack_protection): Optionally disable
9862 after return stack usage.
9863 (instrument_derefs): Optionally disable memory access instrumentation.
9864 (instrument_builtin_call): Likewise.
9865 (instrument_strlen_call): Likewise.
9866 (asan_protect_global): Optionally disable global variables protection.
9867 * doc/invoke.texi: Added doc for new options.
9868 * params.def: Added new options.
9869 * params.h: Likewise.
9871 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9873 PR rtl-optimization/59724
9874 * ifcvt.c (cond_exec_process_if_block): Don't call
9875 flow_find_head_matching_sequence with 0 longest_match.
9876 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9877 non-active insns if !stop_after.
9878 (try_head_merge_bb): Revert 2014-01-07 changes.
9880 2014-01-08 Jeff Law <law@redhat.com>
9882 * ree.c (get_sub_rtx): New function, extracted from...
9883 (merge_def_and_ext): Here.
9884 (combine_reaching_defs): Use get_sub_rtx.
9886 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9888 * cgraph.h (varpool_variable_node): Do not choke on null node.
9890 2014-01-08 Catherine Moore <clm@codesourcery.com>
9892 * config/mips/mips.md (simple_return): Attempt to use JRC
9894 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9896 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9898 PR rtl-optimization/59137
9899 * reorg.c (steal_delay_list_from_target): Call update_block for
9901 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9903 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9905 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9906 two duplicate entries.
9908 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9911 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9913 * config/mips/mips.c (mips_truncated_op_cost): New function.
9914 (mips_rtx_costs): Adjust test for BADDU.
9915 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9917 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9919 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9920 (*baddu_si): ...this new pattern.
9922 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9925 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9927 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9930 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9932 (expand_expr, expand_normal): Adjust.
9933 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9934 inner_reference_p. Use inner_reference_p to expand inner references.
9935 (store_expr): Adjust.
9936 * cfgexpand.c (expand_call_stmt): Adjust.
9938 2014-01-08 Rong Xu <xur@google.com>
9940 * gcov-io.c (gcov_var): Move from gcov-io.h.
9941 (gcov_position): Ditto.
9942 (gcov_is_error): Ditto.
9943 (gcov_rewrite): Ditto.
9944 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9945 only part to libgcc/libgcov.h.
9947 2014-01-08 Marek Polacek <polacek@redhat.com>
9950 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9952 2014-01-08 Marek Polacek <polacek@redhat.com>
9955 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9957 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9959 PR rtl-optimization/59649
9960 * stor-layout.c (get_mode_bounds): For BImode return
9961 0 and STORE_FLAG_VALUE.
9963 2014-01-08 Richard Biener <rguenther@suse.de>
9966 * gimple.h (is_gimple_builtin_call): Remove.
9967 (gimple_builtin_call_types_compatible_p): New.
9968 (gimple_call_builtin_p): New overload.
9969 * gimple.c (is_gimple_builtin_call): Remove.
9970 (validate_call): Rename to ...
9971 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9973 (validate_type): New static function.
9974 (gimple_call_builtin_p): New overload and adjust.
9975 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9976 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9977 (gimple_fold_stmt_to_constant_1): Likewise.
9978 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9980 2014-01-08 Richard Biener <rguenther@suse.de>
9983 * gimplify.c (gimplify_expr): Gimplify register-register type
9984 VIEW_CONVERT_EXPRs to separate stmts.
9986 2014-01-07 Jeff Law <law@redhat.com>
9989 * ree.c (combine_set_extension): Handle case where source
9990 and destination registers in an extension insn are different.
9991 (combine_reaching_defs): Allow source and destination registers
9992 in extension to be different under limited circumstances.
9993 (add_removable_extension): Remove restriction that the
9994 source and destination registers in the extension are the same.
9995 (find_and_remove_re): Emit a copy from the extension's
9996 destination to its source after the defining insn if
9997 the source and destination registers are different.
10000 * ifcvt.c (merge_if_block): If we are merging a block with more than
10001 one successor with a block with no successors, remove any BARRIER
10002 after the second block.
10004 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10006 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10008 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10011 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10012 for 14-bit register offsets when INT14_OK_STRICT is false.
10014 2014-01-07 Roland Stigge <stigge@antcom.de>
10015 Michael Meissner <meissner@linux.vnet.ibm.com>
10018 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10019 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10021 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10023 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10026 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10028 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10029 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10030 rtx is const0_rtx or not.
10032 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10035 * target-globals.c (save_target_globals): Remove this_fn_optab
10037 * toplev.c: Include optabs.h.
10038 (target_reinit): Temporarily restore the global options if another
10039 set of options are in force.
10041 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10043 PR rtl-optimization/58668
10044 * cfgcleanup.c (flow_find_cross_jump): Don't count
10045 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10046 to determine what is counted.
10047 (flow_find_head_matching_sequence): Use active_insn_p to determine
10049 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10051 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10052 determine what is counted.
10054 PR tree-optimization/59643
10055 * tree-predcom.c (split_data_refs_to_components): If one dr is
10056 read and one write, determine_offset fails and the write isn't
10057 in the bad component, just put the read into the bad component.
10059 2014-01-07 Mike Stump <mikestump@comcast.net>
10060 Jakub Jelinek <jakub@redhat.com>
10063 * tree-core.h (struct tree_optimization_option): Change optabs
10064 type from unsigned char * to void *.
10065 * optabs.c (init_tree_optimization_optabs): Adjust
10066 TREE_OPTIMIZATION_OPTABS initialization.
10068 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10071 * config/i386/i386.h (struct machine_function): Add
10072 no_drap_save_restore field.
10073 * config/i386/i386.c (ix86_save_reg): Use
10074 !cfun->machine->no_drap_save_restore instead of
10075 crtl->stack_realign_needed.
10076 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10077 this function clears frame_pointer_needed. Set
10078 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10079 and DRAP reg is needed.
10081 2014-01-06 Marek Polacek <polacek@redhat.com>
10084 * doc/implement-c.texi: Mention that other integer types are
10085 permitted as bit-field types in strictly conforming mode.
10087 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10089 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10090 is newly allocated.
10092 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10094 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10096 2014-01-06 Martin Jambor <mjambor@suse.cz>
10099 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10101 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10103 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10107 * var-tracking.c (add_stores): Preserve the value of the source even if
10108 we don't record the store.
10110 2014-01-06 Terry Guo <terry.guo@arm.com>
10112 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10114 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10117 * config/darwin.c (darwin_function_section): Adjust return values to
10118 correspond to optimisation changes made in r206070.
10120 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10122 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10123 from prefetch_block tune setting.
10124 (nocona_cost): Correct size of prefetch block to 64.
10126 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10128 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10129 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10130 used to save the static chain register in the computation of the offset
10131 from which the FP registers need to be restored.
10133 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10135 PR tree-optimization/59519
10136 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10137 ICE if get_current_def (current_new_name) is already non-NULL, as long
10138 as it is a phi result of some other phi in *new_exit_bb that has
10141 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10142 or vmovdqu* for misaligned_operand.
10143 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10144 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10145 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10146 aligned_mem for AVX512F masked aligned load and store builtins and for
10147 non-temporal moves.
10149 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10151 PR tree-optimization/59651
10152 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10153 Address range for negative step should be added by TYPE_SIZE_UNIT.
10155 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10157 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10158 overlapping registers also for registers other than the stack pointer.
10160 2014-01-03 Marek Polacek <polacek@redhat.com>
10163 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10166 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10169 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10172 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10173 (push splitter): Use <P:MODE_SIZE> instead of
10174 GET_MODE_SIZE (<P:MODE>mode).
10175 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10176 (mov -1, reg peephole2): Likewise.
10177 * config/i386/sse.md (*mov<mode>_internal,
10178 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10179 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10180 *<code><mode>3, *andnot<mode>3<mask_name>,
10181 <mask_codefor><code><mode>3<mask_name>): Likewise.
10182 * config/i386/subst.md (mask_mode512bit_condition,
10183 sd_mask_mode512bit_condition): Likewise.
10185 2014-01-02 Xinliang David Li <davidxl@google.com>
10187 PR tree-optimization/59303
10188 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10189 (dump_predicates): Better output format.
10190 (pred_equal_p): New function.
10191 (is_neq_relop_p): Ditto.
10192 (is_neq_zero_form_p): Ditto.
10193 (pred_expr_equal_p): Ditto.
10194 (pred_neg_p): Ditto.
10195 (simplify_pred): Ditto.
10196 (simplify_preds_2): Ditto.
10197 (simplify_preds_3): Ditto.
10198 (simplify_preds_4): Ditto.
10199 (simplify_preds): Ditto.
10200 (push_pred): Ditto.
10201 (push_to_worklist): Ditto.
10202 (get_pred_info_from_cmp): Ditto.
10203 (is_degenerated_phi): Ditto.
10204 (normalize_one_pred_1): Ditto.
10205 (normalize_one_pred): Ditto.
10206 (normalize_one_pred_chain): Ditto.
10207 (normalize_preds): Ditto.
10208 (normalize_cond_1): Remove function.
10209 (normalize_cond): Ditto.
10210 (is_gcond_subset_of): Ditto.
10211 (is_subset_of_any): Ditto.
10212 (is_or_set_subset_of): Ditto.
10213 (is_and_set_subset_of): Ditto.
10214 (is_norm_cond_subset_of): Ditto.
10215 (pred_chain_length_cmp): Ditto.
10216 (convert_control_dep_chain_into_preds): Type change.
10217 (find_predicates): Ditto.
10218 (find_def_preds): Ditto.
10219 (destroy_predicates_vecs): Ditto.
10220 (find_matching_predicates_in_rest_chains): Ditto.
10221 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10222 (is_pred_expr_subset): Ditto.
10223 (is_pred_chain_subset_of): Ditto.
10224 (is_included_in): Ditto.
10225 (is_superset_of): Ditto.
10227 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10229 Update copyright years.
10231 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10233 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10234 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10235 config/arc/arc.md, config/arc/arc.opt,
10236 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10237 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10238 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10239 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10240 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10241 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10242 vtable-verify.c, vtable-verify.h: Use the standard form for the
10245 2014-01-02 Tobias Burnus <burnus@net-b.de>
10247 * gcc.c (process_command): Update copyright notice dates.
10248 * gcov-dump.c: Ditto.
10250 * doc/cpp.texi: Bump @copying's copyright year.
10251 * doc/cppinternals.texi: Ditto.
10252 * doc/gcc.texi: Ditto.
10253 * doc/gccint.texi: Ditto.
10254 * doc/gcov.texi: Ditto.
10255 * doc/install.texi: Ditto.
10256 * doc/invoke.texi: Ditto.
10258 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10260 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10262 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10264 * config/i386/sse.md (*mov<mode>_internal): Guard
10265 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10267 PR rtl-optimization/59647
10268 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10269 new_rtx into UNSIGNED_FLOAT rtxes.
10271 Copyright (C) 2014 Free Software Foundation, Inc.
10273 Copying and distribution of this file, with or without modification,
10274 are permitted in any medium without royalty provided the copyright
10275 notice and this notice are preserved.