1 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
4 * fold-const.c (fold_binary_loc): Changed width of mask.
6 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
8 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
9 unsigned int initializers for regno_in, regno_out.
11 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
14 * config/avr/avr.md (cc): Add new attribute set_vzn.
15 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
16 Set cc insn attribute to set_vzn instead of set_zn for alternatives
18 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
19 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
20 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
22 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27 * wide-int.cc (UTItype): Define.
28 (UDWtype): Define for appropriate W_TYPE_SIZE.
30 2014-05-09 Richard Biener <rguenther@suse.de>
32 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
33 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
34 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
35 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
36 ssa_propagate): Adjust.
38 2014-05-08 Jeff Law <law@redhat.com>
40 PR tree-optimization/61009
41 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
42 tri-state rather than a boolean. When a block is too big to
43 thread through, inform caller via negative return value.
44 (thread_across_edge): If a block was too big for normal threading,
45 then it's too big for a joiner too, so remove temporary equivalences
46 and return immediately.
48 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
49 Matthias Klose <doko@ubuntu.com>
52 * optc-gen.awk: Fix option handling for -Wunused-parameter.
54 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
57 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
59 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
62 * config/alpha/alpha.c: Include gimple-iterator.h.
63 (alpha_gimple_fold_builtin): New function. Move
64 ALPHA_BUILTIN_UMULH folding from ...
65 (alpha_fold_builtin): ... here.
66 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
68 2014-05-08 Wei Mi <wmi@google.com>
71 * config/i386/i386.c (ix86_compute_frame_layout): Update
72 preferred_stack_boundary for call, expanded from tls descriptor.
73 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
74 to depend on SP register.
75 (*tls_local_dynamic_base_32_gnu): Ditto.
76 (*tls_local_dynamic_32_once): Ditto.
77 (tls_global_dynamic_64_<mode>): Set
78 ix86_tls_descriptor_calls_expanded_in_cfun.
79 (tls_local_dynamic_base_64_<mode>): Ditto.
80 (tls_global_dynamic_32): Set
81 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
82 to depend on SP register.
83 (tls_local_dynamic_base_32): Ditto.
85 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
87 * config/arm/arm_neon.h: Update comment.
88 * config/arm/neon-docgen.ml: Delete.
89 * config/arm/neon-gen.ml: Delete.
90 * doc/arm-neon-intrinsics.texi: Update comment.
92 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
94 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
96 (vand, vorr, veor, vorn, vbic): Remove.
97 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
99 (neon_vsub_unspec): Likewise.
100 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
102 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
104 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
105 (vadd_s16): Likewise.
106 (vadd_s32): Likewise.
107 (vadd_f32): Likewise.
109 (vadd_u16): Likewise.
110 (vadd_u32): Likewise.
111 (vadd_s64): Likewise.
112 (vadd_u64): Likewise.
113 (vaddq_s8): Likewise.
114 (vaddq_s16): Likewise.
115 (vaddq_s32): Likewise.
116 (vaddq_s64): Likewise.
117 (vaddq_f32): Likewise.
118 (vaddq_u8): Likewise.
119 (vaddq_u16): Likewise.
120 (vaddq_u32): Likewise.
121 (vaddq_u64): Likewise.
123 (vmul_s16): Likewise.
124 (vmul_s32): Likewise.
125 (vmul_f32): Likewise.
127 (vmul_u16): Likewise.
128 (vmul_u32): Likewise.
130 (vmulq_s8): Likewise.
131 (vmulq_s16): Likewise.
132 (vmulq_s32): Likewise.
133 (vmulq_f32): Likewise.
134 (vmulq_u8): Likewise.
135 (vmulq_u16): Likewise.
136 (vmulq_u32): Likewise.
138 (vsub_s16): Likewise.
139 (vsub_s32): Likewise.
140 (vsub_f32): Likewise.
142 (vsub_u16): Likewise.
143 (vsub_u32): Likewise.
144 (vsub_s64): Likewise.
145 (vsub_u64): Likewise.
146 (vsubq_s8): Likewise.
147 (vsubq_s16): Likewise.
148 (vsubq_s32): Likewise.
149 (vsubq_s64): Likewise.
150 (vsubq_f32): Likewise.
151 (vsubq_u8): Likewise.
152 (vsubq_u16): Likewise.
153 (vsubq_u32): Likewise.
154 (vsubq_u64): Likewise.
156 (vand_s16): Likewise.
157 (vand_s32): Likewise.
159 (vand_u16): Likewise.
160 (vand_u32): Likewise.
161 (vand_s64): Likewise.
162 (vand_u64): Likewise.
163 (vandq_s8): Likewise.
164 (vandq_s16): Likewise.
165 (vandq_s32): Likewise.
166 (vandq_s64): Likewise.
167 (vandq_u8): Likewise.
168 (vandq_u16): Likewise.
169 (vandq_u32): Likewise.
170 (vandq_u64): Likewise.
172 (vorr_s16): Likewise.
173 (vorr_s32): Likewise.
175 (vorr_u16): Likewise.
176 (vorr_u32): Likewise.
177 (vorr_s64): Likewise.
178 (vorr_u64): Likewise.
179 (vorrq_s8): Likewise.
180 (vorrq_s16): Likewise.
181 (vorrq_s32): Likewise.
182 (vorrq_s64): Likewise.
183 (vorrq_u8): Likewise.
184 (vorrq_u16): Likewise.
185 (vorrq_u32): Likewise.
186 (vorrq_u64): Likewise.
188 (veor_s16): Likewise.
189 (veor_s32): Likewise.
191 (veor_u16): Likewise.
192 (veor_u32): Likewise.
193 (veor_s64): Likewise.
194 (veor_u64): Likewise.
195 (veorq_s8): Likewise.
196 (veorq_s16): Likewise.
197 (veorq_s32): Likewise.
198 (veorq_s64): Likewise.
199 (veorq_u8): Likewise.
200 (veorq_u16): Likewise.
201 (veorq_u32): Likewise.
202 (veorq_u64): Likewise.
204 (vbic_s16): Likewise.
205 (vbic_s32): Likewise.
207 (vbic_u16): Likewise.
208 (vbic_u32): Likewise.
209 (vbic_s64): Likewise.
210 (vbic_u64): Likewise.
211 (vbicq_s8): Likewise.
212 (vbicq_s16): Likewise.
213 (vbicq_s32): Likewise.
214 (vbicq_s64): Likewise.
215 (vbicq_u8): Likewise.
216 (vbicq_u16): Likewise.
217 (vbicq_u32): Likewise.
218 (vbicq_u64): Likewise.
220 (vorn_s16): Likewise.
221 (vorn_s32): Likewise.
223 (vorn_u16): Likewise.
224 (vorn_u32): Likewise.
225 (vorn_s64): Likewise.
226 (vorn_u64): Likewise.
227 (vornq_s8): Likewise.
228 (vornq_s16): Likewise.
229 (vornq_s32): Likewise.
230 (vornq_s64): Likewise.
231 (vornq_u8): Likewise.
232 (vornq_u16): Likewise.
233 (vornq_u32): Likewise.
234 (vornq_u64): Likewise.
236 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
238 * wide-int.cc (UTItype): Define.
239 (UDWtype): Define for appropriate W_TYPE_SIZE.
241 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
243 PR tree-optimization/59100
244 * tree-ssa-phiopt.c: Include tree-inline.h.
245 (neutral_element_p, absorbing_element_p): New functions.
246 (value_replacement): Handle conditional binary operations with a
247 neutral or absorbing element.
249 2014-05-08 Richard Biener <rguenther@suse.de>
251 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
252 folding the expression.
253 (valueize_expr): Remove.
254 (visit_reference_op_load): Do not valueize the result of
256 (simplify_binary_expression): Likewise.
257 (simplify_unary_expression): Likewise.
259 2014-05-08 Richard Biener <rguenther@suse.de>
261 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
262 looking at TYPE_ARG_TYPES.
264 2014-05-08 Richard Biener <rguenther@suse.de>
266 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
267 pointer propagation special-case.
269 2014-05-08 Bin Cheng <bin.cheng@arm.com>
271 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
272 core part of address expressions.
274 2014-05-08 Alan Modra <amodra@gmail.com>
277 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
278 loads and stores when -mno-strict-align at any alignment.
279 (expand_block_clear): Similarly. Also correct calculation of
282 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
285 * tree-complex.c (expand_complex_move): Keep line info when expanding
287 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
288 of complex expression. Use new argument to display correct location
289 for values coming from phi statement.
290 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
291 (warn_uninitialized_phi): Pass location of phi argument to
293 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
294 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
296 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
298 * config/rs6000/predicates.md (indexed_address_mem): New.
299 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
300 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
301 fpstore_ux, fpstore_u.
302 (sign_extend, indexed, update): New.
303 (cell_micro): Adjust.
304 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
305 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
306 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
307 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
308 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
309 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
310 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
311 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
312 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
313 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
314 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
315 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
316 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
317 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
318 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
320 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
321 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
322 *vsx_extract_<mode>_store): Adjust.
323 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
324 is_cracked_insn, insn_must_be_first_in_group,
325 insn_must_be_last_in_group): Adjust.
327 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
329 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
330 ppc440-fpstore): Adjust.
331 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
332 ppc476-fpstore): Adjust.
333 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
334 ppc601-fpstore): Adjust.
335 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
337 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
339 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
340 ppc7450-fpstore): Adjust.
341 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
342 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
343 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
345 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
346 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
347 cell-fpstore, cell-fpstore-update): Adjust.
348 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
349 ppce300c3_store, ppce300c3_fpstore): Adjust.
350 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
351 e500mc_fpstore): Adjust.
352 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
353 e500mc64_store, e500mc64_fpstore): Adjust.
354 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
355 e5500_fpstore): Adjust.
356 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
357 e6500_fpstore): Adjust.
358 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
360 * config/rs6000/power4.md (power4-load, power4-load-ext,
361 power4-load-ext-update, power4-load-ext-update-indexed,
362 power4-load-update-indexed, power4-load-update, power4-fpload,
363 power4-fpload-update, power4-store, power4-store-update,
364 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
366 * config/rs6000/power5.md (power5-load, power5-load-ext,
367 power5-load-ext-update, power5-load-ext-update-indexed,
368 power5-load-update-indexed, power5-load-update, power5-fpload,
369 power5-fpload-update, power5-store, power5-store-update,
370 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
372 * config/rs6000/power6.md (power6-load, power6-load-ext,
373 power6-load-update, power6-load-update-indexed,
374 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
375 power6-fpload-update, power6-store, power6-store-update,
376 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
378 * config/rs6000/power7.md (power7-load, power7-load-ext,
379 power7-load-update, power7-load-update-indexed,
380 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
381 power7-fpload-update, power7-store, power7-store-update,
382 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
384 * config/rs6000/power8.md (power8-load, power8-load-update,
385 power8-load-ext, power8-load-ext-update, power8-fpload,
386 power8-fpload-update, power8-store, power8-store-update-indexed,
387 power8-fpstore, power8-fpstore-update): Adjust.
388 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
390 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
391 titan_lsu_store, titan_lsu_fpstore): Adjust.
392 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
394 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
397 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
398 unrolled byte insns. Emit address increments after move insns.
400 2014-05-07 David Malcolm <dmalcolm@redhat.com>
402 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
403 const_gimple, rather than a gimple.
404 (gimple_call_builtin_p): Likewise, for the three variants.
406 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
407 (gimple_call_builtin_p): Likewise, for the three variants.
409 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
411 PR tree-optimization/61095
412 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
414 2014-05-07 Richard Biener <rguenther@suse.de>
416 PR tree-optimization/61034
417 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
418 (maybe_skip_until): Use translate to take into account
419 lattices when trying to do disambiguations.
420 (get_continuation_for_phi_1): Likewise.
421 (get_continuation_for_phi): Adjust for added translate arguments.
422 (walk_non_aliased_vuses): Likewise.
423 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
424 (walk_non_aliased_vuses): Likewise.
425 (call_may_clobber_ref_p_1): Declare.
426 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
427 calls. Stop early if we are only supposed to disambiguate.
428 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
430 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
432 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
433 Emit an error when the function has arguments.
435 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
437 * cfgloop.h (unswitch_loops): Remove.
438 * doc/passes.texi: Remove references to loop-unswitch.c
439 * timevar.def (TV_LOOP_UNSWITCH): Remove.
441 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
443 * tree-vect-data-refs.c (vect_grouped_load_supported): New
444 check for loads group of length 3.
445 (vect_permute_load_chain): New permutations for loads group of
447 * tree-vect-stmts.c (vect_model_load_cost): Change cost
448 of vec_perm_shuffle for the new permutations.
450 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
452 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
453 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
454 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
455 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
456 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
457 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
458 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
459 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
461 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
463 * loop-unswitch.c: Delete.
465 2014-05-07 Richard Biener <rguenther@suse.de>
467 * config.gcc: Always set need_64bit_hwint to yes.
469 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
471 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
472 of using optimize_size.
474 2014-05-06 Mike Stump <mikestump@comcast.net>
476 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
478 2014-05-06 Joseph Myers <joseph@codesourcery.com>
480 * config/i386/sse.md (*mov<mode>_internal)
481 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
482 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
483 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
484 (*<code><mode>3, *andnot<mode>3<mask_name>)
485 (<mask_codefor><code><mode>3<mask_name>): Only consider
486 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
488 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
491 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
493 * lra-constraints.c (valid_address_p): Move earlier in file.
494 Add a constraint argument to the address_info version.
495 (satisfies_memory_constraint_p): New function.
496 (satisfies_address_constraint_p): Likewise.
497 (process_alt_operands, curr_insn_transform): Use them.
498 (process_address): Pass the constraint to valid_address_p when
499 checking address operands.
501 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
503 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
504 to their respective blocks. Fix inadvertent use of "node".
506 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
508 * emit-rtl.c (init_derived_machine_modes): New functionm, split
510 (init_emit_once): ...here.
511 * rtl.h (init_derived_machine_modes): Declare.
512 * toplev.c (do_compile): Call it even if no_backend.
514 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
515 Mike Stump <mikestump@comcast.net>
516 Richard Sandiford <rdsandiford@googlemail.com>
517 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
519 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
520 (rtx_equal_for_memref_p): Update comment.
521 (adjust_offset_for_component_ref): Use wide-int interfaces.
522 * builtins.c (get_object_alignment_2): Likewise.
523 (c_readstr): Likewise.
524 (target_char_cast): Add comment.
525 (determine_block_size): Use wide-int interfaces.
526 (expand_builtin_signbit): Likewise.
527 (fold_builtin_int_roundingfn): Likewise.
528 (fold_builtin_bitop): Likewise.
529 (fold_builtin_bswap): Likewise.
530 (fold_builtin_logarithm): Use signop.
531 (fold_builtin_pow): Likewise.
532 (fold_builtin_memory_op): Use wide-int interfaces.
533 (fold_builtin_object_size): Likewise.
534 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
535 nb_iterations_estimate.
536 (record_niter_bound): Use wide-int interfaces.
537 (get_estimated_loop_iterations_int): Likewise.
538 (get_estimated_loop_iterations): Likewise.
539 (get_max_loop_iterations): Likewise.
540 * cfgloop.h: Include wide-int.h.
541 (struct nb_iter_bound): Change bound to widest_int.
542 (struct loop): Change nb_iterations_upper_bound and
543 nb_iterations_estimate to widest_int.
544 (record_niter_bound): Switch to use widest_int.
545 (get_estimated_loop_iterations): Likewise.
546 (get_max_loop_iterations): Likewise.
547 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
549 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
550 * combine.c (try_combine): Likewise.
551 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
552 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
554 (aarch64_float_const_representable_p): Likewise.
555 * config/arc/arc.c: Include wide-int.h.
556 (arc_can_use_doloop_p): Use wide-int interfaces.
557 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
558 (vfp3_const_double_index): Likewise.
559 * config/avr/avr.c (avr_out_round): Likewise.
560 (avr_fold_builtin): Likewise.
561 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
562 (bfin_can_use_doloop_p): Likewise.
563 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
564 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
565 * config/i386/i386.c: Include wide-int.h.
566 (ix86_data_alignment): Use wide-int interfaces.
567 (ix86_local_alignment): Likewise.
568 (ix86_emit_swsqrtsf): Update real_from_integer.
569 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
570 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
571 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
572 (zero_constant): Likewise.
573 (input_operand): Likewise.
574 (splat_input_operand): Likewise.
575 (non_logical_cint_operand): Change const_double to const_wide_int.
576 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
577 (easy_altivec_constant): Remove comment.
578 (paired_expand_vector_init): Use CONSTANT_P.
579 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
580 (rs6000_emit_move): Update checks.
581 (rs6000_aggregate_candidate): Use wide-int interfaces.
582 (rs6000_expand_ternop_builtin): Likewise.
583 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
584 (rs6000_assemble_integer): Likewise.
585 (rs6000_hash_constant): Likewise.
586 (output_toc): Likewise.
587 (rs6000_rtx_costs): Likewise.
588 (rs6000_emit_swrsqrt); Update call to real_from_integer.
589 * config/rs6000/rs6000-c.c: Include wide-int.h.
590 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
591 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
592 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
593 Handle CONST_WIDE_INT.
594 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
595 Use tree_fits_uhwi_p.
596 * config/sparc/sparc.c: Include wide-int.h.
597 (sparc_fold_builtin): Use wide-int interfaces.
598 * config/vax/vax.c: Include wide-int.h.
599 (vax_float_literal): Use real_from_integer.
600 * coretypes.h (struct hwivec_def): New.
603 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
604 (equiv_constant): Handle CONST_WIDE_INT.
605 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
606 (cselib_hash_rtx): Handle CONST_WIDE_INT.
607 * dbxout.c (stabstr_U): Use wide-int interfaces.
608 (dbxout_type): Update to use cst_fits_shwi_p.
609 * defaults.h (LOG2_BITS_PER_UNIT): Define.
610 (TARGET_SUPPORTS_WIDE_INT): Add default.
611 * dfp.c: Include wide-int.h.
612 (decimal_real_to_integer2): Use wide-int interfaces and rename to
613 decimal_real_to_integer.
614 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
615 decimal_real_to_integer.
616 * doc/generic.texi (Constant expressions): Update for wide_int.
617 * doc/rtl.texi (const_double): Likewise.
618 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
619 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
620 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
621 (REAL_VALUE_FROM_INT): Remove.
622 (TARGET_SUPPORTS_WIDE_INT): New.
623 * doc/tm.texi: Regenerate.
624 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
625 * double-int.h: Include wide-int.h.
626 (struct wi::int_traits): New.
627 * dwarf2out.c (get_full_len): New.
628 (dw_val_equal_p): Add case dw_val_class_wide_int.
629 (size_of_loc_descr): Likewise.
630 (output_loc_operands): Likewise.
631 (insert_double): Remove.
632 (insert_wide_int): New.
634 (print_die): Add case dw_val_class_wide_int.
635 (attr_checksum): Likewise.
636 (attr_checksum_ordered): Likewise.
637 (same_dw_val_p): Likewise.
638 (size_of_die): Likewise.
639 (value_format): Likewise.
640 (output_die): Likewise.
641 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
643 (clz_loc_descriptor): Use wide-int interfaces.
644 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
645 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
646 (round_up_to_align): Use wide-int interfaces.
647 (field_byte_offset): Likewise.
648 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
649 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
650 CONST_DOUBLE handling. Use wide-int interfaces.
651 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
652 (gen_enumeration_type_die): Use add_AT_wide.
653 (hash_loc_operands): Add case dw_val_class_wide_int.
654 (compare_loc_operands): Likewise.
655 * dwarf2out.h: Include wide-int.h.
657 (enum dw_val_class): Add dw_val_class_wide_int.
658 (struct dw_val_struct): Add val_wide.
659 * emit-rtl.c (const_wide_int_htab): New.
660 (const_wide_int_htab_hash): New.
661 (const_wide_int_htab_eq): New.
662 (lookup_const_wide_int): New.
663 (const_double_htab_hash): Use wide-int interfaces.
664 (const_double_htab_eq): Likewise.
665 (rtx_to_double_int): Conditionally compile for wide-int.
666 (immed_double_int_const): Rename to immed_wide_int_const and
668 (immed_double_const): Conditionally compile for wide-int.
669 (init_emit_once): Use wide-int interfaces.
670 * explow.c (plus_constant): Likewise.
671 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
672 (lshift_value): Use wide-int interfaces.
673 (expand_mult): Likewise.
674 (choose_multiplier): Likewise.
675 (expand_smod_pow2): Likewise.
676 (make_tree): Likewise.
677 * expr.c (convert_modes): Consolidate handling of constants.
678 Use wide-int interfaces.
679 (emit_group_load_1): Add note.
680 (store_expr): Update comment.
681 (get_inner_reference): Use wide-int interfaces.
682 (expand_constructor): Update comment.
683 (expand_expr_real_2): Use wide-int interfaces.
684 (expand_expr_real_1): Likewise.
685 (reduce_to_bit_field_precision): Likewise.
686 (const_vector_from_tree): Likewise.
687 * final.c: Include wide-int-print.h.
688 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
689 * fixed-value.c: Include wide-int.h.
690 (fixed_from_string): Use wide-int interfaces.
691 (fixed_to_decimal): Likewise.
692 (fixed_convert_from_real): Likewise.
693 (real_convert_from_fixed): Likewise.
694 * fold-const.h (mem_ref_offset): Return an offset_int.
695 (div_if_zero_remainder): Remove code parameter.
696 * fold-const.c (div_if_zero_remainder): Remove code parameter.
697 Use wide-int interfaces.
698 (may_negate_without_overflow_p): Use wide-int interfaces.
699 (negate_expr_p): Likewise.
700 (fold_negate_expr): Likewise.
701 (int_const_binop_1): Likewise.
702 (const_binop): Likewise.
703 (fold_convert_const_int_from_int): Likewise.
704 (fold_convert_const_int_from_real): Likewise.
705 (fold_convert_const_int_from_fixed): Likewise.
706 (fold_convert_const_fixed_from_int): Likewise.
707 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
708 (sign_bit_p): Use wide-int interfaces.
709 (make_range_step): Likewise.
710 (build_range_check): Likewise. Pass an integer of the correct type
711 instead of using integer_one_node.
712 (range_predecessor): Pass an integer of the correct type instead
713 of using integer_one_node.
714 (range_successor): Likewise.
715 (merge_ranges): Likewise.
716 (unextend): Use wide-int interfaces.
717 (extract_muldiv_1): Likewise.
718 (fold_div_compare): Likewise.
719 (fold_single_bit_test): Likewise.
720 (fold_sign_changed_comparison): Likewise.
721 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
722 (fold_plusminus_mult_expr): Use wide-int interfaces.
723 (native_encode_int): Likewise.
724 (native_interpret_int): Likewise.
725 (fold_unary_loc): Likewise.
726 (pointer_may_wrap_p): Likewise.
727 (size_low_cst): Likewise.
728 (mask_with_tz): Likewise.
729 (fold_binary_loc): Likewise.
730 (fold_ternary_loc): Likewise.
731 (multiple_of_p): Likewise.
732 (tree_call_nonnegative_warnv_p): Update calls to
733 tree_int_cst_min_precision and real_from_integer.
734 (fold_negate_const): Use wide-int interfaces.
735 (fold_abs_const): Likewise.
736 (fold_relational_const): Use tree_int_cst_lt.
737 (round_up_loc): Use wide-int interfaces.
738 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
739 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
740 * gengtype.c: Remove include of double-int.h.
741 (do_typedef): Use wide-int interfaces.
742 (open_base_files): Add wide-int.h.
743 (main): Add offset_int and widest_int typedefs.
744 * gengtype-lex.l: Handle "^".
745 (CXX_KEYWORD): Add "static".
746 * gengtype-parse.c (require3): New.
747 (require_template_declaration): Handle constant template arguments
748 and nested templates.
749 * gengtype-state.c: Don't include "double-int.h".
750 * genpreds.c (write_one_predicate_function): Update comment.
751 (write_tm_constrs_h): Add check for hval and lval use in
753 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
754 (add_to_sequence): Likewise.
755 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
756 and const_double_operand.
757 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
759 * gimple-fold.c (get_base_constructor): Likewise.
760 (fold_array_ctor_reference): Likewise.
761 (fold_nonarray_ctor_reference): Likewise.
762 (fold_const_aggregate_ref_1): Likewise.
763 (gimple_val_nonnegative_real_p): Likewise.
764 (gimple_fold_indirect_ref): Likewise.
765 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
766 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
767 (struct slsr_cand_d): Change index to be widest_int.
768 (struct incr_info_d): Change incr to be widest_int.
769 (alloc_cand_and_find_basis): Use wide-int interfaces.
770 (slsr_process_phi): Likewise.
771 (backtrace_base_for_ref): Likewise. Return a widest_int.
772 (restructure_reference): Take a widest_int instead of a double_int.
773 (slsr_process_ref): Use wide-int interfaces.
774 (create_mul_ssa_cand): Likewise.
775 (create_mul_imm_cand): Likewise.
776 (create_add_ssa_cand): Likewise.
777 (create_add_imm_cand): Take a widest_int instead of a double_int.
778 (slsr_process_add): Use wide-int interfaces.
779 (slsr_process_cast): Likewise.
780 (slsr_process_copy): Likewise.
781 (dump_candidate): Likewise.
782 (dump_incr_vec): Likewise.
783 (replace_ref): Likewise.
784 (cand_increment): Likewise. Return a widest_int.
785 (cand_abs_increment): Likewise.
786 (replace_mult_candidate): Take a widest_int instead of a double_int.
787 (replace_unconditional_candidate): Use wide-int interfaces.
788 (incr_vec_index): Take a widest_int instead of a double_int.
789 (create_add_on_incoming_edge): Likewise.
790 (create_phi_basis): Use wide-int interfaces.
791 (replace_conditional_candidate): Likewise.
792 (record_increment): Take a widest_int instead of a double_int.
793 (record_phi_increments): Use wide-int interfaces.
794 (phi_incr_cost): Take a widest_int instead of a double_int.
795 (lowest_cost_path): Likewise.
796 (total_savings): Likewise.
797 (analyze_increments): Use wide-int interfaces.
798 (ncd_with_phi): Take a widest_int instead of a double_int.
799 (ncd_of_cand_and_phis): Likewise.
800 (nearest_common_dominator_for_cands): Likewise.
801 (insert_initializers): Use wide-int interfaces.
802 (all_phi_incrs_profitable): Likewise.
803 (replace_one_candidate): Likewise.
804 (replace_profitable_candidates): Likewise.
805 * godump.c: Include wide-int-print.h.
806 (go_output_typedef): Use wide-int interfaces.
807 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
808 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
809 (build_loop_iteration_domains): Likewise.
810 * hooks.h: Include wide-int.h rather than double-int.h.
811 (hook_bool_dint_dint_uint_bool_true): Delete.
812 (hook_bool_wint_wint_uint_bool_true): Declare.
813 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
814 (hook_bool_wint_wint_uint_bool_true): New.
815 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
817 (ubsan_expand_si_overflow_mul_check): Likewise.
818 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
819 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
820 (get_ancestor_addr_info): Likewise.
821 (ipa_modify_call_arguments): Likewise.
822 * loop-doloop.c (doloop_modify): Likewise.
823 (doloop_optimize): Likewise.
824 * loop-iv.c (iv_number_of_iterations): Likewise.
825 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
826 (unroll_loop_constant_iterations): Likewise.
827 (decide_unroll_runtime_iterations): Likewise.
828 (unroll_loop_runtime_iterations): Likewise.
829 (decide_peel_simple): Likewise.
830 (decide_unroll_stupid): Likewise.
831 * lto-streamer-in.c (streamer_read_wi): Add.
832 (input_cfg): Use wide-int interfaces.
833 (lto_input_tree_1): Likewise.
834 * lto-streamer-out.c (streamer_write_wi): Add.
835 (hash_tree): Use wide-int interfaces.
836 (output_cfg): Likewise.
837 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
838 (GTFILES): Add wide-int.h and signop.h.
839 (TAGS): Look for .cc files too.
840 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
841 * optabs.c (expand_subword_shift): Likewise.
842 (expand_doubleword_shift): Likewise.
843 (expand_absneg_bit): Likewise.
844 (expand_copysign_absneg): Likewise.
845 (expand_copysign_bit): Likewise.
846 * postreload.c (reload_cse_simplify_set): Likewise.
847 * predict.c (predict_iv_comparison): Likewise.
848 * pretty-print.h: Include wide-int-print.h.
850 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
851 * print-tree.c: Include wide-int-print.h.
852 (print_node_brief): Use wide-int interfaces.
853 (print_node): Likewise.
854 * read-rtl.c (validate_const_wide_int): New.
855 (read_rtx_code): Add CONST_WIDE_INT case.
856 * real.c: Include wide-int.h.
857 (real_to_integer2): Delete.
858 (real_to_integer): New function, returning a wide_int.
859 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
860 (ten_to_ptwo): Update call to real_from_integer.
861 (real_digit): Likewise.
862 * real.h: Include signop.h, wide-int.h and insn-modes.h.
863 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
864 (REAL_VALUE_TO_INT): Delete.
865 (real_to_integer): Declare a wide-int form.
866 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
867 * recog.c (const_int_operand): Improve comment.
868 (const_scalar_int_operand): New.
869 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
870 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
871 (split_double): Likewise.
872 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
873 (rtx_size): Likewise.
874 (rtx_alloc_stat_v): New.
875 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
876 (cwi_output_hex): New.
877 (iterative_hash_rtx): Handle CONST_WIDE_INT.
878 (cwi_check_failed_bounds): New.
879 * rtl.def (CONST_WIDE_INT): New.
880 * rtl.h: Include <utility> and wide-int.h.
881 (struct hwivec_def): New.
882 (CWI_GET_NUM_ELEM): New.
883 (CWI_PUT_NUM_ELEM): New.
884 (struct rtx_def): Add num_elem and hwiv.
885 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
886 (CASE_CONST_UNIQUE): Likewise.
887 (CASE_CONST_ANY): Likewise.
888 (CONST_SCALAR_INT_P): Likewise.
889 (CONST_WIDE_INT_P): New.
892 (cwi_check_failed_bounds): New.
895 (CONST_WIDE_INT_VEC) New.
896 (CONST_WIDE_INT_NUNITS) New.
897 (CONST_WIDE_INT_ELT) New.
898 (rtx_mode_t): New type.
899 (wi::int_traits <rtx_mode_t>): New.
901 (wi::min_value): New.
902 (wi::max_value): New.
904 (const_wide_int_alloc): New.
905 (immed_wide_int_const): New.
906 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
907 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
908 * signop.h: New file.
909 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
910 (simplify_const_unary_operation): Use wide-int interfaces.
911 (simplify_binary_operation_1): Likewise.
912 (simplify_const_binary_operation): Likewise.
913 (simplify_const_relational_operation): Likewise.
914 (simplify_immed_subreg): Likewise.
915 * stmt.c (expand_case): Likewise.
916 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
917 signop rather than a bool.
918 * stor-layout.c (layout_type): Use wide-int interfaces.
919 (initialize_sizetypes): Update calls to
920 set_min_and_max_values_for_integral_type.
921 (set_min_and_max_values_for_integral_type): Take a signop rather
922 than a bool. Use wide-int interfaces.
923 (fixup_signed_type): Update accordingly. Remove
924 HOST_BITS_PER_DOUBLE_INT limit.
925 (fixup_unsigned_type): Likewise.
926 * system.h (STATIC_CONSTANT_P): New.
927 (STATIC_ASSERT): New.
928 * target.def (can_use_doloop_p): Take widest_ints rather than
930 * target.h: Include wide-int.h rather than double-int.h.
931 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
933 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
934 rather than INT_CST_LT_UNSIGNED.
935 (can_use_doloop_if_innermost): Take widest_ints rather than
937 * tree-affine.c: Include wide-int-print.h.
938 (double_int_ext_for_comb): Delete.
939 (wide_int_ext_for_comb): New.
940 (aff_combination_zero): Use wide-int interfaces.
941 (aff_combination_const): Take a widest_int instead of a double_int.
942 (aff_combination_elt): Use wide-int interfaces.
943 (aff_combination_scale): Take a widest_int instead of a double_int.
944 (aff_combination_add_elt): Likewise.
945 (aff_combination_add_cst): Likewise.
946 (aff_combination_add): Use wide-int interfaces.
947 (aff_combination_convert): Likewise.
948 (tree_to_aff_combination): Likewise.
949 (add_elt_to_tree): Take a widest_int instead of a double_int.
950 (aff_combination_to_tree): Use wide-int interfaces.
951 (aff_combination_remove_elt): Likewise.
952 (aff_combination_add_product): Take a widest_int instead of
954 (aff_combination_mult): Use wide-int interfaces.
955 (aff_combination_expand): Likewise.
956 (double_int_constant_multiple_p): Delete.
957 (wide_int_constant_multiple_p): New.
958 (aff_combination_constant_multiple_p): Take a widest_int pointer
959 instead of a double_int pointer.
960 (print_aff): Use wide-int interfaces.
961 (get_inner_reference_aff): Take a widest_int pointer
962 instead of a double_int pointer.
963 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
964 * tree-affine.h: Include wide-int.h.
965 (struct aff_comb_elt): Change type of coef to widest_int.
966 (struct affine_tree_combination): Change type of offset to widest_int.
967 (double_int_ext_for_comb): Delete.
968 (wide_int_ext_for_comb): New.
969 (aff_combination_const): Use widest_int instead of double_int.
970 (aff_combination_scale): Likewise.
971 (aff_combination_add_elt): Likewise.
972 (aff_combination_constant_multiple_p): Likewise.
973 (get_inner_reference_aff): Likewise.
974 (aff_comb_cannot_overlap_p): Likewise.
975 (aff_combination_zero_p): Use wide-int interfaces.
976 * tree.c: Include tree.h.
977 (init_ttree): Use make_int_cst.
978 (tree_code_size): Removed code for INTEGER_CST case.
979 (tree_size): Add INTEGER_CST case.
980 (make_node_stat): Update comment.
981 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
982 (build_int_cst_type): Use wide-int interfaces.
983 (double_int_to_tree): Likewise.
984 (double_int_fits_to_tree_p): Delete.
985 (force_fit_type_double): Delete.
986 (force_fit_type): New.
987 (int_cst_hash_hash): Use wide-int interfaces.
988 (int_cst_hash_eq): Likewise.
989 (build_int_cst_wide): Delete.
990 (wide_int_to_tree): New.
991 (cache_integer_cst): Use wide-int interfaces.
992 (build_low_bits_mask): Likewise.
993 (cst_and_fits_in_hwi): Likewise.
994 (real_value_from_int_cst): Likewise.
995 (make_int_cst_stat): New.
996 (integer_zerop): Use wide_int interfaces.
997 (integer_onep): Likewise.
998 (integer_all_onesp): Likewise.
999 (integer_pow2p): Likewise.
1000 (integer_nonzerop): Likewise.
1001 (tree_log2): Likewise.
1002 (tree_floor_log2): Likewise.
1003 (tree_ctz): Likewise.
1004 (int_size_in_bytes): Likewise.
1005 (mem_ref_offset): Return an offset_int rather than a double_int.
1006 (build_type_attribute_qual_variant): Use wide_int interfaces.
1007 (type_hash_eq): Likewise
1008 (tree_int_cst_equal): Likewise.
1009 (tree_int_cst_lt): Delete.
1010 (tree_int_cst_compare): Likewise.
1011 (tree_fits_shwi_p): Use wide_int interfaces.
1012 (tree_fits_uhwi_p): Likewise.
1013 (tree_int_cst_sign_bit): Likewise.
1014 (tree_int_cst_sgn): Likewise.
1015 (tree_int_cst_min_precision): Take a signop rather than a bool.
1016 (simple_cst_equal): Use wide_int interfaces.
1017 (compare_tree_int): Likewise.
1018 (iterative_hash_expr): Likewise.
1019 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1021 (get_type_static_bounds): Use wide_int interfaces.
1022 (tree_int_cst_elt_check_failed): New.
1023 (build_common_tree_nodes): Reordered to set prec before filling in
1025 (int_cst_value): Check cst_and_fits_in_hwi.
1026 (widest_int_cst_value): Use wide_int interfaces.
1027 (upper_bound_in_type): Likewise.
1028 (lower_bound_in_type): Likewise.
1029 (num_ending_zeros): Likewise.
1030 (drop_tree_overflow): Likewise.
1031 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1032 (gen_conditions_for_pow_cst_base): Likewise.
1033 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1034 (group_case_labels_stmt): Use wide-int interfaces.
1035 (verify_gimple_assign_binary): Likewise.
1036 (print_loop): Likewise.
1037 * tree-chrec.c (tree_fold_binomial): Likewise.
1038 * tree-core.h (struct tree_base): Add int_length.
1039 (struct tree_int_cst): Change rep of value.
1040 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1041 (dr_may_alias_p): Likewise.
1042 (max_stmt_executions_tree): Likewise.
1043 * tree.def (INTEGER_CST): Update comment.
1044 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1045 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1046 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1047 (dequeue_and_dump): Use wide-int interfaces.
1048 * tree.h: Include wide-int.h.
1049 (NULL_TREE): Moved to earlier loc in file.
1050 (TREE_INT_CST_ELT_CHECK): New.
1051 (tree_int_cst_elt_check_failed): New.
1053 (TREE_INT_CST): Delete.
1054 (TREE_INT_CST_LOW): Use wide-int interfaces.
1055 (TREE_INT_CST_HIGH): Delete.
1056 (TREE_INT_CST_NUNITS): New.
1057 (TREE_INT_CST_EXT_NUNITS): Likewise.
1058 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1059 (TREE_INT_CST_ELT): Likewise.
1060 (INT_CST_LT): Delete.
1061 (tree_int_cst_elt_check): New (two forms).
1062 (type_code_size): Update comment.
1063 (make_int_cst_stat, make_int_cst): New.
1064 (tree_to_double_int): Delete.
1065 (double_int_fits_to_tree_p): Delete.
1066 (force_fit_type_double): Delete.
1067 (build_int_cstu): Replace with out-of-line function.
1068 (build_int_cst_wide): Delete.
1069 (tree_int_cst_lt): Define inline.
1070 (tree_int_cst_le): New.
1071 (tree_int_cst_compare): Define inline.
1072 (tree_int_cst_min_precision): Take a signop rather than a bool.
1073 (wi::int_traits <const_tree>): New.
1074 (wi::int_traits <tree>): New.
1075 (wi::extended_tree): New.
1076 (wi::int_traits <wi::extended_tree>): New.
1077 (wi::to_widest): New.
1078 (wi::to_offset): New.
1079 (wi::fits_to_tree_p): New.
1080 (wi::min_value): New.
1081 (wi::max_value): New.
1082 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1083 (copy_tree_body_r): Likewise.
1084 * tree-object-size.c (compute_object_offset): Likewise.
1085 (addr_object_size): Likewise.
1086 * tree-predcom.c: Include wide-int-print.h.
1087 (struct dref_d): Change type of offset to widest_int.
1088 (dump_dref): Call wide-int printer.
1089 (aff_combination_dr_offset): Use wide-int interfaces.
1090 (determine_offset): Take a widest_int pointer rather than a
1092 (split_data_refs_to_components): Use wide-int interfaces.
1093 (suitable_component_p): Likewise.
1094 (order_drefs): Likewise.
1095 (add_ref_to_chain): Likewise.
1096 (valid_initializer_p): Likewise.
1097 (determine_roots_comp): Likewise.
1098 * tree-pretty-print.c: Include wide-int-print.h.
1099 (dump_generic_node): Use wide-int interfaces.
1100 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1101 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1102 (move_fixed_address_to_symbol): Likewise.
1103 (move_hint_to_base): Likewise.
1104 (move_pointer_to_base): Likewise.
1105 (move_variant_to_index): Likewise.
1106 (most_expensive_mult_to_index): Likewise.
1107 (addr_to_parts): Likewise.
1108 (copy_ref_info): Likewise.
1109 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1110 (indirect_refs_may_alias_p): Likewise.
1111 (stmt_kills_ref_p_1): Likewise.
1112 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1113 * tree-ssa-ccp.c: Update comment at top of file. Include
1115 (struct prop_value_d): Change type of mask to widest_int.
1116 (extend_mask): New function.
1117 (dump_lattice_value): Use wide-int interfaces.
1118 (get_default_value): Likewise.
1119 (set_constant_value): Likewise.
1120 (set_value_varying): Likewise.
1121 (valid_lattice_transition): Likewise.
1122 (set_lattice_value): Likewise.
1123 (value_to_double_int): Delete.
1124 (value_to_wide_int): New.
1125 (get_value_from_alignment): Use wide-int interfaces.
1126 (get_value_for_expr): Likewise.
1127 (do_dbg_cnt): Likewise.
1128 (ccp_finalize): Likewise.
1129 (ccp_lattice_meet): Likewise.
1130 (bit_value_unop_1): Use widest_ints rather than double_ints.
1131 (bit_value_binop_1): Likewise.
1132 (bit_value_unop): Use wide-int interfaces.
1133 (bit_value_binop): Likewise.
1134 (bit_value_assume_aligned): Likewise.
1135 (evaluate_stmt): Likewise.
1136 (ccp_fold_stmt): Likewise.
1137 (visit_cond_stmt): Likewise.
1138 (ccp_visit_stmt): Likewise.
1139 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1140 (constant_pointer_difference): Likewise.
1141 (associate_pointerplus): Likewise.
1142 (combine_conversions): Likewise.
1143 * tree-ssa-loop.h: Include wide-int.h.
1144 (struct tree_niter_desc): Change type of max to widest_int.
1145 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1146 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1147 (remove_redundant_iv_tests): Likewise.
1148 (canonicalize_loop_induction_variables): Likewise.
1149 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1150 (constant_multiple_of): Take a widest_int pointer instead of
1151 a double_int pointer.
1152 (get_computation_aff): Use wide-int interfaces.
1153 (ptr_difference_cost): Likewise.
1154 (difference_cost): Likewise.
1155 (get_loop_invariant_expr_id): Likewise.
1156 (get_computation_cost_at): Likewise.
1157 (iv_elimination_compare_lt): Likewise.
1158 (may_eliminate_iv): Likewise.
1159 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1160 instead of double_int.
1161 (max_loop_iterations): Likewise.
1162 (max_stmt_executions): Likewise.
1163 (estimated_stmt_executions): Likewise.
1164 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1165 (split_to_var_and_offset): Use wide-int interfaces.
1166 (determine_value_range): Likewise.
1167 (bound_difference_of_offsetted_base): Likewise.
1168 (bounds_add): Take a widest_int instead of a double_int.
1169 (number_of_iterations_ne_max): Use wide-int interfaces.
1170 (number_of_iterations_ne): Likewise.
1171 (number_of_iterations_lt_to_ne): Likewise.
1172 (assert_loop_rolls_lt): Likewise.
1173 (number_of_iterations_lt): Likewise.
1174 (number_of_iterations_le): Likewise.
1175 (number_of_iterations_cond): Likewise.
1176 (number_of_iterations_exit): Likewise.
1177 (finite_loop_p): Likewise.
1178 (derive_constant_upper_bound_assign): Likewise.
1179 (derive_constant_upper_bound): Return a widest_int.
1180 (derive_constant_upper_bound_ops): Likewise.
1181 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1182 (record_estimate): Take a widest_int rather than a double_int.
1183 (record_nonwrapping_iv): Use wide-int interfaces.
1184 (double_int_cmp): Delete.
1185 (wide_int_cmp): New.
1186 (bound_index): Take a widest_int rather than a double_int.
1187 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1188 (maybe_lower_iteration_bound): Likewise.
1189 (estimate_numbers_of_iterations_loop): Likewise.
1190 (estimated_loop_iterations): Take a widest_int pointer than than
1191 a double_int pointer.
1192 (estimated_loop_iterations_int): Use wide-int interfaces.
1193 (max_loop_iterations): Take a widest_int pointer than than
1194 a double_int pointer.
1195 (max_loop_iterations_int): Use wide-int interfaces.
1196 (max_stmt_executions): Take a widest_int pointer than than
1197 a double_int pointer.
1198 (estimated_stmt_executions): Likewise.
1199 (n_of_executions_at_most): Use wide-int interfaces.
1200 (scev_probably_wraps_p): Likewise.
1201 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1203 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1205 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1206 double_ints. Adjust for trailing_wide_ints <3> representation.
1207 (set_nonzero_bits): Likewise.
1208 (get_range_info): Return wide_ints rather than double_ints.
1209 Adjust for trailing_wide_ints <3> representation.
1210 (get_nonzero_bits): Likewise.
1211 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1213 * tree-ssanames.h (struct range_info_def): Replace min, max and
1214 nonzero_bits with a trailing_wide_ints <3>.
1215 (set_range_info): Use wide_int_refs rather than double_ints.
1216 (set_nonzero_bits): Likewise.
1217 (get_range_info): Return wide_ints rather than double_ints.
1218 (get_nonzero_bits): Likewise.
1219 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1220 * tree-ssa-pre.c (phi_translate_1): Likewise.
1221 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1222 (acceptable_pow_call): Likewise.
1223 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1225 (vn_reference_fold_indirect): Likewise.
1226 (vn_reference_maybe_forwprop_address): Likewise.
1227 (valueize_refs_1): Likewise.
1228 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1229 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1230 tree_int_cst_lt and tree_int_cst_le.
1231 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1233 (streamer_alloc_tree): Likewise.
1234 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1235 (streamer_write_tree_header): Likewise.
1236 (streamer_write_integer_cst): Likewise.
1237 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1238 (build_constructors): Likewise.
1239 (array_value_type): Likewise.
1240 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1241 (vect_check_gather): Likewise.
1242 * tree-vect-generic.c (build_replicated_const): Likewise.
1243 (expand_vector_divmod): Likewise.
1244 * tree-vect-loop.c (vect_transform_loop): Likewise.
1245 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1246 (vect_do_peeling_for_alignment): Likewise.
1247 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1248 * tree-vrp.c: Include wide-int.h.
1249 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1250 (extract_range_from_assert): Use wide-int interfaces.
1251 (vrp_int_const_binop): Likewise.
1252 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1253 double_int pointers.
1254 (ranges_from_anti_range): Use wide-int interfaces.
1255 (quad_int_cmp): Delete.
1256 (quad_int_pair_sort): Likewise.
1257 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1258 (extract_range_from_unary_expr_1): Likewise.
1259 (adjust_range_with_scev): Likewise.
1260 (masked_increment): Take and return wide_ints rather than double_ints.
1261 (register_edge_assert_for_2): Use wide-int interfaces.
1262 (check_array_ref): Likewise.
1263 (search_for_addr_array): Likewise.
1264 (maybe_set_nonzero_bits): Likewise.
1265 (union_ranges): Pass an integer of the correct type instead of
1266 using integer_one_node.
1267 (intersect_ranges): Likewise.
1268 (simplify_truth_ops_using_ranges): Likewise.
1269 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1270 (range_fits_type_p): Likewise.
1271 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1273 (simplify_conversion_using_ranges): Use wide-int interfaces.
1274 (simplify_float_conversion_using_ranges): Likewise.
1275 (vrp_finalize): Likewise.
1276 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1277 (gimple_stringops_transform): Likewise.
1278 * varasm.c (decode_addr_const): Likewise.
1279 (const_hash_1): Likewise.
1280 (const_rtx_hash_1): Likewise
1281 (output_constant): Likewise.
1282 (array_size_for_constructor): Likewise.
1283 (output_constructor_regular_field): Likewise.
1284 (output_constructor_bitfield): Likewise.
1285 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1286 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1288 * gencheck.c: Define BITS_PER_UNIT.
1291 * wide-int-print.cc: New.
1292 * wide-int-print.h: New.
1294 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1296 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1298 2014-05-06 Richard Biener <rguenther@suse.de>
1300 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1301 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1302 (TODO_verify_all): Adjust.
1303 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1304 TODO_verify_stmts and TODO_verify_rtl_sharing.
1305 * bb-reorder.c: Likewise.
1306 * cfgexpand.c: Likewise.
1307 * cprop.c: Likewise.
1309 * function.c: Likewise.
1310 * fwprop.c: Likewise.
1312 * gimple-ssa-isolate-paths.c: Likewise.
1313 * gimple-ssa-strength-reduction.c: Likewise.
1314 * ipa-split.c: Likewise.
1315 * loop-init.c: Likewise.
1316 * loop-unroll.c: Likewise.
1317 * lower-subreg.c: Likewise.
1318 * modulo-sched.c: Likewise.
1319 * postreload-gcse.c: Likewise.
1320 * predict.c: Likewise.
1321 * recog.c: Likewise.
1322 * sched-rgn.c: Likewise.
1323 * store-motion.c: Likewise.
1324 * tracer.c: Likewise.
1325 * trans-mem.c: Likewise.
1326 * tree-call-cdce.c: Likewise.
1327 * tree-cfg.c: Likewise.
1328 * tree-cfgcleanup.c: Likewise.
1329 * tree-complex.c: Likewise.
1330 * tree-eh.c: Likewise.
1331 * tree-emutls.c: Likewise.
1332 * tree-if-conv.c: Likewise.
1333 * tree-into-ssa.c: Likewise.
1334 * tree-loop-distribution.c: Likewise.
1335 * tree-object-size.c: Likewise.
1336 * tree-parloops.c: Likewise.
1337 * tree-pass.h: Likewise.
1338 * tree-sra.c: Likewise.
1339 * tree-ssa-ccp.c: Likewise.
1340 * tree-ssa-copy.c: Likewise.
1341 * tree-ssa-copyrename.c: Likewise.
1342 * tree-ssa-dce.c: Likewise.
1343 * tree-ssa-dom.c: Likewise.
1344 * tree-ssa-dse.c: Likewise.
1345 * tree-ssa-forwprop.c: Likewise.
1346 * tree-ssa-ifcombine.c: Likewise.
1347 * tree-ssa-loop-ch.c: Likewise.
1348 * tree-ssa-loop-ivcanon.c: Likewise.
1349 * tree-ssa-loop.c: Likewise.
1350 * tree-ssa-math-opts.c: Likewise.
1351 * tree-ssa-phiopt.c: Likewise.
1352 * tree-ssa-phiprop.c: Likewise.
1353 * tree-ssa-pre.c: Likewise.
1354 * tree-ssa-reassoc.c: Likewise.
1355 * tree-ssa-sink.c: Likewise.
1356 * tree-ssa-strlen.c: Likewise.
1357 * tree-ssa-tail-merge.c: Likewise.
1358 * tree-ssa-uncprop.c: Likewise.
1359 * tree-switch-conversion.c: Likewise.
1360 * tree-tailcall.c: Likewise.
1361 * tree-vect-generic.c: Likewise.
1362 * tree-vectorizer.c: Likewise.
1363 * tree-vrp.c: Likewise.
1365 * var-tracking.c: Likewise.
1366 * bt-load.c: Likewise.
1367 * cfgcleanup.c: Likewise.
1368 * combine-stack-adj.c: Likewise.
1369 * combine.c: Likewise.
1370 * compare-elim.c: Likewise.
1371 * config/epiphany/resolve-sw-modes.c: Likewise.
1372 * config/i386/i386.c: Likewise.
1373 * config/mips/mips.c: Likewise.
1374 * config/s390/s390.c: Likewise.
1375 * config/sh/sh_treg_combine.cc: Likewise.
1376 * config/sparc/sparc.c: Likewise.
1379 * final.c: Likewise.
1380 * ifcvt.c: Likewise.
1381 * mode-switching.c: Likewise.
1382 * passes.c: Likewise.
1383 * postreload.c: Likewise.
1385 * reg-stack.c: Likewise.
1386 * regcprop.c: Likewise.
1387 * regrename.c: Likewise.
1390 2014-05-06 Richard Biener <rguenther@suse.de>
1393 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1394 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1396 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1399 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1401 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1402 Tom de Vries <tom@codesourcery.com>
1404 * target.def (call_fusage_contains_non_callee_clobbers): New
1406 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1408 (@node Miscellaneous Register Hooks): New node.
1409 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1410 * doc/tm.texi: Regenerate.
1412 2014-05-05 Marek Polacek <polacek@redhat.com>
1415 * opts.c (common_handle_option): Call error_at instead of warning_at.
1417 2014-05-05 Richard Biener <rguenther@suse.de>
1419 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1420 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1421 under the TODO_verify_il umbrella.
1423 2014-05-05 Richard Biener <rguenther@suse.de>
1425 * passes.c (execute_function_todo): Move TODO_verify_flow under
1426 the TODO_verify_ul umbrella.
1428 2014-05-05 Richard Biener <rguenther@suse.de>
1431 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1432 X & CST away from a CST that is the mask of a mode.
1434 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1436 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1437 int argument to enum machine_mode.
1438 (picochip_class_max_nregs): Ditto.
1439 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1440 (picochip_class_max_nregs): Ditto.
1442 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1444 * target.def: Add new target hook.
1445 * doc/tm.texi: Regenerate.
1446 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1447 * targhooks.c (default_keep_leaf_when_profiled): New function.
1449 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1450 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1452 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1454 PR tree-optimization/60363
1455 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1456 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1457 (update_destination_phis): New parameter.
1458 (create_edge_and_update_destination_phis): Ditto.
1459 (ssa_fix_duplicate_block_edges): Pass new arguments.
1460 (thread_single_edge): Ditto.
1462 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1464 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1465 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1466 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1467 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1468 Use RS6000_BTM_HARD_FLOAT.
1469 (BU_MISC_2): Likewise.
1470 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1471 RS6000_BTM_HARD_FLOAT.
1472 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1474 (rs6000_invalid_builtin): Add hard floating builtin support.
1475 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1476 hard float builtins.
1477 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1479 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1481 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1482 Add missing function* argument.
1484 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1486 * lra-constraints.c (valid_address_p): Move earlier in file.
1487 Add a constraint argument to the address_info version.
1488 (satisfies_memory_constraint_p): New function.
1489 (satisfies_address_constraint_p): Likewise.
1490 (process_alt_operands, curr_insn_transform): Use them.
1491 (process_address): Pass the constraint to valid_address_p when
1492 checking address operands.
1494 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1496 * config/mips/mips.c (mips_isa_rev): New variable.
1497 (mips_set_architecture): Set it.
1498 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1500 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1501 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1502 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1503 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1504 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1505 conditions in terms of mips_isa_rev.
1506 (mips_isa_rev): Declare.
1508 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1510 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1511 (prob_unlikely, prob_likely): Make variables const.
1513 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1515 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1517 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1519 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1521 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1523 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1524 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1525 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1527 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1528 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1530 Replace usage of ROUND_REG with sh_round_reg.
1531 Use CEIL instead of ROUND_ADVANCE.
1533 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1536 * config/sh/sh.c: Include stdlib headers before everything else.
1538 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1540 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1541 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1542 (gimplify_adjust_omp_clauses): Simd region is never
1543 directly nested in combined parallel. Instead, for linear
1544 with copyin/copyout, if in combined for simd loop, make decl
1545 firstprivate/lastprivate on OMP_FOR.
1546 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1547 expand_omp_for_static_chunk): When setting endvar, also set
1548 fd->loop.v to the same value.
1550 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1552 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1554 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1556 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1559 2014-05-02 Marek Polacek <polacek@redhat.com>
1561 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1563 2014-05-02 Kito Cheng <kito@0xlab.org>
1565 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1566 to a C expression marco.
1567 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1568 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1569 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1570 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1571 HONOR_REG_ALLOC_ORDER.
1572 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1574 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1576 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1578 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1580 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1582 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1584 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1585 (convert_scalar_cond_reduction): Likewise.
1586 (predicate_scalar_phi): Add recognition and transformation
1587 of simple conditioanl reduction to be vectorizable.
1589 2014-05-01 Marek Polacek <polacek@redhat.com>
1592 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1594 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1596 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1597 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1598 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1599 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1600 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1601 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1602 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1603 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1605 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1607 * config/arc/arc.opt (mlra): Move comment above option name
1608 to avoid mis-parsing as language options.
1610 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1612 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1613 * config/sol2.h: ... here.
1614 * config/sol2-10.h: Remove.
1616 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1617 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1618 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1619 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1620 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1621 * config/sol2.h: ... here.
1622 (SECTION_NAME_FORMAT): Don't redefine.
1623 (STARTFILE_ARCH32_SPEC): Rename to ...
1624 (STARTFILE_ARCH_SPEC): ... this.
1625 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1626 * config/sparc/sol2.h: ... here.
1627 (SECTION_NAME_FORMAT): Don't undef.
1628 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1629 (SUBTARGET_EXTRA_SPECS): Remove.
1630 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1632 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1633 (MD_STARTFILE_PREFIX): Remove.
1634 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1635 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1636 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1637 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1638 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1639 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1640 * config/i386/sol2.h: ... here.
1641 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1642 * config/i386/sol2-bi.h: Remove.
1643 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1644 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1646 * config/i386/t-sol2-64: Rename to ...
1647 * config/i386/t-sol2: ... this.
1648 * config/sparc/t-sol2-64: Rename to ...
1649 * config/sparc/t-sol2: ... this.
1651 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1652 sol2_tm_file_head, sol2_tm_file_tail.
1653 Include ${cpu_type}/sol2.h before sol2.h.
1655 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1656 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1657 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1658 Reflect i386/t-sol2-64 renaming.
1659 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1660 Reflect sparc/t-sol2-64 renaming.
1662 2014-04-30 Richard Biener <rguenther@suse.de>
1664 * passes.c (execute_function_todo): Move TODO_verify_stmts
1665 and TODO_verify_ssa under the TODO_verify_il umbrella.
1666 * tree-ssa.h (verify_ssa): Adjust prototype.
1667 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1668 we should verify SSA operands.
1669 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1670 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1671 whether we should verify whether not throwing stmts have EH info.
1672 * graphite-scop-detection.c (create_sese_edges): Adjust.
1673 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1674 * tree-eh.c (lower_try_finally_switch): Do not add the
1675 default case label twice.
1677 2014-04-30 Marek Polacek <polacek@redhat.com>
1679 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1680 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1681 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1682 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1684 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1686 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1687 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1688 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1689 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1690 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1691 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1692 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1693 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1695 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1697 * tree-cfg.c (dump_function_to_file): Dump the return type of
1698 functions, in a line to itself before the function body, mimicking
1699 the layout of a C function.
1701 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1703 PR tree-optimization/60971
1704 * tree-tailcall.c (process_assignment): Reject conversions which
1707 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1709 * calls.c (initialize_argument_information): Always treat
1710 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1711 (expand_call): Likewise.
1712 (emit_library_call_calue_1): Likewise.
1713 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1714 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1717 2014-04-29 Nick Clifton <nickc@redhat.com>
1719 * config/msp430/msp430.md (umulsidi): Fix typo.
1720 (mulhisi3): Enable even inside interrupt handlers.
1721 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1722 bigger return address pushed in large mode.
1724 2014-04-29 Nick Clifton <nickc@redhat.com>
1726 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1727 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1729 * config/m32r/m32r.c (init_reg_tables): Likewise.
1730 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1731 enum to hold the modes.
1733 2014-04-29 Richard Biener <rguenther@suse.de>
1735 * dominance.c (free_dominance_info): Add overload with
1737 (dom_info_state): Likewise.
1738 (dom_info_available_p): Likewise.
1739 * basic-block.h (free_dominance_info, dom_info_state,
1740 dom_info_available_p): Declare overloads.
1741 * passes.c (execute_function_todo): Verify that verifiers
1742 don't change dominator info state. Drop dominator info
1743 for IPA pass invocations.
1744 * cgraph.c (release_function_body): Restore asserts that
1745 dominator information is released.
1747 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1749 * doc/invoke.texi: Fix typo.
1750 * tree-vrp.c: Fix typos.
1751 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1753 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1755 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1757 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1759 * config/aarch64/aarch64-builtins.c
1760 (aarch64_types_storestruct_lane_qualifiers): New.
1761 (TYPES_STORESTRUCT_LANE): Likewise.
1762 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1763 (st3_lane): Likewise.
1764 (st4_lane): Likewise.
1765 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1766 (vec_store_lanesci_lane<mode>): Likewise.
1767 (vec_store_lanesxi_lane<mode>): Likewise.
1768 (aarch64_st2_lane<VQ:mode>): Likewise.
1769 (aarch64_st3_lane<VQ:mode>): Likewise.
1770 (aarch64_st4_lane<VQ:mode>): Likewise.
1771 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1772 * config/aarch64/arm_neon.h
1773 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1774 use new macro arguments.
1775 (__ST3_LANE_FUNC): Likewise.
1776 (__ST4_LANE_FUNC): Likewise.
1777 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1778 (V_THREE_ELEM): Likewise.
1779 (V_FOUR_ELEM): Likewise.
1781 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1783 * doc/gimple.texi: Replace the description of the now-defunct
1784 union gimple_statement_d with a diagram showing the
1785 gimple_statement_base class hierarchy and its relationships to
1786 the GSS_ and GIMPLE_ enums.
1788 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1790 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1791 * config/aarch64/aarch64.c
1792 (aarch64_cannot_change_mode_class): Weaken conditions.
1793 (aarch64_modes_tieable_p): New.
1794 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1796 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1798 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1799 (loadsync_<mode>): Change mode.
1800 (load_quadpti, store_quadpti): New.
1801 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1802 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1804 2014-04-28 Martin Jambor <mjambor@suse.cz>
1806 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1807 same alias type as the original statement.
1808 (subreplacement_assignment_data): New type.
1809 (handle_unscalarized_data_in_subtree): New type of parameter,
1810 generate new memory accesses with same alias type as the original
1812 (load_assign_lhs_subreplacements): Likewise.
1813 (sra_modify_constructor_assign): Generate new memory accesses with
1814 same alias type as the original statement.
1816 2014-04-28 Richard Biener <rguenther@suse.de>
1818 * tree-pass.h (TODO_verify_il): Define.
1819 (TODO_verify_all): Complete properly.
1820 * passes.c (execute_function_todo): Move existing loop-closed
1821 SSA verification under TODO_verify_il.
1822 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1823 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1824 Fix tree sharing issue.
1826 2014-04-28 Richard Biener <rguenther@suse.de>
1829 * builtins.def (DEF_C11_BUILTIN): Add.
1830 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1831 * coretypes.h (enum function_class): Add function_c11_misc.
1832 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1833 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1834 (call_may_clobber_ref_p_1): Likewise.
1835 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1836 (mark_all_reaching_defs_necessary_1): Likewise.
1837 (propagate_necessity): Likewise.
1838 (eliminate_unnecessary_stmts): Likewise.
1839 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1841 2014-04-28 Richard Biener <rguenther@suse.de>
1843 * tree-vrp.c (vrp_var_may_overflow): Remove.
1844 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1845 with overflow immediately bump to one before that value and
1846 let iteration figure out overflow status.
1848 2014-04-28 Richard Biener <rguenther@suse.de>
1850 * configure.ac: Do valgrind header checks unconditionally.
1851 Add --enable-valgrind-annotations.
1852 * system.h: Guard valgrind header inclusion with
1853 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1854 * alloc-pool.c (pool_alloc, pool_free): Use
1855 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1856 to guard possibly dead code.
1857 * config.in: Regenerated.
1858 * configure: Likewise.
1860 2014-04-28 Jeff Law <law@redhat.com>
1862 PR tree-optimization/60902
1863 * tree-ssa-threadedge.c
1864 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1865 over real defs when invalidating outputs from statements that do not
1866 produce useful outputs for threading.
1868 2014-04-28 Richard Biener <rguenther@suse.de>
1870 PR tree-optimization/60979
1871 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1872 SCOPs that end in a block with a successor with abnormal
1875 2014-04-28 Richard Biener <rguenther@suse.de>
1877 * tree-pass.h (execute_pass_list): Adjust prototype.
1878 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1879 (do_per_function): Change callback signature, push all actual
1880 work to the callbals.
1881 (do_per_function_toporder): Likewise.
1882 (execute_function_dump): Adjust.
1883 (execute_function_todo): Likewise.
1884 (clear_last_verified): Likewise.
1885 (verify_curr_properties): Likewise.
1886 (update_properties_after_pass): Likewise.
1887 (execute_pass_list_1): Split out from ...
1888 (execute_pass_list): ... here. Adjust.
1889 (execute_ipa_pass_list): Likewise.
1890 * cgraphunit.c (cgraph_add_new_function): Adjust.
1891 (analyze_function): Likewise.
1892 (expand_function): Likewise.
1893 * cgraph.c (release_function_body): Free dominance info
1894 here instead of asserting it was magically freed elsewhere.
1896 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1898 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1899 * configure: Regenerate.
1900 * config/sparc/sparc.opt (muser-mode): New option.
1901 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1903 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1904 * doc/invoke.texi (SPARC options): Document -muser-mode.
1906 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1908 * cselib.c (find_slot_memmode): Delete.
1909 (cselib_hasher): Change compare_type to a struct.
1910 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1912 (preserve_constants_and_equivs): Adjust for new compare_type.
1913 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1914 (wrap_constant): Delete.
1915 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1917 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1919 * doc/install.texi (Building with profile feedback): Remove
1922 2014-04-26 Tom de Vries <tom@codesourcery.com>
1924 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1927 2014-04-25 Cary Coutant <ccoutant@google.com>
1930 * dwarf2out.c (should_move_die_to_comdat): A type definition
1931 can contain a subprogram definition, but don't move it to a
1933 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1934 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1936 (clone_tree_hash): Rename to...
1937 (clone_tree_partial): ...this; change callers. Copy
1938 DW_TAG_subprogram DIEs as declarations.
1939 (copy_decls_walk): Don't copy children of a declaration into a
1942 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1945 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1948 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1950 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1951 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1953 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1956 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1958 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1960 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1962 PR tree-optimization/60930
1963 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1964 creating a multiply candidate by folding two constant
1965 multiplicands when the result overflows.
1967 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1969 PR tree-optimization/60960
1970 * tree-vect-generic.c (expand_vector_operation): Only call
1971 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1973 2014-04-25 Tom de Vries <tom@codesourcery.com>
1975 * expr.c (clobber_reg_mode): New function.
1976 * expr.h (clobber_reg): New function.
1978 2014-04-25 Tom de Vries <tom@codesourcery.com>
1980 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1983 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1984 Tom de Vries <tom@codesourcery.com>
1986 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1988 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1989 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1990 new argument to find_all_hard_reg_sets call.
1992 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1994 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1995 Use HOST_WIDE_INT_C for mask literal.
1996 (aarch_rev16_shleft_mask_imm_p): Likewise.
1998 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2001 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2003 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2005 PR preprocessor/56540
2006 * config/i386/i386-c.c (ix86_target_macros): Define
2007 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2009 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2011 * configure.ac (tga_func): Remove.
2012 (LIB_TLS_SPEC): Remove.
2013 * configure: Regenerate.
2014 * config.in: Regenerate.
2015 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2017 2014-04-25 Richard Biener <rguenther@suse.de>
2020 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2021 call stmt use/clobber sets during stmt walk instead of
2022 walking the possibly incomplete set of caller edges.
2024 2014-04-25 Richard Biener <rguenther@suse.de>
2027 * passes.c (apply_ipa_transforms): Inline into only caller ...
2028 (execute_one_pass): ... here. Properly bring in function
2029 bodies for nodes we want to apply IPA transforms to.
2031 2014-04-24 Cong Hou <congh@google.com>
2033 PR tree-optimization/60896
2034 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2035 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2036 (vect_mark_pattern_stmts): Set the def type of all statements in
2037 PATTERN_DEF_SEQ as vect_internal_def.
2039 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2041 * doc/extend.texi (PowerPC Built-in Functions): Document new
2042 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2043 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2045 * config/rs6000/predicates.md (const_0_to_3_operand): New
2046 predicate to match 0..3 integer constants.
2048 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2049 to support adding miscellaneous builtin functions.
2050 (BU_DFP_MISC_2): Likewise.
2051 (BU_P7_MISC_1): Likewise.
2052 (BU_P7_MISC_2): Likewise.
2053 (BU_P8V_MISC_3): Likewise.
2054 (BU_MISC_1): Likewise.
2055 (BU_MISC_2): Likewise.
2056 (DIVWE): Add extended divide builtin functions.
2059 (DIVWEUO): Likewise.
2063 (DIVDEUO): Likewise.
2064 (DXEX): Add decimal floating-point builtin functions.
2067 (DDEDPDQ): Likewise.
2069 (DENBCDQ): Likewise.
2076 (CDTBCD): Add new BCD builtin functions.
2080 (BCDADD_LT): Likewise.
2081 (BCDADD_EQ): Likewise.
2082 (BCDADD_GT): Likewise.
2083 (BCDADD_OV): Likewise.
2085 (BCDSUB_LT): Likewise.
2086 (BCDSUB_EQ): Likewise.
2087 (BCDSUB_GT): Likewise.
2088 (BCDSUB_OV): Likewise.
2089 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2090 (UNPACK_TD): Likewise.
2091 (PACK_TF): Likewise.
2092 (UNPACK_TF): Likewise.
2093 (UNPACK_TF_0): Likewise.
2094 (UNPACK_TF_1): Likewise.
2095 (PACK_V1TI): Likewise.
2096 (UNPACK_V1TI): Likewise.
2098 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2099 support for decimal floating point builtin functions.
2100 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2101 functions that take constant arguments.
2102 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2103 (rs6000_init_builtins): Setup long double, _Decimal64, and
2104 _Decimal128 types for new builtin functions.
2105 (builtin_function_type): Set the unsigned flags appropriately for
2106 the new builtin functions.
2107 (rs6000_opt_masks): Add support for decimal floating point builtin
2110 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2111 floating point builtin functions.
2112 (RS6000_BTM_COMMON): Likewise.
2113 (RS6000_BTI_long_double): Likewise.
2114 (RS6000_BTI_dfloat64): Likewise.
2115 (RS6000_BTI_dfloat128): Likewise.
2116 (long_double_type_internal_node): Likewise.
2117 (dfloat64_type_internal_node): Likewise.
2118 (dfloat128_type_internal_node): Likewise.
2120 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2121 2.07 bcd arithmetic instructions.
2122 (UNSPEC_BCDSUB): Likewise.
2123 (UNSPEC_BCD_OVERFLOW): Likewise.
2124 (UNSPEC_BCD_ADD_SUB): Likewise.
2125 (bcd_add_sub): Likewise.
2126 (BCD_TEST): Likewise.
2127 (bcd<bcd_add_sub>): Likewise.
2128 (bcd<bcd_add_sub>_test): Likewise.
2129 (bcd<bcd_add_sub>_test2): Likewise.
2130 (bcd<bcd_add_sub>_<code>): Likewise.
2131 (peephole2 for combined bcd ops): Likewise.
2133 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2134 decimal floating point builtin functions.
2135 (UNSPEC_DENBCD): Likewise.
2136 (UNSPEC_DXEX): Likewise.
2137 (UNSPEC_DIEX): Likewise.
2138 (UNSPEC_DSCLI): Likewise.
2139 (UNSPEC_DSCRI): Likewise.
2140 (D64_D128): Likewise.
2141 (dfp_suffix): Likewise.
2142 (dfp_ddedpd_<mode>): Likewise.
2143 (dfp_denbcd_<mode>): Likewise.
2144 (dfp_dxex_<mode>): Likewise.
2145 (dfp_diex_<mode>): Likewise.
2146 (dfp_dscli_<mode>): Likewise.
2147 (dfp_dscri_<mode>): Likewise.
2149 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2151 (UNSPEC_CDTBCD): Likewise.
2152 (UNSPEC_CBCDTD): Likewise.
2153 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2154 (UNSPEC_DIVEO): Likewise.
2155 (UNSPEC_DIVEU): Likewise.
2156 (UNSPEC_DIVEUO): Likewise.
2157 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2158 pack/unpack 128-bit types.
2159 (UNSPEC_PACK_128BIT): Likewise.
2160 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2161 (udiv<mode>3): Use idiv_ldiv mode attribute.
2162 (div<mode>3): Likewise.
2163 (addg6s): Add new BCD builtin functions.
2166 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2167 (div_extend): Likewise.
2168 (div<div_extend>_<mode>"): Likewise.
2169 (FP128_64): Add support for new builtin functions to pack/unpack
2171 (unpack<mode>): Likewise.
2172 (unpacktf_0): Likewise.
2173 (unpacktf_1): Likewise.
2174 (unpack<mode>_dm): Likewise.
2175 (unpack<mode>_nodm): Likewise.
2176 (pack<mode>): Likewise.
2177 (unpackv1ti): Likewise.
2178 (packv1ti): Likewise.
2180 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2182 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2185 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2187 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2188 * gimplify.c (omp_is_private): Change last argument's type to int.
2189 Only diagnose lastprivate if the simd argument is 1, only diagnose
2190 linear if the simd argument is 2.
2191 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2192 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2193 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2194 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2195 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2196 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2197 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2198 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2199 * tree-nested.c (convert_nonlocal_omp_clauses,
2200 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2202 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2205 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2208 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2210 * flag-types.h (enum ivar_visibility): Add.
2212 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2214 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2215 function * argument.
2217 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2219 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2221 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2222 Tom de Vries <tom@codesourcery.com>
2224 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2225 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2227 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2228 * emit-rtl.c (try_split): Same.
2230 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2231 Tom de Vries <tom@codesourcery.com>
2233 * common.opt (fuse-caller-save): New option.
2235 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2237 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2238 elements for big-endian.
2240 2014-04-24 Richard Biener <rguenther@suse.de>
2242 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2243 during TER and instead use the sepops interface for expanding
2244 non-GIMPLE_SINGLE_RHS.
2246 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2248 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2249 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2251 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2253 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2254 assembler 64-bit option.
2255 * configure: Regenerate.
2257 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2259 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2260 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2261 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2262 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2263 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2265 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2267 * config/aarch64/aarch64-builtins.c
2268 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2269 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2270 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2271 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2273 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2274 (Vrevsuff): New mode attribute.
2276 2014-04-24 Terry Guo <terry.guo@arm.com>
2278 * config/arm/arm.h (machine_function): Define variable
2279 after_arm_reorg here.
2280 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2281 (arm_split_constant): Update the way to access variable
2284 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2286 2014-04-23 Tom de Vries <tom@codesourcery.com>
2288 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2290 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2292 * is-a.h: Update comments to reflect the following changes to the
2293 "pointerness" of the API, making the template parameter match the
2294 return type, allowing use of is-a.h with typedefs of pointers.
2295 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2296 that the return type matches the parameter to the is_a_helper.
2298 (dyn_cast): Likewise.
2300 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2301 pointer from the is-a.h API.
2303 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2304 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2306 (is_a_helper <varpool_node>::test): Likewise, convert to...
2307 (is_a_helper <varpool_node *>::test): ...this.
2309 (varpool_first_variable): Update for removal of implicit pointer
2310 from the is-a.h API.
2311 (varpool_next_variable): Likewise.
2312 (varpool_first_static_initializer): Likewise.
2313 (varpool_next_static_initializer): Likewise.
2314 (varpool_first_defined_variable): Likewise.
2315 (varpool_next_defined_variable): Likewise.
2316 (cgraph_first_defined_function): Likewise.
2317 (cgraph_next_defined_function): Likewise.
2318 (cgraph_first_function): Likewise.
2319 (cgraph_next_function): Likewise.
2320 (cgraph_first_function_with_gimple_body): Likewise.
2321 (cgraph_next_function_with_gimple_body): Likewise.
2322 (cgraph_alias_target): Likewise.
2323 (varpool_alias_target): Likewise.
2324 (cgraph_function_or_thunk_node): Likewise.
2325 (varpool_variable_node): Likewise.
2326 (symtab_real_symbol_p): Likewise.
2327 * cgraphunit.c (referred_to_p): Likewise.
2328 (analyze_functions): Likewise.
2329 (handle_alias_pairs): Likewise.
2330 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2331 * gimple-ssa.h (gimple_vuse_op): Likewise.
2332 (gimple_vdef_op): Likewise.
2333 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2334 * gimple.c (gimple_build_asm_1): Likewise.
2335 (gimple_build_try): Likewise.
2336 (gimple_build_resx): Likewise.
2337 (gimple_build_eh_dispatch): Likewise.
2338 (gimple_build_omp_for): Likewise.
2339 (gimple_omp_for_set_clauses): Likewise.
2341 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2342 (is_a_helper <gimple_statement_asm *>::test): ...this.
2343 (is_a_helper <gimple_statement_bind>::test): Convert to...
2344 (is_a_helper <gimple_statement_bind *>::test): ...this.
2345 (is_a_helper <gimple_statement_call>::test): Convert to...
2346 (is_a_helper <gimple_statement_call *>::test): ...this.
2347 (is_a_helper <gimple_statement_catch>::test): Convert to...
2348 (is_a_helper <gimple_statement_catch *>::test): ...this.
2349 (is_a_helper <gimple_statement_resx>::test): Convert to...
2350 (is_a_helper <gimple_statement_resx *>::test): ...this.
2351 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2352 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2353 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2354 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2355 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2356 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2357 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2358 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2359 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2360 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2361 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2362 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2363 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2364 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2365 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2366 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2367 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2368 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2369 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2370 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2371 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2372 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2373 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2374 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2375 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2376 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2377 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2378 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2379 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2380 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2381 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2382 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2383 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2384 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2385 (is_a_helper <gimple_statement_phi>::test): Convert to...
2386 (is_a_helper <gimple_statement_phi *>::test): ...this.
2387 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2388 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2389 (is_a_helper <gimple_statement_try>::test): Convert to...
2390 (is_a_helper <gimple_statement_try *>::test): ...this.
2391 (is_a_helper <gimple_statement_wce>::test): Convert to...
2392 (is_a_helper <gimple_statement_wce *>::test): ...this.
2393 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2394 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2395 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2396 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2397 (is_a_helper <const gimple_statement_call>::test): Convert to...
2398 (is_a_helper <const gimple_statement_call *>::test): ...this.
2399 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2400 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2401 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2402 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2403 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2404 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2405 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2406 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2407 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2409 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2411 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2413 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2415 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2416 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2417 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2419 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2420 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2422 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2423 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2424 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2425 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2426 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2427 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2429 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2430 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2431 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2432 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2434 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2435 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2436 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2437 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2438 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2439 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2440 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2441 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2442 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2443 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2444 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2445 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2446 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2447 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2448 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2449 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2451 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2453 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2454 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2456 (gimple_use_ops): Update for removal of implicit pointer from the
2458 (gimple_set_use_ops): Likewise.
2459 (gimple_vuse): Likewise.
2460 (gimple_vdef): Likewise.
2461 (gimple_vuse_ptr): Likewise.
2462 (gimple_vdef_ptr): Likewise.
2463 (gimple_set_vuse): Likewise.
2464 (gimple_set_vdef): Likewise.
2465 (gimple_omp_return_set_lhs): Likewise.
2466 (gimple_omp_return_lhs): Likewise.
2467 (gimple_omp_return_lhs_ptr): Likewise.
2468 (gimple_call_fntype): Likewise.
2469 (gimple_call_set_fntype): Likewise.
2470 (gimple_call_set_internal_fn): Likewise.
2471 (gimple_call_use_set): Likewise.
2472 (gimple_call_clobber_set): Likewise.
2473 (gimple_bind_vars): Likewise.
2474 (gimple_bind_set_vars): Likewise.
2475 (gimple_bind_body_ptr): Likewise.
2476 (gimple_bind_set_body): Likewise.
2477 (gimple_bind_add_stmt): Likewise.
2478 (gimple_bind_block): Likewise.
2479 (gimple_bind_set_block): Likewise.
2480 (gimple_asm_ninputs): Likewise.
2481 (gimple_asm_noutputs): Likewise.
2482 (gimple_asm_nclobbers): Likewise.
2483 (gimple_asm_nlabels): Likewise.
2484 (gimple_asm_input_op): Likewise.
2485 (gimple_asm_input_op_ptr): Likewise.
2486 (gimple_asm_output_op): Likewise.
2487 (gimple_asm_output_op_ptr): Likewise.
2488 (gimple_asm_set_output_op): Likewise.
2489 (gimple_asm_clobber_op): Likewise.
2490 (gimple_asm_set_clobber_op): Likewise.
2491 (gimple_asm_label_op): Likewise.
2492 (gimple_asm_set_label_op): Likewise.
2493 (gimple_asm_string): Likewise.
2494 (gimple_catch_types): Likewise.
2495 (gimple_catch_types_ptr): Likewise.
2496 (gimple_catch_handler_ptr): Likewise.
2497 (gimple_catch_set_types): Likewise.
2498 (gimple_catch_set_handler): Likewise.
2499 (gimple_eh_filter_types): Likewise.
2500 (gimple_eh_filter_types_ptr): Likewise.
2501 (gimple_eh_filter_failure_ptr): Likewise.
2502 (gimple_eh_filter_set_types): Likewise.
2503 (gimple_eh_filter_set_failure): Likewise.
2504 (gimple_eh_must_not_throw_fndecl): Likewise.
2505 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2506 (gimple_eh_else_n_body_ptr): Likewise.
2507 (gimple_eh_else_e_body_ptr): Likewise.
2508 (gimple_eh_else_set_n_body): Likewise.
2509 (gimple_eh_else_set_e_body): Likewise.
2510 (gimple_try_eval_ptr): Likewise.
2511 (gimple_try_cleanup_ptr): Likewise.
2512 (gimple_try_set_eval): Likewise.
2513 (gimple_try_set_cleanup): Likewise.
2514 (gimple_wce_cleanup_ptr): Likewise.
2515 (gimple_wce_set_cleanup): Likewise.
2516 (gimple_phi_capacity): Likewise.
2517 (gimple_phi_num_args): Likewise.
2518 (gimple_phi_result): Likewise.
2519 (gimple_phi_result_ptr): Likewise.
2520 (gimple_phi_set_result): Likewise.
2521 (gimple_phi_arg): Likewise.
2522 (gimple_phi_set_arg): Likewise.
2523 (gimple_resx_region): Likewise.
2524 (gimple_resx_set_region): Likewise.
2525 (gimple_eh_dispatch_region): Likewise.
2526 (gimple_eh_dispatch_set_region): Likewise.
2527 (gimple_omp_critical_name): Likewise.
2528 (gimple_omp_critical_name_ptr): Likewise.
2529 (gimple_omp_critical_set_name): Likewise.
2530 (gimple_omp_for_clauses): Likewise.
2531 (gimple_omp_for_clauses_ptr): Likewise.
2532 (gimple_omp_for_set_clauses): Likewise.
2533 (gimple_omp_for_collapse): Likewise.
2534 (gimple_omp_for_index): Likewise.
2535 (gimple_omp_for_index_ptr): Likewise.
2536 (gimple_omp_for_set_index): Likewise.
2537 (gimple_omp_for_initial): Likewise.
2538 (gimple_omp_for_initial_ptr): Likewise.
2539 (gimple_omp_for_set_initial): Likewise.
2540 (gimple_omp_for_final): Likewise.
2541 (gimple_omp_for_final_ptr): Likewise.
2542 (gimple_omp_for_set_final): Likewise.
2543 (gimple_omp_for_incr): Likewise.
2544 (gimple_omp_for_incr_ptr): Likewise.
2545 (gimple_omp_for_set_incr): Likewise.
2546 (gimple_omp_for_pre_body_ptr): Likewise.
2547 (gimple_omp_for_set_pre_body): Likewise.
2548 (gimple_omp_parallel_clauses): Likewise.
2549 (gimple_omp_parallel_clauses_ptr): Likewise.
2550 (gimple_omp_parallel_set_clauses): Likewise.
2551 (gimple_omp_parallel_child_fn): Likewise.
2552 (gimple_omp_parallel_child_fn_ptr): Likewise.
2553 (gimple_omp_parallel_set_child_fn): Likewise.
2554 (gimple_omp_parallel_data_arg): Likewise.
2555 (gimple_omp_parallel_data_arg_ptr): Likewise.
2556 (gimple_omp_parallel_set_data_arg): Likewise.
2557 (gimple_omp_task_clauses): Likewise.
2558 (gimple_omp_task_clauses_ptr): Likewise.
2559 (gimple_omp_task_set_clauses): Likewise.
2560 (gimple_omp_task_child_fn): Likewise.
2561 (gimple_omp_task_child_fn_ptr): Likewise.
2562 (gimple_omp_task_set_child_fn): Likewise.
2563 (gimple_omp_task_data_arg): Likewise.
2564 (gimple_omp_task_data_arg_ptr): Likewise.
2565 (gimple_omp_task_set_data_arg): Likewise.
2566 (gimple_omp_taskreg_clauses): Likewise.
2567 (gimple_omp_taskreg_clauses_ptr): Likewise.
2568 (gimple_omp_taskreg_set_clauses): Likewise.
2569 (gimple_omp_taskreg_child_fn): Likewise.
2570 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2571 (gimple_omp_taskreg_set_child_fn): Likewise.
2572 (gimple_omp_taskreg_data_arg): Likewise.
2573 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2574 (gimple_omp_taskreg_set_data_arg): Likewise.
2575 (gimple_omp_task_copy_fn): Likewise.
2576 (gimple_omp_task_copy_fn_ptr): Likewise.
2577 (gimple_omp_task_set_copy_fn): Likewise.
2578 (gimple_omp_task_arg_size): Likewise.
2579 (gimple_omp_task_arg_size_ptr): Likewise.
2580 (gimple_omp_task_set_arg_size): Likewise.
2581 (gimple_omp_task_arg_align): Likewise.
2582 (gimple_omp_task_arg_align_ptr): Likewise.
2583 (gimple_omp_task_set_arg_align): Likewise.
2584 (gimple_omp_single_clauses): Likewise.
2585 (gimple_omp_single_clauses_ptr): Likewise.
2586 (gimple_omp_single_set_clauses): Likewise.
2587 (gimple_omp_target_clauses): Likewise.
2588 (gimple_omp_target_clauses_ptr): Likewise.
2589 (gimple_omp_target_set_clauses): Likewise.
2590 (gimple_omp_target_child_fn): Likewise.
2591 (gimple_omp_target_child_fn_ptr): Likewise.
2592 (gimple_omp_target_set_child_fn): Likewise.
2593 (gimple_omp_target_data_arg): Likewise.
2594 (gimple_omp_target_data_arg_ptr): Likewise.
2595 (gimple_omp_target_set_data_arg): Likewise.
2596 (gimple_omp_teams_clauses): Likewise.
2597 (gimple_omp_teams_clauses_ptr): Likewise.
2598 (gimple_omp_teams_set_clauses): Likewise.
2599 (gimple_omp_sections_clauses): Likewise.
2600 (gimple_omp_sections_clauses_ptr): Likewise.
2601 (gimple_omp_sections_set_clauses): Likewise.
2602 (gimple_omp_sections_control): Likewise.
2603 (gimple_omp_sections_control_ptr): Likewise.
2604 (gimple_omp_sections_set_control): Likewise.
2605 (gimple_omp_for_set_cond): Likewise.
2606 (gimple_omp_for_cond): Likewise.
2607 (gimple_omp_atomic_store_set_val): Likewise.
2608 (gimple_omp_atomic_store_val): Likewise.
2609 (gimple_omp_atomic_store_val_ptr): Likewise.
2610 (gimple_omp_atomic_load_set_lhs): Likewise.
2611 (gimple_omp_atomic_load_lhs): Likewise.
2612 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2613 (gimple_omp_atomic_load_set_rhs): Likewise.
2614 (gimple_omp_atomic_load_rhs): Likewise.
2615 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2616 (gimple_omp_continue_control_def): Likewise.
2617 (gimple_omp_continue_control_def_ptr): Likewise.
2618 (gimple_omp_continue_set_control_def): Likewise.
2619 (gimple_omp_continue_control_use): Likewise.
2620 (gimple_omp_continue_control_use_ptr): Likewise.
2621 (gimple_omp_continue_set_control_use): Likewise.
2622 (gimple_transaction_body_ptr): Likewise.
2623 (gimple_transaction_label): Likewise.
2624 (gimple_transaction_label_ptr): Likewise.
2625 (gimple_transaction_set_body): Likewise.
2626 (gimple_transaction_set_label): Likewise.
2628 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2629 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2630 * ipa-ref.c (ipa_record_reference): Likewise.
2631 * ipa-reference.c (analyze_function): Likewise.
2632 (ipa_reference_write_optimization_summary): Likewise.
2633 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2634 (address_taken_from_non_vtable_p): Likewise.
2635 (comdat_can_be_unshared_p_1): Likewise.
2636 * lto-cgraph.c (lto_output_ref): Likewise.
2637 (add_references): Likewise.
2638 (compute_ltrans_boundary): Likewise.
2639 (output_symtab): Likewise.
2640 (input_ref): Likewise.
2641 (input_cgraph_1): Likewise.
2642 (output_cgraph_opt_summary): Likewise.
2643 * lto-streamer-out.c (lto_output): Likewise.
2644 (output_symbol_p): Likewise.
2645 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2646 (lsei_start_function_in_partition): Likewise.
2647 (lsei_next_variable_in_partition): Likewise.
2648 (lsei_start_variable_in_partition): Likewise.
2649 * symtab.c (insert_to_assembler_name_hash): Likewise.
2650 (unlink_from_assembler_name_hash): Likewise.
2651 (symtab_unregister_node): Likewise.
2652 (symtab_remove_node): Likewise.
2653 (dump_symtab_node): Likewise.
2654 (verify_symtab_base): Likewise.
2655 (verify_symtab_node): Likewise.
2656 (symtab_make_decl_local): Likewise.
2657 (symtab_alias_ultimate_target): Likewise.
2658 (symtab_resolve_alias): Likewise.
2659 (symtab_get_symbol_partitioning_class): Likewise.
2660 * tree-phinodes.c (allocate_phi_node): Likewise.
2661 (reserve_phi_args_for_new_edge): Likewise.
2662 (remove_phi_args): Likewise.
2663 * varpool.c (varpool_node_for_asm): Likewise.
2664 (varpool_remove_unreferenced_decls): Likewise.
2666 2014-04-23 Jeff Law <law@redhat.com>
2668 PR tree-optimization/60902
2669 * tree-ssa-threadedge.c
2670 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2671 invalidate outputs from statements that do not produce useful
2672 outputs for threading.
2674 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2676 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2677 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2678 machine descriptions for Stack Smashing Protector.
2680 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2682 * aarch64.md (<optab>_rol<mode>3): New pattern.
2683 (<optab>_rolsi3_uxtw): Likewise.
2684 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2686 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2688 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2689 (arm_cortex_a12_tune): Likewise.
2691 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2693 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2695 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2697 * config/arm/arm.md (arm_rev16si2): New pattern.
2698 (arm_rev16si2_alt): Likewise.
2699 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2701 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2703 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2704 (rev16<mode>2_alt): Likewise.
2705 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2706 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2707 (aarch_rev16_shleft_mask_imm_p): Likewise.
2708 (aarch_rev16_p_1): Likewise.
2709 (aarch_rev16_p): Likewise.
2710 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2711 (aarch_rev16_shright_mask_imm_p): Likewise.
2712 (aarch_rev16_shleft_mask_imm_p): Likewise.
2714 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2716 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2717 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2719 (cortex_a53_extra_costs): Likewise.
2720 (cortex_a57_extra_costs): Likewise.
2721 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2722 (cortexa7_extra_costs): Likewise.
2723 (cortexa8_extra_costs): Likewise.
2724 (cortexa12_extra_costs): Likewise.
2725 (cortexa15_extra_costs): Likewise.
2726 (v7m_extra_costs): Likewise.
2727 (arm_new_rtx_costs): Handle BSWAP.
2729 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2731 * config/arm/arm.c (cortexa8_extra_costs): New table.
2732 (arm_cortex_a8_tune): New tuning struct.
2733 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2735 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2737 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2739 2014-04-23 Richard Biener <rguenther@suse.de>
2741 * Makefile.in (OBJS): Remove loop-unswitch.o.
2742 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2743 * passes.def (pass_rtl_unswitch): Likewise.
2744 * loop-init.c (gate_rtl_unswitch): Likewise.
2745 (rtl_unswitch): Likewise.
2746 (pass_data_rtl_unswitch): Likewise.
2747 (pass_rtl_unswitch): Likewise.
2748 (make_pass_rtl_unswitch): Likewise.
2749 * rtl.h (reversed_condition): Likewise.
2750 (compare_and_jump_seq): Likewise.
2751 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2753 * loop-unroll.c (compare_and_jump_seq): Likewise.
2755 2014-04-23 Richard Biener <rguenther@suse.de>
2757 PR tree-optimization/60903
2758 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2759 commented code block.
2760 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2761 loop flags to newly created BBs and edges.
2763 2014-04-23 Nick Clifton <nickc@redhat.com>
2765 * config/msp430/msp430.c (msp430_handle_option): Move function
2767 (msp430_option_override): Simplify mcu and mcpu option handling.
2768 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2769 support for -mhwmult command line option.
2770 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2771 -mhwmult command line option.
2772 (msp430_hwmult_enabled): Delete.
2773 (msp43o_output_labelref): Add support for -mhwmult command line option.
2774 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2775 (umulsidi3): Likewise.
2776 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2777 (mcpu, mlarge, msmall): Likewise.
2778 (mhwmult): New option.
2779 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2781 (msp430_is_f5_mcu): Remove prototype.
2782 (msp430_use_f5_series_hwmult): Add prototype.
2783 * config/msp430/msp430-opts.h: New file.
2784 * common/config/msp430: New directory.
2785 * common/config/msp430/msp430-common.c: New file.
2786 * config.gcc (msp430): Remove target_has_targetm_common.
2787 * doc/invoke.texi: Document -mhwmult command line option.
2789 2014-04-23 Nick Clifton <nickc@redhat.com>
2791 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2792 default-manifest.o if it can be found in the search path.
2793 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2795 2014-04-23 Terry Guo <terry.guo@arm.com>
2797 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2799 2014-04-23 Richard Biener <rguenther@suse.de>
2802 * tree-inline.c (declare_return_variable): Use mark_addressable.
2804 2014-04-23 Richard Biener <rguenther@suse.de>
2807 * loop-init.c (loop_optimizer_init): Make sure to apply
2808 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2810 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2813 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2815 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2816 if flag_sanitize_undefined_trap_on_error.
2817 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2818 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2819 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2820 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2821 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2822 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2823 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2824 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2825 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2826 * ubsan.c (ubsan_instrument_unreachable): Return
2827 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2828 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2829 if flag_sanitize_undefined_trap_on_error and
2830 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2831 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2832 instrument_bool_enum_load): Emit __builtin_trap () if
2833 flag_sanitize_undefined_trap_on_error and
2834 __builtin_handle_*_abort () if !flag_sanitize_recover.
2835 * doc/invoke.texi (-fsanitize-recover,
2836 -fsanitize-undefined-trap-on-error): Document.
2838 2014-04-22 Christian Bruel <christian.bruel@st.com>
2840 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2841 Force immediates to SImode.
2843 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2845 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2847 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2848 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2849 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2850 (nios2_fpu_insn): Add entry for round.
2851 (N2FPU_NO_ERRNO_P): Define.
2852 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2854 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2856 2014-04-22 Richard Henderson <rth@redhat.com>
2858 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2859 (add<GPI>3_compare0): Remove leading * from name.
2860 (add<GPI>3_carryin): Likewise.
2861 (sub<GPI>3_compare0): Likewise.
2862 (sub<GPI>3_carryin): Likewise.
2863 (<su_optab>mulditi3): New expander.
2864 (multi3): New expander.
2865 (madd<GPI>): Remove leading * from name.
2867 2014-04-22 Martin Jambor <mjambor@suse.cz>
2869 * cgraphclones.c (cgraph_function_versioning): Copy
2870 ipa_transforms_to_apply instead of asserting it is empty.
2872 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2875 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2876 on count_exp to get mode.
2878 2014-04-22 Andrew Pinski <apinski@cavium.com>
2880 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2881 Handle TLS for ILP32.
2882 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2883 (tlsie_small_<mode>): this and handle PTR.
2884 (tlsie_small_sidi): New pattern.
2885 (tlsle_small): Change to an expand to handle ILP32.
2886 (tlsle_small_<mode>): New pattern.
2887 (tlsdesc_small): Rename to ...
2888 (tlsdesc_small_<mode>): this and handle PTR.
2890 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2892 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2894 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2896 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2897 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2898 (aarch64_types_signed_poly_qualifiers): Likewise.
2899 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2900 (aarch64_types_poly_signed_qualifiers): Likewise.
2901 (TYPES_REINTERP_SS): Type macro added.
2902 (TYPES_REINTERP_SU): Likewise.
2903 (TYPES_REINTERP_SP): Likewise.
2904 (TYPES_REINTERP_US): Likewise.
2905 (TYPES_REINTERP_PS): Likewise.
2906 (aarch64_fold_builtin): New expression folding added.
2907 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2908 Declarations removed.
2909 (REINTERP_SS): Declarations added.
2910 (REINTERP_US): Likewise.
2911 (REINTERP_PS): Likewise.
2912 (REINTERP_SU): Likewise.
2913 (REINTERP_SP): Likewise.
2914 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2915 (vreinterpretq_p8_f64): Likewise.
2916 (vreinterpret_p16_f64): Likewise.
2917 (vreinterpretq_p16_f64): Likewise.
2918 (vreinterpret_f32_f64): Likewise.
2919 (vreinterpretq_f32_f64): Likewise.
2920 (vreinterpret_f64_f32): Likewise.
2921 (vreinterpret_f64_p8): Likewise.
2922 (vreinterpret_f64_p16): Likewise.
2923 (vreinterpret_f64_s8): Likewise.
2924 (vreinterpret_f64_s16): Likewise.
2925 (vreinterpret_f64_s32): Likewise.
2926 (vreinterpret_f64_s64): Likewise.
2927 (vreinterpret_f64_u8): Likewise.
2928 (vreinterpret_f64_u16): Likewise.
2929 (vreinterpret_f64_u32): Likewise.
2930 (vreinterpret_f64_u64): Likewise.
2931 (vreinterpretq_f64_f32): Likewise.
2932 (vreinterpretq_f64_p8): Likewise.
2933 (vreinterpretq_f64_p16): Likewise.
2934 (vreinterpretq_f64_s8): Likewise.
2935 (vreinterpretq_f64_s16): Likewise.
2936 (vreinterpretq_f64_s32): Likewise.
2937 (vreinterpretq_f64_s64): Likewise.
2938 (vreinterpretq_f64_u8): Likewise.
2939 (vreinterpretq_f64_u16): Likewise.
2940 (vreinterpretq_f64_u32): Likewise.
2941 (vreinterpretq_f64_u64): Likewise.
2942 (vreinterpret_s64_f64): Likewise.
2943 (vreinterpretq_s64_f64): Likewise.
2944 (vreinterpret_u64_f64): Likewise.
2945 (vreinterpretq_u64_f64): Likewise.
2946 (vreinterpret_s8_f64): Likewise.
2947 (vreinterpretq_s8_f64): Likewise.
2948 (vreinterpret_s16_f64): Likewise.
2949 (vreinterpretq_s16_f64): Likewise.
2950 (vreinterpret_s32_f64): Likewise.
2951 (vreinterpretq_s32_f64): Likewise.
2952 (vreinterpret_u8_f64): Likewise.
2953 (vreinterpretq_u8_f64): Likewise.
2954 (vreinterpret_u16_f64): Likewise.
2955 (vreinterpretq_u16_f64): Likewise.
2956 (vreinterpret_u32_f64): Likewise.
2957 (vreinterpretq_u32_f64): Likewise.
2959 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2961 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2962 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2963 (vreinterpret_p8_s8): Likewise.
2964 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2965 (vreinterpret_p8_s16): Likewise.
2966 (vreinterpret_p8_s32): Likewise.
2967 (vreinterpret_p8_s64): Likewise.
2968 (vreinterpret_p8_f32): Likewise.
2969 (vreinterpret_p8_u8): Likewise.
2970 (vreinterpret_p8_u16): Likewise.
2971 (vreinterpret_p8_u32): Likewise.
2972 (vreinterpret_p8_u64): Likewise.
2973 (vreinterpret_p8_p16): Likewise.
2974 (vreinterpretq_p8_s8): Likewise.
2975 (vreinterpretq_p8_s16): Likewise.
2976 (vreinterpretq_p8_s32): Likewise.
2977 (vreinterpretq_p8_s64): Likewise.
2978 (vreinterpretq_p8_f32): Likewise.
2979 (vreinterpretq_p8_u8): Likewise.
2980 (vreinterpretq_p8_u16): Likewise.
2981 (vreinterpretq_p8_u32): Likewise.
2982 (vreinterpretq_p8_u64): Likewise.
2983 (vreinterpretq_p8_p16): Likewise.
2984 (vreinterpret_p16_s8): Likewise.
2985 (vreinterpret_p16_s16): Likewise.
2986 (vreinterpret_p16_s32): Likewise.
2987 (vreinterpret_p16_s64): Likewise.
2988 (vreinterpret_p16_f32): Likewise.
2989 (vreinterpret_p16_u8): Likewise.
2990 (vreinterpret_p16_u16): Likewise.
2991 (vreinterpret_p16_u32): Likewise.
2992 (vreinterpret_p16_u64): Likewise.
2993 (vreinterpret_p16_p8): Likewise.
2994 (vreinterpretq_p16_s8): Likewise.
2995 (vreinterpretq_p16_s16): Likewise.
2996 (vreinterpretq_p16_s32): Likewise.
2997 (vreinterpretq_p16_s64): Likewise.
2998 (vreinterpretq_p16_f32): Likewise.
2999 (vreinterpretq_p16_u8): Likewise.
3000 (vreinterpretq_p16_u16): Likewise.
3001 (vreinterpretq_p16_u32): Likewise.
3002 (vreinterpretq_p16_u64): Likewise.
3003 (vreinterpretq_p16_p8): Likewise.
3004 (vreinterpret_f32_s8): Likewise.
3005 (vreinterpret_f32_s16): Likewise.
3006 (vreinterpret_f32_s32): Likewise.
3007 (vreinterpret_f32_s64): Likewise.
3008 (vreinterpret_f32_u8): Likewise.
3009 (vreinterpret_f32_u16): Likewise.
3010 (vreinterpret_f32_u32): Likewise.
3011 (vreinterpret_f32_u64): Likewise.
3012 (vreinterpret_f32_p8): Likewise.
3013 (vreinterpret_f32_p16): Likewise.
3014 (vreinterpretq_f32_s8): Likewise.
3015 (vreinterpretq_f32_s16): Likewise.
3016 (vreinterpretq_f32_s32): Likewise.
3017 (vreinterpretq_f32_s64): Likewise.
3018 (vreinterpretq_f32_u8): Likewise.
3019 (vreinterpretq_f32_u16): Likewise.
3020 (vreinterpretq_f32_u32): Likewise.
3021 (vreinterpretq_f32_u64): Likewise.
3022 (vreinterpretq_f32_p8): Likewise.
3023 (vreinterpretq_f32_p16): Likewise.
3024 (vreinterpret_s64_s8): Likewise.
3025 (vreinterpret_s64_s16): Likewise.
3026 (vreinterpret_s64_s32): Likewise.
3027 (vreinterpret_s64_f32): Likewise.
3028 (vreinterpret_s64_u8): Likewise.
3029 (vreinterpret_s64_u16): Likewise.
3030 (vreinterpret_s64_u32): Likewise.
3031 (vreinterpret_s64_u64): Likewise.
3032 (vreinterpret_s64_p8): Likewise.
3033 (vreinterpret_s64_p16): Likewise.
3034 (vreinterpretq_s64_s8): Likewise.
3035 (vreinterpretq_s64_s16): Likewise.
3036 (vreinterpretq_s64_s32): Likewise.
3037 (vreinterpretq_s64_f32): Likewise.
3038 (vreinterpretq_s64_u8): Likewise.
3039 (vreinterpretq_s64_u16): Likewise.
3040 (vreinterpretq_s64_u32): Likewise.
3041 (vreinterpretq_s64_u64): Likewise.
3042 (vreinterpretq_s64_p8): Likewise.
3043 (vreinterpretq_s64_p16): Likewise.
3044 (vreinterpret_u64_s8): Likewise.
3045 (vreinterpret_u64_s16): Likewise.
3046 (vreinterpret_u64_s32): Likewise.
3047 (vreinterpret_u64_s64): Likewise.
3048 (vreinterpret_u64_f32): Likewise.
3049 (vreinterpret_u64_u8): Likewise.
3050 (vreinterpret_u64_u16): Likewise.
3051 (vreinterpret_u64_u32): Likewise.
3052 (vreinterpret_u64_p8): Likewise.
3053 (vreinterpret_u64_p16): Likewise.
3054 (vreinterpretq_u64_s8): Likewise.
3055 (vreinterpretq_u64_s16): Likewise.
3056 (vreinterpretq_u64_s32): Likewise.
3057 (vreinterpretq_u64_s64): Likewise.
3058 (vreinterpretq_u64_f32): Likewise.
3059 (vreinterpretq_u64_u8): Likewise.
3060 (vreinterpretq_u64_u16): Likewise.
3061 (vreinterpretq_u64_u32): Likewise.
3062 (vreinterpretq_u64_p8): Likewise.
3063 (vreinterpretq_u64_p16): Likewise.
3064 (vreinterpret_s8_s16): Likewise.
3065 (vreinterpret_s8_s32): Likewise.
3066 (vreinterpret_s8_s64): Likewise.
3067 (vreinterpret_s8_f32): Likewise.
3068 (vreinterpret_s8_u8): Likewise.
3069 (vreinterpret_s8_u16): Likewise.
3070 (vreinterpret_s8_u32): Likewise.
3071 (vreinterpret_s8_u64): Likewise.
3072 (vreinterpret_s8_p8): Likewise.
3073 (vreinterpret_s8_p16): Likewise.
3074 (vreinterpretq_s8_s16): Likewise.
3075 (vreinterpretq_s8_s32): Likewise.
3076 (vreinterpretq_s8_s64): Likewise.
3077 (vreinterpretq_s8_f32): Likewise.
3078 (vreinterpretq_s8_u8): Likewise.
3079 (vreinterpretq_s8_u16): Likewise.
3080 (vreinterpretq_s8_u32): Likewise.
3081 (vreinterpretq_s8_u64): Likewise.
3082 (vreinterpretq_s8_p8): Likewise.
3083 (vreinterpretq_s8_p16): Likewise.
3084 (vreinterpret_s16_s8): Likewise.
3085 (vreinterpret_s16_s32): Likewise.
3086 (vreinterpret_s16_s64): Likewise.
3087 (vreinterpret_s16_f32): Likewise.
3088 (vreinterpret_s16_u8): Likewise.
3089 (vreinterpret_s16_u16): Likewise.
3090 (vreinterpret_s16_u32): Likewise.
3091 (vreinterpret_s16_u64): Likewise.
3092 (vreinterpret_s16_p8): Likewise.
3093 (vreinterpret_s16_p16): Likewise.
3094 (vreinterpretq_s16_s8): Likewise.
3095 (vreinterpretq_s16_s32): Likewise.
3096 (vreinterpretq_s16_s64): Likewise.
3097 (vreinterpretq_s16_f32): Likewise.
3098 (vreinterpretq_s16_u8): Likewise.
3099 (vreinterpretq_s16_u16): Likewise.
3100 (vreinterpretq_s16_u32): Likewise.
3101 (vreinterpretq_s16_u64): Likewise.
3102 (vreinterpretq_s16_p8): Likewise.
3103 (vreinterpretq_s16_p16): Likewise.
3104 (vreinterpret_s32_s8): Likewise.
3105 (vreinterpret_s32_s16): Likewise.
3106 (vreinterpret_s32_s64): Likewise.
3107 (vreinterpret_s32_f32): Likewise.
3108 (vreinterpret_s32_u8): Likewise.
3109 (vreinterpret_s32_u16): Likewise.
3110 (vreinterpret_s32_u32): Likewise.
3111 (vreinterpret_s32_u64): Likewise.
3112 (vreinterpret_s32_p8): Likewise.
3113 (vreinterpret_s32_p16): Likewise.
3114 (vreinterpretq_s32_s8): Likewise.
3115 (vreinterpretq_s32_s16): Likewise.
3116 (vreinterpretq_s32_s64): Likewise.
3117 (vreinterpretq_s32_f32): Likewise.
3118 (vreinterpretq_s32_u8): Likewise.
3119 (vreinterpretq_s32_u16): Likewise.
3120 (vreinterpretq_s32_u32): Likewise.
3121 (vreinterpretq_s32_u64): Likewise.
3122 (vreinterpretq_s32_p8): Likewise.
3123 (vreinterpretq_s32_p16): Likewise.
3124 (vreinterpret_u8_s8): Likewise.
3125 (vreinterpret_u8_s16): Likewise.
3126 (vreinterpret_u8_s32): Likewise.
3127 (vreinterpret_u8_s64): Likewise.
3128 (vreinterpret_u8_f32): Likewise.
3129 (vreinterpret_u8_u16): Likewise.
3130 (vreinterpret_u8_u32): Likewise.
3131 (vreinterpret_u8_u64): Likewise.
3132 (vreinterpret_u8_p8): Likewise.
3133 (vreinterpret_u8_p16): Likewise.
3134 (vreinterpretq_u8_s8): Likewise.
3135 (vreinterpretq_u8_s16): Likewise.
3136 (vreinterpretq_u8_s32): Likewise.
3137 (vreinterpretq_u8_s64): Likewise.
3138 (vreinterpretq_u8_f32): Likewise.
3139 (vreinterpretq_u8_u16): Likewise.
3140 (vreinterpretq_u8_u32): Likewise.
3141 (vreinterpretq_u8_u64): Likewise.
3142 (vreinterpretq_u8_p8): Likewise.
3143 (vreinterpretq_u8_p16): Likewise.
3144 (vreinterpret_u16_s8): Likewise.
3145 (vreinterpret_u16_s16): Likewise.
3146 (vreinterpret_u16_s32): Likewise.
3147 (vreinterpret_u16_s64): Likewise.
3148 (vreinterpret_u16_f32): Likewise.
3149 (vreinterpret_u16_u8): Likewise.
3150 (vreinterpret_u16_u32): Likewise.
3151 (vreinterpret_u16_u64): Likewise.
3152 (vreinterpret_u16_p8): Likewise.
3153 (vreinterpret_u16_p16): Likewise.
3154 (vreinterpretq_u16_s8): Likewise.
3155 (vreinterpretq_u16_s16): Likewise.
3156 (vreinterpretq_u16_s32): Likewise.
3157 (vreinterpretq_u16_s64): Likewise.
3158 (vreinterpretq_u16_f32): Likewise.
3159 (vreinterpretq_u16_u8): Likewise.
3160 (vreinterpretq_u16_u32): Likewise.
3161 (vreinterpretq_u16_u64): Likewise.
3162 (vreinterpretq_u16_p8): Likewise.
3163 (vreinterpretq_u16_p16): Likewise.
3164 (vreinterpret_u32_s8): Likewise.
3165 (vreinterpret_u32_s16): Likewise.
3166 (vreinterpret_u32_s32): Likewise.
3167 (vreinterpret_u32_s64): Likewise.
3168 (vreinterpret_u32_f32): Likewise.
3169 (vreinterpret_u32_u8): Likewise.
3170 (vreinterpret_u32_u16): Likewise.
3171 (vreinterpret_u32_u64): Likewise.
3172 (vreinterpret_u32_p8): Likewise.
3173 (vreinterpret_u32_p16): Likewise.
3174 (vreinterpretq_u32_s8): Likewise.
3175 (vreinterpretq_u32_s16): Likewise.
3176 (vreinterpretq_u32_s32): Likewise.
3177 (vreinterpretq_u32_s64): Likewise.
3178 (vreinterpretq_u32_f32): Likewise.
3179 (vreinterpretq_u32_u8): Likewise.
3180 (vreinterpretq_u32_u16): Likewise.
3181 (vreinterpretq_u32_u64): Likewise.
3182 (vreinterpretq_u32_p8): Likewise.
3183 (vreinterpretq_u32_p16): Likewise.
3185 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3187 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3189 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3191 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3192 (vqnegd_s64): Likewise.
3193 (vqabs_s64): Likewise.
3194 (vqabsd_s64): Likewise.
3196 2014-04-22 Richard Henderson <rth@redhat.com>
3198 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3199 computation to the top of the loop.
3201 2014-04-22 Renlin <renlin.li@arm.com>
3202 Jiong Wang <jiong.wang@arm.com>
3204 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3205 * config/aarch64/aarch64.c (aarch64_layout_frame)
3206 (aarch64_initial_elimination_offset): Likewise.
3208 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3210 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3213 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3215 * machmode.h (bitwise_mode_for_mode): Declare.
3216 * stor-layout.h (bitwise_type_for_mode): Likewise.
3217 * stor-layout.c (bitwise_mode_for_mode): New function.
3218 (bitwise_type_for_mode): Likewise.
3219 * builtins.c (fold_builtin_memory_op): Use it instead of
3220 int_mode_for_mode and build_nonstandard_integer_type.
3222 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3224 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3225 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3226 (*-*-solaris2*): Simplify.
3227 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3228 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3229 *-*-solaris2.9* handling.
3231 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3233 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3234 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3236 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3237 * configure: Regenerate.
3239 * config/i386/sol2-9.h: Remove.
3241 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3242 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3243 Remove Solaris 9 references.
3245 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3247 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3248 (floatuns<GPI:mode><GPF:mode>2): Remove.
3249 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3250 and floatuns conversions.
3251 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3252 and floatuns conversions.
3253 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3254 (w1,w2): New mode attributes for inequal width conversions.
3256 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3258 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3259 the output asm format.
3261 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3263 * config/aarch64/aarch64-simd.md
3264 (aarch64_cm<optab>di): Always split.
3265 (*aarch64_cm<optab>di): New.
3266 (aarch64_cmtstdi): Always split.
3267 (*aarch64_cmtstdi): New.
3269 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3271 PR tree-optimization/60823
3272 * omp-low.c (ipa_simd_modify_function_body): Go through
3273 all SSA_NAMEs and for those refering to vector arguments
3274 which are going to be replaced adjust SSA_NAME_VAR and,
3275 if it is a default definition, change it into a non-default
3276 definition assigned at the beginning of function from new_decl.
3277 (ipa_simd_modify_stmt_ops): Rewritten.
3278 * tree-dfa.c (set_ssa_default_def): When removing default def,
3279 check for NULL loc instead of NULL *loc.
3281 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3283 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3284 restrictions on core registers for DImode values in Thumb2.
3286 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3288 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3289 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3291 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3293 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3294 (*iordi_notzesidi_di): Likewise.
3295 (*iordi_notsesidi_di): Likewise.
3297 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3299 * config/arm/arm-protos.h (tune_params): New struct members.
3300 * config/arm/arm.c: Initialise tune_params per processor.
3301 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3302 for speed, based on new tune_params.
3304 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3306 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3307 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3308 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3309 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3310 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3311 (vrnda_f64): Likewise.
3312 (vrndi_f64): Likewise.
3313 (vrndm_f64): Likewise.
3314 (vrndn_f64): Likewise.
3315 (vrndp_f64): Likewise.
3316 (vrndx_f64): Likewise.
3318 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3320 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3321 GET_MODE_SIZE argument is enum machine_mode.
3323 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3326 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3327 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3329 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3332 * asan.c (asan_emit_stack_protection): Force the base to align to
3333 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3334 appropriate bits if STRICT_ALIGNMENT.
3335 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3337 (expand_used_vars): Leave a space in the stack frame for alignment
3338 if STRICT_ALIGNMENT.
3340 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3342 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3344 (gimple_store_p): Likewise.
3345 (gimple_assign_load_p): Likewise.
3346 (gimple_assign_cast_p): Likewise.
3347 (gimple_clobber_p): Likewise.
3349 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3350 rather than a gimple.
3351 (gimple_assign_cast_p): Likewise.
3353 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3356 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3357 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3359 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3360 more debug information for E500 if -mdebug=reg.
3362 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3365 * config/i386/i386.c (ix86_expand_builtin)
3366 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3367 register for target RTX.
3368 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3370 2014-04-18 Cong Hou <congh@google.com>
3372 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3373 the widen-mult pattern by handling two operands with different sizes,
3374 and operands whose size is smaller than half of the result type.
3376 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3378 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3379 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3380 (do_estimate_edge_time): Compute it.
3381 * ipa-inline.c (want_inline_small_function_p): Bypass
3382 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3384 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3386 * ipa-inline.c (spec_rem): New static variable.
3387 (dump_overall_stats): New function.
3388 (dump_inline_stats): New function.
3390 2014-04-18 Richard Henderson <rth@redhat.com>
3392 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3393 to GET_MODE_SIZE, not a reg_class_t.
3395 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3397 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3398 (vsx_xxmrglw_<mode>): Likewise.
3400 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3403 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3404 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3405 (rs6000_init_hard_regno_mode_ok): Likewise.
3407 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3409 * ipa-inline.c (inline_small_functions): Account only non-cold
3411 * doc/invoke.texi (inline-unit-growth): Update documentation.
3413 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3415 * config/rs6000/rs6000.md (addti3, subti3): New.
3417 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3420 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3421 comment. Check optimize_insn_for_size_p instead of
3422 optimize_insn_for_speed_p.
3424 2014-04-17 Martin Jambor <mjambor@suse.cz>
3426 * gimple-iterator.c (gsi_start_edge): New function.
3427 * gimple-iterator.h (gsi_start_edge): Declare.
3428 * tree-sra.c (single_non_eh_succ): New function.
3429 (disqualify_ops_if_throwing_stmt): Renamed to
3430 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3431 having one non-EH successor BB.
3432 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3433 generate loads into replacements.
3434 (sra_modify_assign): Likewise and and also use the simple path for
3436 (sra_modify_function_body): Commit statements on edges.
3438 2014-04-17 Richard Biener <rguenther@suse.de>
3441 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3442 comparison results and add clarifying comment.
3444 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3446 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3447 (blank_mode): Initialize it.
3448 (emit_mode_size_inline, emit_mode_nunits_inline,
3449 emit_mode_inner_inline): New functions.
3450 (emit_insn_modes_h): Call them and surround their output with
3451 #if GCC_VERSION >= 4001 ... #endif.
3452 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3453 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3454 mode_* arrays if the argument is __builtin_constant_p.
3455 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3456 is enum machine_mode.
3458 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3460 * passes.c (opt_pass::execute): Adjust.
3461 (pass_manager::execute_pass_mode_switching): Likewise.
3462 (early_local_passes::execute): Likewise.
3463 (execute_one_pass): Pass cfun to the pass's execute method.
3464 * tree-pass.h (opt_pass::execute): Add function * argument.
3465 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3466 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3467 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3468 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3469 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3470 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3471 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3472 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3473 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3474 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3475 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3476 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3477 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3478 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3479 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3480 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3481 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3482 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3483 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3484 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3485 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3486 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3487 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3488 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3489 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3490 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3491 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3492 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3493 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3496 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3498 * passes.c (opt_pass::gate): Take function * argument.
3499 (gate_all_early_local_passes): Merge into
3500 (early_local_passes::gate): this.
3501 (gate_all_early_optimizations): Merge into
3502 (all_early_optimizations::gate): this.
3503 (gate_all_optimizations): Mege into
3504 (all_optimizations::gate): this.
3505 (gate_all_optimizations_g): Merge into
3506 (all_optimizations_g::gate): this.
3507 (gate_rest_of_compilation): Mege into
3508 (rest_of_compilation::gate): this.
3509 (gate_postreload): Merge into
3510 (postreload::gate): this.
3511 (dump_one_pass): Pass cfun to the pass's gate method.
3512 (execute_ipa_summary_passes): Likewise.
3513 (execute_one_pass): Likewise.
3514 (ipa_write_summaries_2): Likewise.
3515 (ipa_write_optimization_summaries_1): Likewise.
3516 (ipa_read_summaries_1): Likewise.
3517 (ipa_read_optimization_summaries_1): Likewise.
3518 (execute_ipa_stmt_fixups): Likewise.
3519 * tree-pass.h (opt_pass::gate): Add function * argument.
3520 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3521 combine-stack-adj.c, combine.c, compare-elim.c,
3522 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3523 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3524 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3525 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3526 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3527 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3528 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3529 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3530 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3531 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3532 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3533 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3534 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3535 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3536 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3537 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3538 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3539 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3540 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3541 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3542 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3543 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3544 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3545 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3546 var-tracking.c, vtable-verify.c, web.c: Adjust.
3548 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3550 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3551 * configure: Regenerate.
3553 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3555 * passes.c (dump_one_pass): don't check pass->has_gate.
3556 (execute_ipa_summary_passes): Likewise.
3557 (execute_one_pass): Likewise.
3558 (ipa_write_summaries_2): Likewise.
3559 (ipa_write_optimization_summaries_1): Likewise.
3560 (ipa_read_optimization_summaries_1): Likewise.
3561 (execute_ipa_stmt_fixups): Likewise.
3562 * tree-pass.h (pass_data::has_gate): Remove.
3563 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3564 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3565 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3566 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3567 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3568 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3569 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3570 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3571 gimple-low.c, gimple-ssa-isolate-paths.c,
3572 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3573 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3574 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3575 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3576 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3577 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3578 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3579 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3580 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3581 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3582 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3583 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3584 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3585 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3586 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3587 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3588 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3589 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3590 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3591 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3592 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3593 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3596 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3598 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3600 * passes.c (pass_manager::register_dump_files_1): Merge into
3601 (pass_manager::register_dump_files): this, and remove its handling of
3602 properties since the pass always has the properties anyway.
3603 (pass_manager::pass_manager): Adjust.
3605 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3607 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3608 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3609 dealing with properties.
3610 (pass_manager::register_dump_files): Adjust.
3612 2014-03-20 Mark Wielaard <mjw@redhat.com>
3614 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3615 then represent the bound as normal constant value.
3617 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3620 Forward port from 4.8 branch
3621 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3623 * config/i386/bmiintrin.h (_blsi_u32): New.
3627 (_blsmsk_u32): Ditto.
3628 (_blsmsk_u64): Ditto.
3629 (_tzcnt_u32): Ditto.
3630 (_tzcnt_u64): Ditto.
3632 2014-04-17 Kito Cheng <kito@0xlab.org>
3634 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3636 2014-04-17 Richard Biener <rguenther@suse.de>
3639 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3640 boolean results for comparisons.
3642 2014-04-17 Richard Biener <rguenther@suse.de>
3644 PR tree-optimization/60836
3645 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3646 initial PHI args to be gimple values.
3648 2014-04-17 Richard Biener <rguenther@suse.de>
3650 PR tree-optimization/60841
3651 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3652 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3653 of stmts to SLP build.
3654 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3655 (vect_analyze_slp): Likewise.
3656 (vect_analyze_slp_instance): Likewise.
3657 (vect_build_slp_tree): Limit overall SLP tree growth.
3658 * tree-vectorizer.h (vect_analyze_data_refs,
3659 vect_analyze_slp): Adjust prototypes.
3661 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3663 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3666 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3668 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3669 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3670 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3671 for TARGET_SLOW_PSHUFB
3673 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3675 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3676 * config/i386/i386.c (intel_cost): Ditto.
3678 2014-04-17 Joey Ye <joey.ye@arm.com>
3680 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3682 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3684 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3687 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3689 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3690 (type_all_derivations_known_p): New predicate.
3691 (type_all_ctors_visible_p): New predicate.
3692 (type_possibly_instantiated_p): New predicate.
3693 (get_odr_type): Compute all_derivations_known.
3694 (dump_odr_type): Dump the flag.
3695 (maybe_record_type): Cleanup.
3696 (record_target_from_binfo): Add bases_to_consider array;
3697 record bases for types w/o instances and skip CXX destructor.
3698 (possible_polymorphic_call_targets_1): Add bases_to_consider
3699 and consider_construction parameters; check if type may have instance.
3700 (get_polymorphic_call_info): Set maybe_in_construction to true
3701 when we know nothing.
3702 (record_targets_from_bases): Skip CXX destructors; they are
3703 never called for types in construction.
3704 (possible_polymorphic_call_targets): Do not record target when
3705 type may not have instance.
3707 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3710 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3711 external aliases alive, too.
3713 2014-04-16 Andrew Pinski <apinski@cavium.com>
3715 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3718 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3720 * final.c (compute_alignments): Do not apply loop alignment to a block
3721 falling through to the exit.
3723 2014-04-16 Catherine Moore <clm@codesourcery.com>
3725 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3726 Adjust constraints for microMIPS store patterns.
3728 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3730 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3732 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3734 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3735 (append_use): Run at -O0.
3736 (append_vdef): Likewise.
3737 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3738 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3740 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3742 PR tree-optimization/60844
3743 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3744 (propagate_op_to_single_use, remove_visited_stmt_chain,
3745 linearize_expr, repropagate_negates, reassociate_bb): Use it
3746 instead of gsi_remove.
3748 2014-04-16 Martin Jambor <mjambor@suse.cz>
3750 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3751 ipa_transforms_to_apply.
3752 (cgraph_function_versioning): Assert that old_node has empty
3753 ipa_transforms_to_apply.
3754 * trans-mem.c (ipa_tm_create_version): Likewise.
3755 * tree-inline.c (tree_function_versioning): Do not duplicate
3756 ipa_transforms_to_apply.
3758 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3761 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3763 Pass necessary as flags on 64-bit Solaris/x86.
3764 Use lowercase relocs for x86_64-*-*.
3765 * configure: Regenerate.
3767 2014-04-15 Jan Hubicka <jh@suse.cz>
3769 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3770 (maybe_record_node, likely_target_p): Use it.
3772 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3775 Revert following patch
3777 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3780 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3781 software floating point or no floating point registers, do not
3782 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3783 in GPRs that occurs after we tested for GPRs that would never be
3786 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3787 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3788 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3789 specifically allow DDmode, since that does not use the SPE SIMD
3792 2014-03-21 Mark Wielaard <mjw@redhat.com>
3794 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3795 as unsigned or int depending on type and value used.
3797 2014-04-15 Richard Biener <rguenther@suse.de>
3799 PR rtl-optimization/56965
3800 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3801 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3803 * alias.c (true_dependence_1): Do not call
3804 nonoverlapping_component_refs_p.
3805 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3806 nonoverlapping_component_refs_p.
3807 (indirect_refs_may_alias_p): Likewise.
3809 2014-04-15 Teresa Johnson <tejohnson@google.com>
3811 * cfg.c (dump_bb_info): Fix flags check.
3812 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3814 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3816 PR rtl-optimization/60663
3817 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3820 2014-04-15 Richard Biener <rguenther@suse.de>
3822 * lto-streamer.h (LTO_major_version): Bump to 4.
3824 2014-04-15 Richard Biener <rguenther@suse.de>
3826 * common.opt (lto_partition_model): New enum.
3827 (flto-partition=): Merge separate options with a single with argument,
3828 add -flto-partition=one support.
3829 * flag-types.h (enum lto_partition_model): Declare.
3830 * opts.c (finish_options): Remove duplicate -flto-partition=
3832 * lto-wrapper.c (run_gcc): Adjust.
3834 2014-04-15 Richard Biener <rguenther@suse.de>
3836 * alias.c (ncr_compar): New function.
3837 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3839 2014-04-15 Richard Biener <rguenther@suse.de>
3841 * alias.c (record_component_aliases): Do not walk BINFOs.
3843 2014-04-15 Richard Biener <rguenther@suse.de>
3845 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3846 Add struct function argument and adjust.
3847 (find_func_aliases_for_call): Likewise.
3848 (find_func_aliases): Likewise.
3849 (find_func_clobbers): Likewise.
3850 (intra_create_variable_infos): Likewise.
3851 (compute_points_to_sets): Likewise.
3852 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3854 2014-04-15 Richard Biener <rguenther@suse.de>
3856 * tree.c (iterative_hash_expr): Use enum tree_code_class
3857 to store TREE_CODE_CLASS.
3858 (tree_block): Likewise.
3859 (tree_set_block): Likewise.
3860 * tree.h (fold_build_pointer_plus_loc): Use
3861 convert_to_ptrofftype_loc.
3863 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3866 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3869 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3871 * cfgloop.h (struct loop): Move force_vectorize down.
3872 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3873 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3874 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3875 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3876 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3877 * tree-core.h (enum annot_expr_kind): Add new kind values.
3878 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3879 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3881 * tree.def (ANNOTATE_EXPR): Tweak comment.
3883 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3885 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3888 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3890 * tree.h (TYPE_IDENTIFIER): Declare.
3891 * tree.c (subrange_type_for_debug_p): Use it.
3892 * godump.c (go_format_type): Likewise.
3893 * dwarf2out.c (is_cxx_auto, modified_type_die,
3894 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3895 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3897 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3900 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3902 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3904 * config/i386/i386.c (examine_argument): Return bool. Return true if
3905 parameter should be passed in memory.
3906 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3907 (construct_container): Update calls to examine_argument.
3908 (function_arg_advance_64): Ditto.
3909 (return_in_memory_32): Merge with ix86_return_in_memory.
3910 (return_in_memory_64): Ditto.
3911 (return_in_memory_ms_64): Ditto.
3913 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3915 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3916 * coverage.c (coverage_compute_profile_id): Handle externally visible
3919 2014-04-14 Martin Jambor <mjambor@suse.cz>
3921 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3922 DECL_DISREGARD_INLINE_LIMITS functions.
3924 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3927 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3929 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3932 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3933 optimize_insn_for_speed_p instead of
3934 optimize_function_for_speed_p.
3936 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3938 * doc/invoke.texi (free): Document AArch64.
3940 2014-04-14 Richard Biener <rguenther@suse.de>
3942 PR tree-optimization/60042
3943 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3944 (insert_into_preds_of_block): Do not prevent PHI insertion
3945 for REFERENCE exprs here ...
3946 (eliminate_dom_walker::before_dom_children): ... but prevent
3947 their use here under similar conditions when applied to the
3948 IL after PRE optimizations.
3950 2014-04-14 Richard Biener <rguenther@suse.de>
3952 * passes.def: Move early points-to after early SRA.
3954 2014-04-14 Richard Biener <rguenther@suse.de>
3956 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3957 check for which sign-changes we allow when forwarding
3958 a converted value into a switch.
3960 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3962 * stor-layout.c (place_field): Finalize non-constant offset for the
3965 2014-04-14 Richard Biener <rguenther@suse.de>
3967 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3969 (expand_switch_using_bit_tests_p): Likewise.
3970 (process_switch): Compute and pass on speed_p based on the
3972 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3973 optimize_bb_for_speed_p.
3975 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3977 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3978 * function.h (struct function): Rename has_force_vect_loops into
3979 has_force_vectorize_loops.
3980 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3981 (input_struct_function_base): Likewise.
3982 * lto-streamer-out.c (output_cfg): Likewise.
3983 (output_struct_function_base): Likewise.
3984 * omp-low.c (expand_omp_simd): Likewise.
3985 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3986 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3987 (version_loop_for_if_conversion): Likewise.
3988 (tree_if_conversion): Likewise.
3989 (main_tree_if_conversion): Likewise.
3990 (gate_tree_if_conversion): Likewise.
3991 * tree-inline.c (copy_loops): Likewise.
3992 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3993 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3994 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3995 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3996 * tree-vectorizer.c (vectorize_loops): Likewise.
3997 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3999 2014-04-14 Richard Biener <rguenther@suse.de>
4002 * lto-streamer-out.c (wrap_refs): New function.
4003 (lto_output): Wrap symbol references in global initializes in
4004 type-preserving MEM_REFs.
4006 2014-04-14 Christian Bruel <christian.bruel@st.com>
4008 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4010 2014-04-14 Christian Bruel <christian.bruel@st.com>
4012 * config/sh/sh.md (setmemqi): New expand pattern.
4013 * config/sh/sh.h (CLEAR_RATIO): Define.
4014 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4015 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4017 2014-04-14 Richard Biener <rguenther@suse.de>
4020 * fold-const.c (negate_expr_p): Don't negate directional rounding
4022 (fold_negate_expr): Likewise.
4024 2014-04-14 Richard Biener <rguenther@suse.de>
4026 PR tree-optimization/59817
4027 PR tree-optimization/60453
4028 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4029 recursion to catch all CHRECs in the scalar evolution and restrict
4030 the predicate for the remains appropriately.
4032 2014-04-12 Catherine Moore <clm@codesourcery.com>
4034 * config/mips/constraints.md: Add new register constraint "kb".
4035 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4036 (*movhi_internal): Likewise.
4037 (*movqi_internal): Likewise.
4038 * config/mips/mips.h (M16_STORE_REGS): New register class.
4039 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4040 (REG_CLASS_CONTENTS): Likewise.
4041 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4043 2014-04-11 Tobias Burnus <burnus@net-b.de>
4046 * doc/invoke.texi (-Wformat-signedness): Document it.
4047 (Wformat=2): Mention that this enables -Wformat-signedness.
4049 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4051 * common/config/epiphany/epiphany-common.c
4052 (epiphany_option_optimization_table): Enable section anchors by
4053 default at -O1 or higher.
4054 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4055 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4056 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4057 carries no extra cost.
4058 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4059 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4060 * config/epiphany/predicates.md (memclob_operand): New predicate.
4061 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4062 Use memclob_operand predicate and X constraint for operand 3.
4064 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4066 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4067 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4070 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4072 PR rtl-optimization/60651
4073 * mode-switching.c (optimize_mode_switching): Make sure to emit
4074 sets of a lower numbered entity before sets of a higher numbered
4075 entity to a mode of the same or lower priority.
4076 When creating a seginfo for a basic block that starts with a code
4077 label, move the insertion point past the code label.
4078 (new_seginfo): Document and enforce requirement that
4079 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4080 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4081 * doc/tm.texi: Regenerate.
4083 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4086 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4088 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4090 * BASE-VER: Set to 4.10.0.
4092 2014-04-11 Tobias Burnus <burnus@net-b.de>
4095 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4096 * doc/gcc.texi (Service): Update description in the @menu
4097 * doc/invoke.texi (Option Summary): Remove misplaced and
4100 2014-04-11 Steve Ellcey <sellcey@mips.com>
4101 Jakub Jelinek <jakub@redhat.com>
4104 * expr.c (convert_move): Use emit_store_flag_force instead of
4105 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4108 2014-04-11 Richard Biener <rguenther@suse.de>
4111 * varasm.c (assemble_alias): Avoid endless error reporting
4112 recursion by setting TREE_ASM_WRITTEN.
4114 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4116 * config/s390/s390.md: Add a splitter for NOT rtx.
4118 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4120 PR rtl-optimization/60663
4121 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4123 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4124 Jakub Jelinek <jakub@redhat.com>
4127 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4128 flag from decl_node to node.
4130 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4133 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4134 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4135 ameliorating the cases where it can be.
4137 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4140 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4142 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4143 (loadsync_<mode>): Change mode.
4144 (load_quadpti, store_quadpti): New.
4145 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4146 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4147 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4149 2014-04-09 Cong Hou <congh@google.com>
4152 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4155 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4157 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4158 instead of vnor to exploit possible fusion opportunity in the
4160 (altivec_expand_vec_perm_const_le): Likewise.
4162 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4164 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4165 (loadsync_<mode>): Change mode.
4166 (load_quadpti, store_quadpti): New.
4167 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4168 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4170 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4173 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4174 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4175 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4177 2014-04-08 Richard Biener <rguenther@suse.de>
4180 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4181 a 64bit widest int print double-int similar to on HWI64 hosts.
4183 2014-04-08 Richard Biener <rguenther@suse.de>
4185 PR tree-optimization/60785
4186 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4187 default defs properly.
4189 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4191 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4192 (Weffc++): Likewise.
4194 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4196 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4197 set completep to false rather than true.
4199 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4202 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4203 ARM_TARGET2_DWARF_FORMAT.
4205 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4208 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4209 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4212 2014-04-07 Richard Biener <rguenther@suse.de>
4214 PR tree-optimization/60766
4215 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4216 (may_eliminate_iv): Convert cand_value_at result to desired type.
4218 2014-04-07 Jason Merrill <jason@redhat.com>
4221 * common.opt (-fno-gnu-unique): Add.
4222 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4224 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4226 * haifa-sched.c: Fix outdated function reference and minor
4227 grammar errors in introductory comment.
4229 2014-04-07 Richard Biener <rguenther@suse.de>
4232 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4234 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4236 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4239 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4240 size accounting for thunks.
4241 (pa_asm_output_mi_thunk): Use final_start_function() and
4242 final_end_function() to output function start and end directives.
4244 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4246 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4247 device specific ISA/ feature information. Remove short_sp and
4248 errata_skip ds. Add avr_device_specific_features enum to have device
4250 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4251 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4252 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4253 updated device specific info.
4254 * config/avr/avr-mcus.def: Merge device specific details to
4255 dev_attribute field.
4256 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4258 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4259 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4260 assembler if RMW isa supported by current device.
4261 * config/avr/genmultilib.awk: Update as device info structure changed.
4262 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4264 2014-04-04 Cong Hou <congh@google.com>
4266 PR tree-optimization/60656
4267 * tree-vect-stmts.c (supportable_widening_operation):
4268 Fix a bug that elements in a vector with vect_used_by_reduction
4269 property are incorrectly reordered when the operation on it is not
4270 consistant with the one in reduction operation.
4272 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4274 PR rtl-optimization/60155
4275 * gcse.c (record_set_data): New function.
4276 (single_set_gcse): New function.
4277 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4278 (hoist_code): Likewise.
4279 (get_pressure_class_and_nregs): Likewise.
4281 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4283 * explow.c (probe_stack_range): Emit a final optimization blockage.
4285 2014-04-04 Anthony Green <green@moxielogic.com>
4287 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4290 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4293 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4294 flags are set only during streaming.
4295 * ipa.c (process_references, walk_polymorphic_call_targets,
4296 symtab_remove_unreachable_nodes): Drop bodies of always inline
4297 after early inlining.
4298 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4300 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4301 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4304 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4307 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4310 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4312 (cortex_a53_fdivd): Likewise.
4314 2014-04-04 Martin Jambor <mjambor@suse.cz>
4317 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4319 * cgraph.c (clone_of_p): Also return true if thunks match.
4320 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4321 cgraph_function_or_thunk_node and an obsolete comment.
4322 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4324 (build_function_decl_skip_args): Likewise.
4325 (set_new_clone_decl_and_node_flags): New function.
4326 (duplicate_thunk_for_node): Likewise.
4327 (redirect_edge_duplicating_thunks): Likewise.
4328 (cgraph_clone_node): New parameter args_to_skip, pass it to
4329 redirect_edge_duplicating_thunks which is called instead of
4330 cgraph_redirect_edge_callee.
4331 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4332 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4334 2014-04-04 Jeff Law <law@redhat.com>
4337 * config/arm/predicates.md (const_int_I_operand): New predicate.
4338 (const_int_M_operand): Similarly.
4339 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4341 (insv_t2, extv_reg, extzv_t2): Likewise.
4342 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4343 (pop_multiple_with_writeback_and_return): Likewise.
4344 (vfp_pop_multiple_with_writeback): Likewise
4346 2014-04-04 Richard Biener <rguenther@suse.de>
4349 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4350 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4352 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4353 * gimplify.c (gimple_add_tmp_var_fn): New function.
4354 * gimple-expr.h (create_tmp_reg_fn): Declare.
4355 * gimple-expr.c (create_tmp_reg_fn): New function.
4356 * gimple-low.c (record_vars_into): Don't change cfun.
4357 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4358 code generation without cfun.
4360 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4363 * Makefile.in (install-driver): Fix shell scripting.
4365 2014-04-03 Cong Hou <congh@google.com>
4367 PR tree-optimization/60505
4368 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4369 threshold of number of iterations below which no vectorization
4371 * tree-vect-loop.c (new_loop_vec_info):
4372 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4373 * tree-vect-loop.c (vect_analyze_loop_operations):
4374 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4375 * tree-vect-loop.c (vect_transform_loop):
4376 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4377 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4378 of iterations of the loop and see if we should build the epilogue.
4380 2014-04-03 Richard Biener <rguenther@suse.de>
4382 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4383 (streamer_tree_cache_create): Adjust.
4384 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4385 to allow optional nodes array.
4386 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4387 (streamer_tree_cache_append): Likewise.
4388 (streamer_tree_cache_create): Create nodes array optionally
4389 as specified by parameter.
4390 * lto-streamer-out.c (create_output_block): Avoid maintaining
4391 the node array in the writer cache.
4392 (DFS_write_tree): Remove assertion.
4393 (produce_asm_for_decls): Free the out decl state hash table early.
4394 * lto-streamer-in.c (lto_data_in_create): Adjust for
4395 streamer_tree_cache_create prototype change.
4397 2014-04-03 Richard Biener <rguenther@suse.de>
4399 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4400 set TREE_CHAIN to NULL_TREE.
4402 2014-04-03 Richard Biener <rguenther@suse.de>
4404 PR tree-optimization/60740
4405 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4406 over all GIMPLE_COND operands.
4408 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4410 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4411 (Weffc++): Remove Scott's numbering, merge lists and reference
4414 2014-04-03 Nick Clifton <nickc@redhat.com>
4416 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4419 2014-04-03 Martin Jambor <mjambor@suse.cz>
4421 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4422 mention gcc_unreachable before failing.
4423 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4426 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4429 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4430 inconsistent code and instead mark the context inconsistent.
4431 (possible_polymorphic_call_targets): For inconsistent contexts
4432 return empty complete list.
4434 2014-04-02 Anthony Green <green@moxielogic.com>
4436 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4437 (extendqisi2, extendhisi2): Define.
4438 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4439 (WCHAR_TYPE): Change to unsigned int.
4441 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4443 PR tree-optimization/60733
4444 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4445 insertion point for PHI candidates to be the end of the feeding
4446 block for the PHI argument.
4448 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4450 PR rtl-optimization/60650
4451 * lra-constraints.c (process_alt_operands): Decrease reject for
4452 earlyclobber matching.
4454 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4456 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4458 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4460 * config/spu/spu.c (pad_bb): Do not crash when the last
4461 insn is CODE_FOR_blockage.
4463 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4465 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4466 lies outside the target mode.
4468 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4471 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4472 software floating point or no floating point registers, do not
4473 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4474 in GPRs that occurs after we tested for GPRs that would never be
4477 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4478 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4479 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4480 specifically allow DDmode, since that does not use the SPE SIMD
4483 2014-04-02 Richard Biener <rguenther@suse.de>
4486 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4487 MODE_INTs. Properly use negv_optab.
4488 (expand_abs): Likewise.
4490 2014-04-02 Richard Biener <rguenther@suse.de>
4493 * Makefile.in (install-driver): Guard extra installs with special
4496 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4498 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4501 2014-04-01 Richard Henderson <rth@redhat.com>
4504 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4505 alternative enabled before register allocation.
4507 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4509 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4510 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4512 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4513 (nios2_got_address): Update nios2_large_got_address call site.
4514 (nios2_delegitimize_address): New function.
4515 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4516 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4517 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4519 2014-04-01 Martin Husemann <martin@duskware.de>
4521 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4524 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4526 PR rtl-optimization/60604
4527 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4528 check from register_operand.
4529 (register_operand): Redefine in terms of general_operand.
4530 (nonmemory_operand): Use register_operand for the non-constant cases.
4532 2014-04-01 Richard Biener <rguenther@suse.de>
4534 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4536 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4538 * doc/invoke.texi (mapp-regs): Clarify.
4540 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4542 * config/i386/avx512fintrin.h (__v32hi): Define type.
4543 (__v64qi): Likewise.
4544 (_mm512_set1_epi8): Define.
4545 (_mm512_set1_epi16): Define.
4546 (_mm512_set4_epi32): Define.
4547 (_mm512_set4_epi64): Define.
4548 (_mm512_set4_pd): Define.
4549 (_mm512_set4_ps): Define.
4550 (_mm512_setr4_epi64): Define.
4551 (_mm512_setr4_epi32): Define.
4552 (_mm512_setr4_pd): Define.
4553 (_mm512_setr4_ps): Define.
4554 (_mm512_setzero_epi32): Define.
4556 2014-03-31 Martin Jambor <mjambor@suse.cz>
4559 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4560 callsite_arguments_match_p. Updated all callers. Also check types of
4561 corresponding formal parameters and actual arguments.
4562 (not_all_callers_have_enough_arguments_p) Renamed to
4563 some_callers_have_mismatched_arguments_p.
4565 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4567 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4569 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4572 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4575 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4577 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4578 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4580 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4581 Use FMAMODE_NOVF512 mode iterator.
4582 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4583 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4584 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4585 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4587 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4588 Use VF_128_256 mode iterator.
4589 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4592 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4594 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4595 static chain if needed.
4597 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4600 * lra-constraints.c (index_part_to_reg): New.
4601 (process_address): Use it.
4603 2014-03-27 Jeff Law <law@redhat.com>
4604 Jakub Jelinek <jakub@redhat.com>
4607 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4608 gen_rtx_{PLUS,MULT} to build up the address expression.
4610 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4611 creating non-canonical RTL.
4613 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4616 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4617 functions; reorganize to make cheap checks first.
4618 (inline_small_functions): Do not estimate growth when dumping;
4620 * ipa-inline.h (inline_summary): Add min_size.
4621 (growth_likely_positive): New function.
4622 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4623 (set_cond_stmt_execution_predicate): Cleanup.
4624 (estimate_edge_size_and_time): Compute min_size.
4625 (estimate_calls_size_and_time): Likewise.
4626 (estimate_node_size_and_time): Likewise.
4627 (inline_update_overall_summary): Update min_size.
4628 (do_estimate_edge_time): Likewise.
4629 (do_estimate_edge_size): Update.
4630 (do_estimate_edge_hints): Update.
4631 (growth_likely_positive): New function.
4633 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4636 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4637 also if addr has VOIDmode.
4639 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4641 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4642 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4644 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4645 instructions as well as AdvancedSIMD loads.
4647 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4649 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4650 Use crypto_aese type.
4651 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4652 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4653 crypto_aese, crypto_aesmc. Move to types.md.
4654 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4656 * config/arm/iterators.md (crypto_type): Likewise.
4658 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4660 * cgraph.c: Include expr.h and tree-dfa.h.
4661 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4664 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4667 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4668 regs from checking multi-reg pseudos.
4670 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4672 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4674 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4676 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4677 if it would clobber the stack pointer, even temporarily.
4679 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4681 * mode-switching.c: Make small adjustments to the top comment.
4683 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4685 * config/rs6000/constraints.md (wD constraint): New constraint to
4686 match the constant integer to get the top DImode/DFmode out of a
4687 vector in a VSX register.
4689 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4690 match the constant integer to get the top DImode/DFmode out of a
4691 vector in a VSX register.
4693 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4696 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4699 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4700 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4702 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4703 Optimize vec_extract of 64-bit values, where the value being
4704 extracted is in the top word, where we can use scalar
4705 instructions. Add direct move and store support. Combine the big
4706 endian/little endian vector select load support into a single insn.
4707 (vsx_extract_<mode>_internal1): Likewise.
4708 (vsx_extract_<mode>_internal2): Likewise.
4709 (vsx_extract_<mode>_load): Likewise.
4710 (vsx_extract_<mode>_store): Likewise.
4711 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4712 combined into vsx_extract_<mode>_load.
4713 (vsx_extract_<mode>_one_le): Likewise.
4715 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4716 define the top 64-bit vector element.
4718 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4721 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4722 Document vec_vbpermq builtin.
4725 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4726 enable use of xxsldwi and xxpermdi builtin functions.
4727 (vec_xxpermdi): Likewise.
4729 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4730 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4732 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4734 PR rtl-optimization/60650
4735 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4737 (find_spills_for): New.
4738 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4739 Spill all pseudos on the second iteration.
4741 2014-03-27 Marek Polacek <polacek@redhat.com>
4744 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4747 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4749 * config/s390/s390.c (s390_can_use_return_insn): Check for
4750 call-saved FPRs on 31 bit.
4752 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4755 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4756 if they need regimplification, just drop them instead of
4757 calling gimple_regimplify_operands on them.
4759 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4762 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4763 (aarch64_frame_pointer_required): Adjust logic.
4764 (aarch64_can_eliminate): Adjust logic.
4765 (aarch64_override_options_after_change): Adjust logic.
4767 2014-03-27 Dehao Chen <dehao@google.com>
4769 * ipa-inline.c (early_inliner): Update node's inline info.
4771 2014-03-26 Dehao Chen <dehao@google.com>
4773 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4774 compiler inserted conditional jumps for NAN float check.
4776 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4778 * ubsan.h (ubsan_create_data): Change second argument's type
4779 to const location_t *.
4780 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4782 (ubsan_create_data): Change second argument to const location_t *PLOC.
4783 Create Loc field whenever PLOC is non-NULL.
4784 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4785 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4789 * real.c (real_to_integer2): Change type of low to UHWI.
4791 2014-03-26 Tobias Burnus <burnus@net-b.de>
4793 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4794 (CILK_SELF_SPECS): New define.
4795 (driver_self_specs): Use it.
4797 2014-03-26 Richard Biener <rguenther@suse.de>
4799 * tree-pretty-print.c (percent_K_format): Implement special
4800 case for LTO and its stripped down BLOCK tree.
4802 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4805 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4807 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4808 one range is range_int_cst_p, but not both, at least optimize
4809 addition/subtraction of 0 and multiplication by 0 or 1.
4810 * gimple-fold.c (gimple_fold_call): Fold
4811 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4812 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4813 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4815 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4817 PR rtl-optimization/60452
4818 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4819 <case REG>: Return 1 for invalid offsets from the frame pointer.
4821 2014-03-26 Marek Polacek <polacek@redhat.com>
4824 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4827 2014-03-26 Marek Polacek <polacek@redhat.com>
4830 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4833 2014-03-26 Marek Polacek <polacek@redhat.com>
4836 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4837 multiplication in unsigned type.
4839 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4841 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4843 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4845 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4847 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4850 * cif-code.def (UNREACHABLE) New code.
4851 * ipa-inline.c (inline_small_functions): Skip edges to
4852 __builtlin_unreachable.
4853 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4854 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4855 predicate to __bulitin_unreachable.
4856 (set_cond_stmt_execution_predicate): Fix issue when
4857 invert_tree_comparison returns ERROR_MARK.
4858 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4859 propagate to inline clones.
4860 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4862 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4863 * cgraphclones.c (cgraph_clone_node): If call destination is already
4864 ureachable, do not redirect it back.
4865 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4868 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4870 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4871 Do not modify inline clones.
4873 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4875 * config/i386/i386.md (general_sext_operand): New mode attr.
4876 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4877 don't generate (sign_extend (const_int)).
4878 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4879 operands[2]. Use We constraint instead of <i> and
4880 <general_sext_operand> predicate instead of <general_operand>.
4881 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4882 * config/i386/constraints.md (We): New constraint.
4883 * config/i386/predicates.md (x86_64_sext_operand,
4884 sext_operand): New predicates.
4886 2014-03-25 Martin Jambor <mjambor@suse.cz>
4889 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4890 inconsistent devirtualizations to __builtin_unreachable.
4892 2014-03-25 Marek Polacek <polacek@redhat.com>
4895 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4897 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4899 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4900 order of elements for big-endian.
4902 2014-03-25 Richard Biener <rguenther@suse.de>
4905 * gimplify-me.c (gimple_regimplify_operands): Update the
4908 2014-03-25 Martin Jambor <mjambor@suse.cz>
4911 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4912 (lto_output_varpool_node): Likewise.
4913 (input_overwrite_node): Likewise.
4914 (input_varpool_node): Likewise.
4916 2014-03-25 Richard Biener <rguenther@suse.de>
4918 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4919 (run_gcc): Likewise.
4921 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4923 * combine.c (simplify_compare_const): Add MODE argument.
4924 Handle mode_width 0 as very large mode_width.
4925 (try_combine, simplify_comparison): Adjust callers.
4927 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4928 type to avoid signed integer overflow.
4929 * explow.c (plus_constant): Likewise.
4931 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4933 * doc/generic.texi: Correct typos.
4935 2014-03-24 Tobias Burnus <burnus@net-b.de>
4937 * doc/invoke.texi (-flto): Expand section about
4938 using static libraries with LTO.
4940 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4942 PR rtl-optimization/60501
4943 * optabs.def (addptr3_optab): New optab.
4944 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4945 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4946 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4948 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4950 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4952 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4954 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4957 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4958 (_mm256_undefined_ps): Define.
4959 (_mm256_undefined_pd): Define.
4960 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4961 (_mm_undefined_pd): Define.
4962 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4963 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4964 (_mm512_undefined_ps): Define.
4965 (_mm512_undefined_pd): Define.
4966 Use _mm*_undefined_*.
4967 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4969 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4971 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4972 (lshr_simd): DI mode added.
4973 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4974 (aarch64_ushr_simddi): Likewise.
4975 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4976 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4977 (vshrd_n_u64): Likewise.
4979 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4981 * Makefile.in (s-macro_list): Depend on cc1.
4983 2014-03-23 Teresa Johnson <tejohnson@google.com>
4985 * ipa-utils.c (ipa_print_order): Use specified dump file.
4987 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4989 PR rtl-optimization/60601
4990 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4992 * gcc.c (eval_spec_function): Initialize save_growing_value.
4994 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4997 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4998 code == MINUS_EXPR, never swap op0 with op1.
5000 * toplev.c (init_local_tick): Avoid signed integer multiplication
5002 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5003 shift by first operand's bitsize.
5005 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5008 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5010 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5011 TARGET_ISA_64BIT_P(x).
5013 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5015 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5016 pattern for vector nor instead of subtract from splat(-1).
5017 (altivec_expand_vec_perm_const_le): Likewise.
5019 2014-03-21 Richard Henderson <rth@twiddle.net>
5022 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5023 related insns after epilogue_completed.
5025 2014-03-21 Martin Jambor <mjambor@suse.cz>
5028 * cgraph.h (symtab_node): New flag body_removed.
5029 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5030 when removing bodies.
5031 * symtab.c (dump_symtab_base): Dump body_removed flag.
5032 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5033 had their bodies removed.
5035 2014-03-21 Martin Jambor <mjambor@suse.cz>
5038 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5041 2014-03-21 Richard Biener <rguenther@suse.de>
5043 PR tree-optimization/60577
5044 * tree-core.h (struct tree_base): Document nothrow_flag use
5045 in VAR_DECL_NONALIASED.
5046 * tree.h (VAR_DECL_NONALIASED): New.
5047 (may_be_aliased): Adjust.
5048 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5050 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5052 * expr.c (expand_expr_real_1): Remove outdated comment.
5054 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5057 * ira.c (adjust_cleared_regs): Call copy_rtx on
5058 *reg_equiv[REGNO (loc)].src_p before passing it to
5059 simplify_replace_fn_rtx.
5062 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5063 into CONST, put pic register as first operand of PLUS. Use
5064 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5066 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5068 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5070 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5072 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5073 around for store forwarding issue in the FPU on the UT699.
5074 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5075 loads and operations if -mfix-ut699 is specified.
5076 (divtf3_hq): Tweak attribute.
5077 (sqrttf2_hq): Likewise.
5079 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5081 * calls.c (store_one_arg): Remove incorrect const qualification on the
5082 type of the temporary.
5083 * cfgexpand.c (expand_return): Likewise.
5084 * expr.c (expand_constructor): Likewise.
5085 (expand_expr_real_1): Likewise.
5087 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5089 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5092 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5095 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5097 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5099 * config/arm/aarch-common-protos.h
5100 (alu_cost_table): Fix spelling of "extend".
5101 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5103 2014-03-19 Richard Biener <rguenther@suse.de>
5106 * tree-core.h (tree_type_common): Re-order pointer members
5107 to reduce recursion depth during GC walks.
5109 2014-03-19 Marek Polacek <polacek@redhat.com>
5112 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5113 before accessing it.
5115 2014-03-19 Richard Biener <rguenther@suse.de>
5118 * lto-streamer-in.c (input_function): In WPA stage do not drop
5121 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5123 PR tree-optimization/60559
5124 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5125 with build_zero_cst assignment.
5127 2014-03-18 Kai Tietz <ktietz@redhat.com>
5129 PR rtl-optimization/56356
5130 * sdbout.c (sdbout_parms): Verify that parms'
5131 incoming argument is valid.
5132 (sdbout_reg_parms): Likewise.
5134 2014-03-18 Richard Henderson <rth@redhat.com>
5137 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5138 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5139 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5141 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5143 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5144 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5145 Italicize plugin event names in description. Explain that
5146 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5147 Remind that no GCC functions should be called after PLUGIN_FINISH.
5148 Explain what pragmas with expansion are.
5150 2014-03-18 Martin Liska <mliska@suse.cz>
5152 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5153 gimple call statement is update.
5154 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5155 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5157 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5160 * ubsan.c (ubsan_instrument_unreachable): Call
5161 initialize_sanitizer_builtins.
5162 (ubsan_pass): Likewise.
5165 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5166 varpool_finalize_decl instead of rest_of_decl_compilation.
5168 2014-03-18 Richard Biener <rguenther@suse.de>
5170 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5171 by using bitmap_and_compl instead of bitmap_and_compl_into.
5172 (df_rd_transfer_function): Likewise.
5174 2014-03-18 Richard Biener <rguenther@suse.de>
5176 * doc/lto.texi (fresolution): Fix typo.
5178 2014-03-18 Richard Biener <rguenther@suse.de>
5180 * doc/invoke.texi (flto): Update for changes in 4.9.
5182 2014-03-18 Richard Biener <rguenther@suse.de>
5184 * doc/loop.texi: Remove section on the removed lambda framework.
5185 Update loop docs with recent changes in preserving loop structure.
5187 2014-03-18 Richard Biener <rguenther@suse.de>
5189 * doc/lto.texi (-fresolution): Document.
5191 2014-03-18 Richard Biener <rguenther@suse.de>
5193 * doc/contrib.texi: Adjust my name.
5195 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5198 * internal-fn.c: Include diagnostic-core.h.
5199 (expand_BUILTIN_EXPECT): New function.
5200 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5201 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5202 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5203 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5205 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5206 Revert 3 argument __builtin_expect code.
5207 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5208 * gimple-fold.c (gimple_fold_call): Likewise.
5209 * tree.h (fold_builtin_expect): New prototype.
5210 * builtins.c (build_builtin_expect_predicate): Add predictor
5211 argument, if non-NULL, create 3 argument __builtin_expect.
5212 (fold_builtin_expect): No longer static. Add ARG2 argument,
5213 pass it through to build_builtin_expect_predicate.
5214 (fold_builtin_2): Adjust caller.
5215 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5216 * internal-fn.def (BUILTIN_EXPECT): New.
5218 2014-03-18 Tobias Burnus <burnus@net-b.de>
5221 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5222 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5223 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5225 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5228 * predict.c (combine_predictions_for_bb): Fix up formatting.
5229 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5230 fill what it points to if non-NULL.
5231 (tree_predict_by_opcode): Adjust caller, use the predictor.
5232 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5234 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5236 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5237 proper constant for the store mode.
5239 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5241 * symtab.c (change_decl_assembler_name): Fix transparent alias
5244 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5246 * config/aarch64/aarch64.c: Correct the comments about the
5247 aarch64 stack layout.
5249 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5251 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5252 check for GF_OMP_FOR_KIND_FOR.
5254 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5256 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5257 ymm and zmm register names.
5259 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5262 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5263 note creation for the 2010-08-31 changes.
5265 2014-03-17 Marek Polacek <polacek@redhat.com>
5268 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5269 as -fno-tree-loop-vectorize.
5270 (expand_omp_simd): Likewise.
5272 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5274 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5275 (eligible_for_call_delay): New prototype.
5276 * config/sparc/sparc.c (tls_call_delay): Rename into...
5277 (eligible_for_call_delay): ...this. Return false if the instruction
5278 cannot be put in the delay slot of a branch.
5279 (eligible_for_restore_insn): Simplify.
5280 (eligible_for_return_delay): Return false if the instruction cannot be
5281 put in the delay slot of a branch and simplify.
5282 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5283 put in the delay slot of a branch.
5284 * config/sparc/sparc.md (fix_ut699): New attribute.
5285 (tls_call_delay): Delete.
5286 (in_call_delay): Reimplement.
5287 (eligible_for_sibcall_delay): Rename into...
5288 (in_sibcall_delay): ...this.
5289 (eligible_for_return_delay): Rename into...
5290 (in_return_delay): ...this.
5291 (in_branch_delay): Reimplement.
5292 (in_uncond_branch_delay): Delete.
5293 (in_annul_branch_delay): Delete.
5295 2014-03-14 Richard Henderson <rth@redhat.com>
5298 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5299 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5300 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5301 (floathi splitters): Remove.
5302 (float<SWI48x>xf2): New pattern.
5303 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5304 code that tried to handle DImode for 32-bit, but which was excluded
5305 by the pattern's condition. Drop allocation of stack temporary.
5306 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5307 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5308 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5309 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5310 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5311 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5312 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5313 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5314 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5315 (*float<SWI48x><X87MODEF>2_i387): Remove.
5316 (all float _with_temp splitters): Remove.
5317 (*float<SWI48x><MODEF>2_i387): New pattern.
5318 (*float<SWI48><MODEF>2_sse): New pattern.
5319 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5320 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5322 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5323 Marek Polacek <polacek@redhat.com>
5326 * common.opt (dump_base_name_prefixed): New Variable.
5327 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5328 if x_dump_base_name_prefixed is already set, set it at the end.
5330 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5332 PR rtl-optimization/60508
5333 * lra-constraints.c (get_reload_reg): Add new parameter
5335 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5336 Pass the new parameter values.
5338 2014-03-14 Richard Biener <rguenther@suse.de>
5340 * common.opt: Revert unintented changes from r205065.
5343 2014-03-14 Richard Biener <rguenther@suse.de>
5346 * cfghooks.c (split_block): Properly adjust all loops the
5347 block was a latch of.
5349 2014-03-14 Martin Jambor <mjambor@suse.cz>
5352 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5355 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5358 * config/avr/avr.c (avr_set_current_function): Pass function name
5359 through default_strip_name_encoding before sanity checking instead
5360 of skipping the first char of the assembler name.
5362 2014-03-13 Richard Henderson <rth@redhat.com>
5365 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5366 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5367 * config/i386/i386-protos.h: Likewise.
5368 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5369 in the expander instead of a splitter.
5370 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5371 any possibility of requiring a memory.
5372 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5373 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5374 (fp branch splitters): Update for ix86_split_fp_branch.
5375 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5376 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5377 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5378 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5379 (*fop_<MODEF>_3_i387): Likewise.
5380 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5381 (splitters for the fop_* register patterns): Remove.
5382 (fscalexf4_i387): Rename from *fscalexf4_i387.
5383 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5385 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5387 PR tree-optimization/59779
5388 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5389 type for bitsize and maxsize instead of HOST_WIDE_INT.
5391 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5393 PR rtl-optimization/57320
5394 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5395 the CFG after thread_prologue_and_epilogue_insns.
5397 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5399 PR rtl-optimization/57189
5400 * lra-constraints.c (process_alt_operands): Disfavor spilling
5403 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5405 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5407 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5409 PR tree-optimization/59025
5411 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5412 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5414 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5417 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5418 calls of avr_out_plus_1.
5420 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5422 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5423 BB's single pred and update the father loop's latch info later.
5425 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5427 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5431 (VEC_base): Likewise.
5432 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5433 registers, we need to swap double words in little endian mode.
5435 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5436 to be a container mode for 128-bit integer operations added in ISA
5437 2.07. Unlike TImode and PTImode, the preferred register set is
5438 the Altivec/VMX registers for the 128-bit operations.
5440 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5442 (rs6000_split_128bit_ok_p): Likewise.
5444 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5445 macros for creating ISA 2.07 normal and overloaded builtin
5446 functions with 3 arguments.
5447 (BU_P8V_OVERLOAD_3): Likewise.
5448 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5449 for use as overloaded functions.
5450 (VPERM_1TI_UNS): Likewise.
5451 (VSEL_1TI): Likewise.
5452 (VSEL_1TI_UNS): Likewise.
5453 (ST_INTERNAL_1ti): Likewise.
5454 (LD_INTERNAL_1ti): Likewise.
5455 (XXSEL_1TI): Likewise.
5456 (XXSEL_1TI_UNS): Likewise.
5457 (VPERM_1TI): Likewise.
5458 (VPERM_1TI_UNS): Likewise.
5459 (XXPERMDI_1TI): Likewise.
5460 (SET_1TI): Likewise.
5461 (LXVD2X_V1TI): Likewise.
5462 (STXVD2X_V1TI): Likewise.
5463 (VEC_INIT_V1TI): Likewise.
5464 (VEC_SET_V1TI): Likewise.
5465 (VEC_EXT_V1TI): Likewise.
5466 (EQV_V1TI): Likewise.
5467 (NAND_V1TI): Likewise.
5468 (ORC_V1TI): Likewise.
5469 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5470 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5472 (VADDUQM): Likewise.
5473 (VSUBCUQ): Likewise.
5474 (VADDEUQM): Likewise.
5475 (VADDECUQ): Likewise.
5476 (VSUBEUQM): Likewise.
5477 (VSUBECUQ): Likewise.
5479 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5480 __int128_t and __uint128_t types.
5481 (__uint128_type): Likewise.
5482 (altivec_categorize_keyword): Add support for vector __int128_t,
5483 vector __uint128_t, vector __int128, and vector unsigned __int128
5484 as a container type for TImode operations that need to be done in
5485 VSX/Altivec registers.
5486 (rs6000_macro_to_expand): Likewise.
5487 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5488 to support 128-bit integer instructions vaddcuq, vadduqm,
5489 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5490 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5492 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5493 for V1TImode, and set up preferences to use VSX/Altivec registers.
5494 Setup VSX reload handlers.
5495 (rs6000_debug_reg_global): Likewise.
5496 (rs6000_init_hard_regno_mode_ok): Likewise.
5497 (rs6000_preferred_simd_mode): Likewise.
5498 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5499 (easy_altivec_constant): Likewise.
5500 (output_vec_const_move): Likewise.
5501 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5503 (rs6000_expand_vector_extract): Likewise.
5504 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5506 (rs6000_const_vec): Add support for V1TImode.
5507 (rs6000_emit_le_vsx_load): Swap double words when loading or
5508 storing TImode/V1TImode.
5509 (rs6000_emit_le_vsx_store): Likewise.
5510 (rs6000_emit_le_vsx_move): Likewise.
5511 (rs6000_emit_move): Add support for V1TImode.
5512 (altivec_expand_ld_builtin): Likewise.
5513 (altivec_expand_st_builtin): Likewise.
5514 (altivec_expand_vec_init_builtin): Likewise.
5515 (altivec_expand_builtin): Likewise.
5516 (rs6000_init_builtins): Add support for V1TImode type. Add
5517 support for ISA 2.07 128-bit integer builtins. Define type names
5518 for the VSX/Altivec vector types.
5519 (altivec_init_builtins): Add support for overloaded vector
5520 functions with V1TImode type.
5521 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5522 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5524 (rs6000_split_128bit_ok_p): Likewise.
5525 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5526 __int128_t and vector __uint128_t.
5528 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5529 and mode attributes.
5536 (VS_scalar): Likewise.
5537 (VS_double): Likewise.
5538 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5540 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5541 we support the ISA 2.07 128-bit integer arithmetic instructions.
5542 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5543 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5544 and TImode types for use with the builtin functions.
5545 (V1TI_type_node): Likewise.
5546 (unsigned_V1TI_type_node): Likewise.
5547 (intTI_type_internal_node): Likewise.
5548 (uintTI_type_internal_node): Likewise.
5550 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5551 128-bit builtin functions.
5552 (UNSPEC_VADDEUQM): Likewise.
5553 (UNSPEC_VADDECUQ): Likewise.
5554 (UNSPEC_VSUBCUQ): Likewise.
5555 (UNSPEC_VSUBEUQM): Likewise.
5556 (UNSPEC_VSUBECUQ): Likewise.
5557 (VM): Add V1TImode to vector mode iterators.
5559 (VI_unit): Likewise.
5560 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5561 (altivec_vaddcuq): Likewise.
5562 (altivec_vsubuqm): Likewise.
5563 (altivec_vsubcuq): Likewise.
5564 (altivec_vaddeuqm): Likewise.
5565 (altivec_vaddecuq): Likewise.
5566 (altivec_vsubeuqm): Likewise.
5567 (altivec_vsubecuq): Likewise.
5569 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5571 (BOOL_128): Likewise.
5572 (BOOL_REGS_OUTPUT): Likewise.
5573 (BOOL_REGS_OP1): Likewise.
5574 (BOOL_REGS_OP2): Likewise.
5575 (BOOL_REGS_UNARY): Likewise.
5576 (BOOL_REGS_AND_CR0): Likewise.
5578 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5579 128-bit integer builtin support.
5580 (vec_vadduqm): Likewise.
5581 (vec_vaddecuq): Likewise.
5582 (vec_vaddeuqm): Likewise.
5583 (vec_vsubecuq): Likewise.
5584 (vec_vsubeuqm): Likewise.
5585 (vec_vsubcuq): Likewise.
5586 (vec_vsubuqm): Likewise.
5588 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5589 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5590 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5591 128-bit integer add/subtract to ISA 2.07.
5593 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5595 * config/arc/arc.c (arc_predicate_delay_insns):
5596 Fix third argument passed to conditionalize_nonjump.
5598 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5600 * config/aarch64/aarch64-builtins.c
5601 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5602 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5603 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5604 instead of __builtin_lfloor.
5605 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5607 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5609 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5610 (tree_ssa_ifcombine_bb_1): New function.
5611 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5612 is an empty forwarder block to then_bb or vice versa and then_bb
5613 and else_bb are effectively swapped.
5615 2014-03-12 Christian Bruel <christian.bruel@st.com>
5618 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5619 REG_CFA_DEF_CFA note.
5620 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5621 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5623 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5625 PR tree-optimization/60454
5626 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5628 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5630 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5631 Do not define target_cpu_default2 to generic.
5632 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5633 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5634 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5636 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5637 Marc Glisse <marc.glisse@inria.fr>
5639 PR tree-optimization/60502
5640 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5641 instead of build_low_bits_mask.
5643 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5646 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5647 if there are multiple uses, but op doesn't live on E edge.
5648 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5649 clobber stmts before __builtin_unreachable.
5651 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5653 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5654 hard_frame_pointer_rtx.
5655 * cse.c (cse_insn): Remove volatile check.
5656 * cselib.c (cselib_process_insn): Likewise.
5657 * dse.c (scan_insn): Likewise.
5659 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5661 * config/arc/arc.c (conditionalize_nonjump): New function,
5663 (arc_ifcvt): ... this.
5664 (arc_predicate_delay_insns): Use it.
5666 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5668 * config/arc/predicates.md (extend_operand): During/after reload,
5669 allow const_int_operand.
5670 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5671 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5672 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5674 (umulsi3_highpart_i): Likewise.
5676 2014-03-11 Richard Biener <rguenther@suse.de>
5678 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5679 Add asserts to guard possible wrong-code bugs.
5681 2014-03-11 Richard Biener <rguenther@suse.de>
5683 PR tree-optimization/60429
5684 PR tree-optimization/60485
5685 * tree-ssa-structalias.c (set_union_with_increment): Properly
5686 take into account all fields that overlap the shifted vars.
5687 (do_sd_constraint): Likewise.
5688 (do_ds_constraint): Likewise.
5689 (get_constraint_for_ptr_offset): Likewise.
5691 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5693 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5694 (nios2_compute_frame_layout):
5695 Add calculation of cfun->machine->fp_save_offset.
5696 (nios2_expand_prologue): Correct setting of frame pointer register
5698 (nios2_expand_epilogue): Update recovery of stack pointer from
5699 frame pointer accordingly.
5700 (nios2_initial_elimination_offset): Update calculation of offset
5701 for eliminating to HARD_FRAME_POINTER_REGNUM.
5703 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5706 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5707 cgraph_get_create_node on VAR_DECLs.
5709 2014-03-10 Richard Biener <rguenther@suse.de>
5712 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5714 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5716 * config/vms/vms.opt (vms_float_format): New variable.
5718 2014-03-08 Tobias Burnus <burnus@net-b.de>
5720 * doc/invoke.texi (-fcilkplus): Update implementation status.
5722 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5723 Richard Biener <rguenther@suse.de>
5725 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5726 consistently accross all TUs.
5727 (run_gcc): Enable -fshort-double automatically at link at link-time
5728 and disallow override.
5730 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5733 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5734 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5735 if they can't be used.
5737 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5739 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5740 for Solaris 11/x86 ld.
5741 * configure: Regenerate.
5743 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5745 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5746 (LIB_TLS_SPEC): Save as ld_tls_libs.
5747 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5748 (HAVE_AS_IX86_TLSLDM): New test.
5749 * configure, config.in: Regenerate.
5750 * config/i386/i386.c (legitimize_tls_address): Fall back to
5751 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5752 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5753 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5754 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5756 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5758 * common.opt (fira-loop-pressure): Mark as optimization.
5760 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5762 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5763 an OpenMP mappable type.
5765 2014-03-06 Matthias Klose <doko@ubuntu.com>
5767 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5768 MULTILIB_OSDIRNAMES is not defined.
5770 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5771 Meador Inge <meadori@codesourcery.com>
5774 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5775 (arm_legitimize_address): Call legitimize_tls_address for any
5776 arm_tls_referenced_p expression, handle constant addend. Call it
5777 before testing for !TARGET_ARM.
5778 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5780 2014-03-06 Richard Biener <rguenther@suse.de>
5786 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5788 * tree-streamer.c (record_common_node): Assert we don't record
5789 nodes with type double.
5790 (preload_common_node): Skip type double, complex double and double
5791 pointer since it is now frontend dependent due to fshort-double option.
5793 2014-03-06 Richard Biener <rguenther@suse.de>
5795 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5796 or -fno-lto is specified and the linker has full plugin support.
5797 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5798 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5799 * lto-wrapper.c (merge_and_complain): Merge compile-time
5800 optimization levels.
5801 (run_gcc): And pass it through to the link options.
5803 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5807 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5809 * cselib.c (remove_useless_values): Skip to avoid quadratic
5810 behavior if the condition moved from...
5811 (cselib_process_insn): ... here holds.
5813 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5816 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5817 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5820 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5821 (TM_H): Add x86-tune.def.
5823 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5825 * config/aarch64/aarch64.c (generic_tunings):
5826 Use cortexa57_extra_costs.
5828 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5831 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5832 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5833 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5836 2014-03-04 Heiher <r@hev.cc>
5838 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5839 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5841 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5843 * config/i386/predicates.md (const2356_operand): Change to ...
5844 (const2367_operand): ... this.
5845 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5847 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5848 (*avx512pf_scatterpf<mode>sf): Ditto.
5849 (avx512pf_scatterpf<mode>df): Ditto.
5850 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5851 (*avx512pf_scatterpf<mode>df): Ditto.
5852 * config/i386/i386.c (ix86_expand_builtin): Update
5853 incorrect hint operand error message.
5855 2014-03-04 Richard Biener <rguenther@suse.de>
5857 * lto-section-in.c (lto_get_section_data): Fix const cast.
5859 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5861 * tree-streamer.c (record_common_node): Assert we don't record
5862 nodes with type double.
5863 (preload_common_node): Skip type double, complex double and double
5864 pointer since it is now frontend dependent due to fshort-double option.
5866 2014-03-04 Richard Biener <rguenther@suse.de>
5869 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5870 (lto_input_toplevel_asms): Likewise.
5871 * lto-section-in.c (lto_get_section_data): Instead do it here
5874 2014-03-04 Richard Biener <rguenther@suse.de>
5876 PR tree-optimization/60382
5877 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5878 dead PHIs a reduction.
5880 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5882 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5884 (_mm_prefetch): Move out of GCC target("sse") pragma.
5885 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5886 GCC target("prfchw") pragma.
5887 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5889 * config/i386/i386.c (ix86_option_override_internal): Enable
5890 -mprfchw with -mprefetchwt1.
5892 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5894 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5897 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5899 * opts.h (CL_PCH_IGNORE): Define.
5900 * targhooks.c (option_affects_pch_p):
5901 Return false for options that have CL_PCH_IGNORE set.
5902 * opt-functions.awk: Process PchIgnore.
5903 * doc/options.texi: Document PchIgnore.
5905 * config/arc/arc.opt (misize): Add PchIgnore property.
5907 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5909 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5910 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5911 constraint on constants to permit them being loaded into
5912 GENERAL_REGS or BASE_REGS.
5914 2014-03-03 Nick Clifton <nickc@redhat.com>
5916 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5917 anti-cacnonical alternatives.
5918 (negandhi3_real): New pattern.
5919 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5921 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5923 * config/avr/avr-mcus.def: Remove atxmega16x1.
5924 * config/avr/avr-tables.opt: Regenerate.
5925 * config/avr/t-multilib: Regenerate.
5926 * doc/avr-mmcu.texi: Regenerate.
5928 2014-03-03 Tobias Grosser <tobias@grosser.es>
5929 Mircea Namolaru <mircea.namolaru@inria.fr>
5931 PR tree-optimization/58028
5932 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5935 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5937 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5938 not handled by recognizers.
5940 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5943 * function.c (expand_function_end): Don't emit
5944 clobber_return_register sequence if clobber_after is a BARRIER.
5945 * cfgexpand.c (construct_exit_block): Append instructions before
5946 return_label to prev_bb.
5948 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5950 * config/rs6000/constraints.md: Document reserved use of "wc".
5952 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5955 * ipa.c (function_and_variable_visibility): When dissolving comdat
5956 group, also set all symbols to local.
5958 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5963 2013-12-14 Jan Hubicka <jh@suse.cz>
5965 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5967 2014-03-02 Jon Beniston <jon@beniston.com>
5973 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5974 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5975 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5976 (simple_return, *simple_return): New patterns
5977 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5978 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5980 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5982 * dwarf2out.c (gen_subprogram_die): Tidy.
5984 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5987 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5988 (*mov_t_msb_neg_negc): ... this new insn.
5990 2014-02-28 Jason Merrill <jason@redhat.com>
5993 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5996 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5999 * dwarf2out.c (decltype_auto_die): New static.
6000 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6001 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6002 (is_cxx_auto): Likewise.
6004 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6006 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6007 we are not using general regs only.
6009 2014-02-28 Richard Biener <rguenther@suse.de>
6012 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6013 previous fix and only allow to remove trivial pre-headers
6014 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6015 (remove_forwarder_block): Properly update the latch of a loop.
6017 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6020 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6021 (cselib_preserved_hash_table): New.
6022 (preserve_constants_and_equivs): Move preserved vals to it.
6023 (cselib_find_slot): Look it up first.
6024 (cselib_init): Initialize it.
6025 (cselib_finish): Release it.
6026 (dump_cselib_table): Dump it.
6028 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6031 * cselib.c (remove_useless_values): Skip to avoid quadratic
6032 behavior if the condition moved from...
6033 (cselib_process_insn): ... here holds.
6035 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6038 * var-tracking.c (vt_initialize): Apply the same condition to
6039 preserve the CFA base value.
6041 2014-02-28 Joey Ye <joey.ye@arm.com>
6044 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6045 if reload in progress or completed.
6047 2014-02-28 Tobias Burnus <burnus@net-b.de>
6050 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6053 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6055 * doc/tm.texi.in (Condition Code Status): Update documention for
6056 relative locations of cc0-setter and cc0-user.
6058 2014-02-27 Jeff Law <law@redhat.com>
6060 PR rtl-optimization/52714
6061 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6062 into two independent simple sets, if I3 is a jump, ensure the
6063 pattern we place into I3 is a (set (pc) ...).
6065 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6066 Jeff Law <law@redhat.com>
6068 PR rtl-optimization/49847
6069 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6070 are in different blocks.
6071 * doc/tm.texi (Condition Code Status): Update documention for
6072 relative locations of cc0-setter and cc0-user.
6074 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6077 * lra.c (lra_emit_add): Check SUBREG too.
6079 2014-02-27 Andreas Schwab <schwab@suse.de>
6081 * config/m68k/m68k.c (m68k_option_override): Disable
6082 -flive-range-shrinkage for classic m68k.
6083 (m68k_override_options_after_change): Likewise.
6085 2014-02-27 Marek Polacek <polacek@redhat.com>
6088 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6089 -Wmaybe-uninitialized.
6091 2014-02-27 Alan Modra <amodra@gmail.com>
6094 * reload1.c (emit_input_reload_insns): When reload_override_in,
6095 set old to rl->in_reg when rl->in_reg is a subreg.
6097 2014-02-26 Richard Biener <rguenther@suse.de>
6100 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6102 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6104 * common/config/i386/predicates.md (const1256_operand): Remove.
6105 (const2356_operand): New.
6106 (const_1_to_2_operand): Remove.
6107 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6108 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6109 (*avx512pf_gatherpf<mode>sf): Ditto.
6110 (avx512pf_gatherpf<mode>df): Ditto.
6111 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6112 (*avx512pf_gatherpf<mode>df): Ditto.
6113 (avx512pf_scatterpf<mode>sf): Ditto.
6114 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6115 (*avx512pf_scatterpf<mode>sf): Ditto.
6116 (avx512pf_scatterpf<mode>df): Ditto.
6117 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6118 (*avx512pf_scatterpf<mode>df): Ditto.
6119 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6121 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6123 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6124 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6125 (_mm512_mask_testn_epi64_mask): Move to ...
6126 * config/i386/avx512cdintrin.h: Here.
6127 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6128 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6129 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6130 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6131 TARGET_AVX512F from TARGET_AVX512CD.
6133 2014-02-26 Richard Biener <rguenther@suse.de>
6136 * ipa.c (walk_polymorphic_call_targets): Properly guard
6137 call to inline_update_overall_summary.
6139 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6142 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6143 and latches only if requested. Fix latch if it is removed.
6144 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6145 LOOPS_HAVE_PREHEADERS.
6147 2014-02-25 Andrew Pinski <apinski@cavium.com>
6149 * builtins.c (expand_builtin_thread_pointer): Create a new target
6150 when the target is NULL.
6152 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6154 PR rtl-optimization/60317
6155 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6156 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6157 * lra-assigns.c: Include params.h.
6158 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6159 other reload pseudos considerations.
6161 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6163 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6164 to use canonical form for nor<mode>3.
6166 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6169 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6172 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6174 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6175 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6176 (ix86_handle_option): Handle OPT_mprefetchwt1.
6177 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6178 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6180 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6181 OPTION_MASK_ISA_PREFETCHWT1.
6182 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6183 (PTA_PREFETCHWT1): New.
6184 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6185 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6186 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6187 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6188 (*prefetch_avx512pf_<mode>_: Change into ...
6189 (*prefetch_prefetchwt1_<mode>: This.
6190 * config/i386/i386.opt (mprefetchwt1): New.
6191 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6192 (_mm_prefetch): Handle intent to write.
6193 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6195 2014-02-25 Richard Biener <rguenther@suse.de>
6198 * emit-rtl.c (mem_attrs_htab): Remove.
6199 (mem_attrs_htab_hash): Likewise.
6200 (mem_attrs_htab_eq): Likewise.
6201 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6202 (init_emit_once): Do not allocate mem_attrs_htab.
6204 2014-02-25 Richard Biener <rguenther@suse.de>
6207 * lto-opts.c (lto_write_options): Output non-explicit conservative
6208 -fwrapv, -fno-trapv and -fno-strict-overflow.
6209 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6210 (run_gcc): And pass them through.
6212 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6214 * sel-sched.c (calculate_new_fences): New parameter ptime.
6215 Calculate it as a maximum over all fence cycles.
6216 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6217 Print the final schedule timing when sched_verbose.
6219 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6221 PR rtl-optimization/60292
6222 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6223 bit fot the fence instruction.
6225 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6227 * calls.h: Fix typo in comment.
6229 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6231 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6232 adjusting offsetable addresses.
6234 2014-02-24 Guozhi Wei <carrot@google.com>
6236 * sparseset.h (sparseset_pop): Fix the wrong index.
6238 2014-02-24 Walter Lee <walt@tilera.com>
6240 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6241 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6243 * common/config/tilegx/tilegx-common.c
6244 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6245 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6247 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6248 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6249 (tilegx_gimplify_va_arg_expr): Handle big endian.
6250 (tilegx_expand_unaligned_load): Ditto.
6251 (tilegx_expand_unaligned_store): Ditto.
6252 (TARGET_RETURN_IN_MSB): New.
6253 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6254 (TARGET_ENDIAN_DEFAULT): New.
6255 (TARGET_BIG_ENDIAN): Handle big endian.
6256 (BYTES_BIG_ENDIAN): Ditto.
6257 (WORDS_BIG_ENDIAN): Ditto.
6258 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6261 * config/tilegx/tilegx.md (extv): Handle big endian.
6263 (insn_st<n>): Ditto.
6264 (insn_st<n>_add<bitsuffix>): Ditto.
6265 (insn_stnt<n>): Ditto.
6266 (insn_stnt<n>_add<bitsuffix>):Ditto.
6267 (vec_interleave_highv8qi): Handle big endian.
6268 (vec_interleave_highv8qi_be): New.
6269 (vec_interleave_highv8qi_le): New.
6270 (insn_v1int_h): Handle big endian.
6271 (vec_interleave_lowv8qi): Handle big endian.
6272 (vec_interleave_lowv8qi_be): New.
6273 (vec_interleave_lowv8qi_le): New.
6274 (insn_v1int_l): Handle big endian.
6275 (vec_interleave_highv4hi): Handle big endian.
6276 (vec_interleave_highv4hi_be): New.
6277 (vec_interleave_highv4hi_le): New.
6278 (insn_v2int_h): Handle big endian.
6279 (vec_interleave_lowv4hi): Handle big endian.
6280 (vec_interleave_lowv4hi_be): New.
6281 (vec_interleave_lowv4hi_le): New.
6282 (insn_v2int_l): Handle big endian.
6283 (vec_interleave_highv2si): Handle big endian.
6284 (vec_interleave_highv2si_be): New.
6285 (vec_interleave_highv2si_le): New.
6286 (insn_v4int_h): Handle big endian.
6287 (vec_interleave_lowv2si): Handle big endian.
6288 (vec_interleave_lowv2si_be): New.
6289 (vec_interleave_lowv2si_le): New.
6290 (insn_v4int_l): Handle big endian.
6291 * config/tilegx/tilegx.opt (mbig-endian): New option.
6292 (mlittle-endian): New option.
6293 * doc/install.texi: Document tilegxbe-linux.
6294 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6296 2014-02-24 Martin Jambor <mjambor@suse.cz>
6299 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6300 there are no parameter descriptors.
6302 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6304 PR rtl-optimization/60268
6305 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6306 initialization to ...
6307 (sched_rgn_init): ... here.
6308 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6310 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6312 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6315 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6317 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6320 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6322 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6323 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6325 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6327 * config/microblaze/predicates.md: Add cmp_op predicate.
6328 * config/microblaze/microblaze.md: Add branch_compare instruction
6329 which uses cmp_op predicate and emits cmp insn before branch.
6330 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6331 to microblaze_expand_conditional_branch and consolidate logic.
6332 (microblaze_expand_conditional_branch): emit branch_compare
6333 insn instead of handling cmp op separate from branch insn.
6335 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6337 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6340 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6342 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6343 define_insn with define_expand and new define_insn
6344 *altivec_lve<VI_char>x_internal.
6345 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6346 and new define_insn *altivec_stve<VI_char>x_internal.
6347 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6349 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6351 (altivec_expand_stvex_be): New function.
6353 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6355 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6356 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6357 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6358 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6360 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6363 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6364 instead of emit_move_insn.
6366 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6368 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6370 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6371 gen_altivec_vsumsws.
6373 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6375 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6376 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6377 (altivec_lvxl_<mode>): New define_expand incorporating
6378 -maltivec=be semantics where needed.
6379 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6380 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6381 semantics where needed.
6382 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6383 (altivec_stvx_<mode>): New define_expand incorporating
6384 -maltivec=be semantics where needed.
6385 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6386 VM2 iterator instead of V4SI.
6387 (altivec_stvxl_<mode>): New define_expand incorporating
6388 -maltivec=be semantics where needed.
6389 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6390 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6391 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6392 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6393 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6394 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6395 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6396 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6397 ALTIVEC_BUILTIN_STVXL.
6398 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6399 (altivec_expand_stvx_be): Likewise.
6400 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6401 (altivec_expand_lvx_be): Likewise.
6402 (altivec_expand_stvx_be): Likewise.
6403 (altivec_expand_builtin): Add cases for
6404 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6405 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6406 (altivec_init_builtins): Add definitions for
6407 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6408 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6410 2014-02-21 Catherine Moore <clm@codesourcery.com>
6412 * doc/invoke.texi (mvirt, mno-virt): Document.
6413 * config/mips/mips.opt (mvirt): New option.
6414 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6416 2014-02-21 Richard Biener <rguenther@suse.de>
6418 PR tree-optimization/60276
6419 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6420 (STMT_VINFO_MIN_NEG_DIST): New macro.
6421 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6422 STMT_VINFO_MIN_NEG_DIST.
6423 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6424 made for negative dependence distances still hold.
6426 2014-02-21 Richard Biener <rguenther@suse.de>
6429 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6430 DECL_INITIAL for globals not in the current function context.
6432 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6434 PR tree-optimization/56490
6435 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6436 * tree-ssa-uninit.c: Include params.h.
6437 (compute_control_dep_chain): Add num_calls argument, return false
6438 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6439 num_calls to recursive call.
6440 (find_predicates): Change dep_chain into normal array,
6441 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6442 variable and adjust compute_control_dep_chain caller.
6443 (find_def_preds): Likewise.
6445 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6447 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6448 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6450 2014-02-21 Nick Clifton <nickc@redhat.com>
6452 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6453 (pushhi1): Likewise.
6454 (popqi1): Add mode to pre_dec.
6457 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6459 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6460 mode for mask of V8SFmode permutation.
6462 2014-02-20 Richard Henderson <rth@redhat.com>
6465 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6466 a new pseudo for OLDVAL.
6468 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6471 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6472 gen_reg_rtx if d->testing_p.
6473 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6474 if d->testing_p and we will certainly return true.
6475 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6478 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6480 * emit-rtl.c (gen_reg_rtx): Assert that
6481 crtl->emit.regno_pointer_align_length is non-zero.
6483 2014-02-20 Richard Henderson <rth@redhat.com>
6486 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6487 on failure the store back into EXPECT.
6489 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6490 Sandra Loosemore <sandra@codesourcery.com>
6492 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6493 * config/nios2/nios2.c (nios2_function_profiler): Add
6494 -fPIC (flag_pic == 2) support.
6495 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6496 (nios2_large_offset_p): New function.
6497 (nios2_unspec_reloc_p): Move up position, update to use
6498 nios2_large_offset_p.
6499 (nios2_unspec_address): Remove function.
6500 (nios2_unspec_offset): New function.
6501 (nios2_large_got_address): New function.
6502 (nios2_got_address): Add large offset support.
6503 (nios2_legitimize_tls_address): Update usage of removed and new
6505 (nios2_symbol_binds_local_p): New function.
6506 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6507 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6508 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6509 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6510 processing for (const (unspec ...)).
6511 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6513 2014-02-20 Richard Biener <rguenther@suse.de>
6515 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6516 doing the substitution.
6517 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6519 2014-02-20 Martin Jambor <mjambor@suse.cz>
6522 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6523 info when checking whether lattices are bottom.
6525 2014-02-20 Richard Biener <rguenther@suse.de>
6528 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6531 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6534 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6535 parameter specifying the scaling.
6536 (inline_call): Update.
6537 (want_inline_recursively): Guard division by zero.
6538 (recursive_inlining): Update.
6539 * ipa-inline.h (clone_inlined_nodes): Update.
6541 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6544 * config/i386/i386.c (classify_argument): Pass structures of size
6545 64 bytes or less in register.
6547 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6548 Kirill Yukhin <kirill.yukhin@intel.com>
6550 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6551 (_mm_rcp28_round_ss): Ditto.
6552 (_mm_rsqrt28_round_sd): Ditto.
6553 (_mm_rsqrt28_round_ss): Ditto.
6554 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6555 (_mm_rcp14_round_ss): Ditto.
6556 (_mm_rsqrt14_round_sd): Ditto.
6557 (_mm_rsqrt14_round_ss): Ditto.
6558 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6559 the first input operand, get rid of match_dup.
6560 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6562 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6564 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6565 operand as the first input operand, set type attribute.
6566 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6568 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6569 operand as the first input operand, set type attribute.
6571 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6573 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6576 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6579 * config/i386/i386.c (construct_container): Remove TFmode check
6580 for X86_64_INTEGER_CLASS.
6582 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6585 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6586 only when -Wpsabi is enabled.
6588 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6591 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6592 passing arrays in registers are the same as for structs, so remove the
6593 special case for them.
6595 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6597 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6598 destination type, extract only the valid bits if the source type is not
6599 integral and has a different mode.
6601 2014-02-19 Richard Biener <rguenther@suse.de>
6604 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6607 2014-02-19 Richard Biener <rguenther@suse.de>
6610 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6611 (ipa_modify_call_arguments): Emit an argument load explicitely and
6612 preserve virtual SSA form there and for the replacement call.
6613 Do not update SSA form nor free dominance info.
6615 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6617 * ipa.c (function_and_variable_visibility): Also clear WEAK
6618 flag when disolving COMDAT_GROUP.
6620 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6622 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6623 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6625 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6626 do more sanity checks.
6627 (detect_type_change): Return true when giving up early.
6628 (compute_complex_assign_jump_func): Fix type parameter of
6629 ipa_set_ancestor_jf.
6630 (compute_complex_ancestor_jump_func): Likewise.
6631 (update_jump_functions_after_inlining): Fix updating of
6633 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6635 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6637 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6638 inline clones when edge disappears.
6640 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6643 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6644 Split 64-bit moves into 2 patterns. Do not allow the use of
6645 direct move for TDmode in little endian, since the decimal value
6646 has little endian bytes within a word, but the 64-bit pieces are
6647 ordered in a big endian fashion, and normal subreg's of TDmode are
6649 (mov<mode>_64bit_dm): Likewise.
6650 (movtd_64bit_nodm): Likewise.
6652 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6654 PR tree-optimization/60174
6655 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6656 statement of an SSA_NAME that occurs in an abnormal PHI node.
6658 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6661 * final.c (SEEN_BB): Remove.
6662 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6663 (final_scan_insn): Don't force_source_line on second
6664 NOTE_INSN_BASIC_BLOCK.
6666 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6669 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6670 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6671 (type_natural_mode): Warn ABI change when %zmm register is not
6672 available for AVX512F vector value passing.
6674 2014-02-18 Kai Tietz <ktietz@redhat.com>
6677 * config/i386/i386.c (ix86_expand_prologue): Use value in
6678 rax register as displacement when restoring %r10 or %rax.
6679 Fix wrong offset when restoring both registers.
6681 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6683 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6684 assertion with conditional return.
6686 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6687 Uros Bizjak <ubizjak@gmail.com>
6690 * config/i386/driver-i386.c (host_detect_local_cpu): If
6691 YMM state is not saved by the OS, also clear has_f16c. Move
6692 CPUID 0x80000001 handling before YMM state saving checking.
6694 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6696 PR rtl-optimization/58960
6697 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6698 factored out from ...
6699 (sched_init): ... here.
6700 (free_global_sched_pressure_data): New, factored out from ...
6701 (sched_finish): ... here.
6702 * sched-int.h (free_global_sched_pressure_data): Declare.
6703 * sched-rgn.c (nr_regions_initial): New static global.
6704 (haifa_find_rgns): Initialize it.
6705 (schedule_region): Disable sched-pressure for the newly
6708 2014-02-17 Richard Biener <rguenther@suse.de>
6710 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6711 release SSA defs of pattern stmts.
6713 2014-02-17 Richard Biener <rguenther@suse.de>
6715 * tree-inline.c (expand_call_inline): Release the virtual
6716 operand defined by the call we are about to inline.
6718 2014-02-17 Richard Biener <rguenther@suse.de>
6720 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6722 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6723 Ilya Tocar <ilya.tocar@intel.com>
6725 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6726 arguments order in builtin.
6727 (_mm512_permutexvar_epi64): Ditto.
6728 (_mm512_mask_permutexvar_epi64): Ditto
6729 (_mm512_maskz_permutexvar_epi32): Ditto
6730 (_mm512_permutexvar_epi32): Ditto
6731 (_mm512_mask_permutexvar_epi32): Ditto
6733 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6735 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6736 (p8_vmrgow): Likewise.
6738 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6740 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6743 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6746 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6747 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6748 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6749 using direct move instructions on ISA 2.07. Also adjust
6750 instruction length for 64-bit.
6751 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6752 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6754 2014-02-15 Alan Modra <amodra@gmail.com>
6758 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6759 find_replacement on parts of insn rtl that might be reloaded.
6761 2014-02-15 Richard Biener <rguenther@suse.de>
6763 PR tree-optimization/60183
6764 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6765 (tree_ssa_phiprop): Calculate and free post-dominators.
6767 2014-02-14 Jeff Law <law@redhat.com>
6769 PR rtl-optimization/60131
6770 * ree.c (get_extended_src_reg): New function.
6771 (combine_reaching_defs): Use it rather than assuming location of REG.
6772 (find_and_remove_re): Verify first operand of extension is
6773 a REG before adding the insns to the copy list.
6775 2014-02-14 Roland McGrath <mcgrathr@google.com>
6777 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6778 * configure: Regenerated.
6779 * config.in: Regenerated.
6780 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6781 instead of ASM_SHORT.
6783 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6784 Richard Earnshaw <rearnsha@arm.com>
6786 PR rtl-optimization/59535
6787 * lra-constraints.c (process_alt_operands): Encourage alternative
6788 when unassigned pseudo class is superset of the alternative class.
6789 (inherit_reload_reg): Don't inherit when optimizing for code size.
6790 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6791 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6792 modes not less than 4 for Thumb1.
6794 2014-02-14 Kyle McMartin <kyle@redhat.com>
6797 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6799 2014-02-14 Richard Biener <rguenther@suse.de>
6801 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6802 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6803 Do not strip INDIRECT_REFs.
6805 2014-02-14 Richard Biener <rguenther@suse.de>
6808 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6809 DECL_FUNCTION_SPECIFIC_TARGET.
6810 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6811 * tree-streamer-out.c (pack_ts_target_option): Remove.
6812 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6813 (write_ts_function_decl_tree_pointers): Do not stream
6814 DECL_FUNCTION_SPECIFIC_TARGET.
6815 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6816 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6817 (lto_input_ts_function_decl_tree_pointers): Do not stream
6818 DECL_FUNCTION_SPECIFIC_TARGET.
6820 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6822 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6823 (get_initial_def_for_induction, vectorizable_induction): Ignore
6824 debug stmts when looking for exit_phi.
6825 (vectorizable_live_operation): Fix up condition.
6827 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6829 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6830 nreverse() because it changes the content of original tree list.
6832 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6834 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6835 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6837 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6839 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6840 GNU coding standards.
6842 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6845 * dwarf2out.c (gen_subprogram_die): Don't call
6846 add_calling_convention_attribute if subr_die is old_die.
6848 2014-02-13 Sharad Singhai <singhai@google.com>
6850 * doc/optinfo.texi: Fix order of nodes.
6852 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6854 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6855 operands[2], not operands[3].
6857 2014-02-13 Richard Biener <rguenther@suse.de>
6860 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6861 mention the possibility of an in-tree build.
6862 (CLooG): Update recommended version to 0.18.1, mention the
6863 possibility of an in-tree build and clarify that the ISL
6864 bundled with CLooG does not work.
6866 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6869 * expr.c (compress_float_constant): If x is a hard register,
6870 extend into a pseudo and then move to x.
6872 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6874 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6875 caused by bad second argument to warning_at() with -mhotpatch and
6876 nested functions (e.g. with gfortran).
6878 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6880 * opts.c (option_name): Remove "enabled by default" rider.
6882 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6884 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6886 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6887 Uros Bizjak <ubizjak@gmail.com>
6890 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6891 * configure: Regenerated.
6893 2014-02-12 Richard Biener <rguenther@suse.de>
6895 * vec.c (vec_prefix::calculate_allocation): Move as
6896 inline variant to vec.h.
6897 (vec_prefix::calculate_allocation_1): New out-of-line version.
6898 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6899 (vec_prefix::m_has_auto_buf): Rename to ...
6900 (vec_prefix::m_using_auto_storage): ... this.
6901 (vec_prefix::calculate_allocation): Inline the easy cases
6902 and dispatch to calculate_allocation_1 which doesn't need the
6904 (va_heap::reserve): Use gcc_checking_assert.
6905 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6906 m_using_auto_storage.
6907 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6909 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6910 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6911 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6913 2014-02-12 Richard Biener <rguenther@suse.de>
6915 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6916 when we found a dependence.
6918 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6920 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6922 (maybe_fold_stmt): ... into this new function.
6923 * omp-low.c (lower_omp): Update comment.
6925 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6928 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6931 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6933 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6934 identifiers in comments.
6935 (cortexa53_extra_costs): Likewise.
6936 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6937 (cortexa7_extra_costs): Likewise.
6938 (cortexa12_extra_costs): Likewise.
6939 (cortexa15_extra_costs): Likewise.
6940 (v7m_extra_costs): Likewise.
6942 2014-02-12 Richard Biener <rguenther@suse.de>
6945 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6946 of posix_memalign being successful.
6947 (lower_stmt): Restrict lowering of posix_memalign to when
6948 -ftree-bit-ccp is enabled.
6950 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6952 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6954 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6956 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6958 PR rtl-optimization/60116
6959 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6960 other_insn once the combination has been validated.
6962 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6965 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6967 * ipa-devirt.c: Include demangle.h
6968 (odr_violation_reported): New static variable.
6969 (add_type_duplicate): Update odr_violations.
6970 (maybe_record_node): Add completep parameter; update it.
6971 (record_target_from_binfo): Add COMPLETEP parameter;
6972 update it as needed.
6973 (possible_polymorphic_call_targets_1): Likewise.
6974 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6975 rename FINAL to COMPLETE.
6976 (record_targets_from_bases): Sanity check we found the binfo;
6977 fix COMPLETEP updating.
6978 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6979 parameter, fix computing of COMPLETEP.
6980 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6981 at LTO time do demangling.
6982 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6983 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6985 (gimple_get_virt_method_for_binfo): Likewise.
6986 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6987 gimple_get_virt_method_for_vtable): Update prototypes.
6989 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6992 * genautomata.c (add_presence_absence): Fix typo with
6993 {final_}presence_list.
6995 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6998 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6999 for VSX/Altivec vectors that land in GPR registers.
7001 2014-02-11 Richard Henderson <rth@redhat.com>
7002 Jakub Jelinek <jakub@redhat.com>
7005 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7006 around drhs if type conversion to lacc->type is not useless.
7008 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7010 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7012 (cortex-a57.cortex-a53): Likewise.
7013 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7015 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7017 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7020 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7022 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7023 add_options_for_arm_vfp3.
7025 2014-02-11 Jeff Law <law@redhat.com>
7028 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7029 object with an undesirable mode.
7031 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7034 * config/i386/sol2-9.h: New file.
7035 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7036 *-*-solaris2.9*): Use it.
7038 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7040 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7041 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7043 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7045 * config/microblaze/microblaze.c: Extend mcpu version format
7047 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7049 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7051 2014-02-10 Richard Henderson <rth@redhat.com>
7054 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7055 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7056 ms-abi vs -mno-accumulate-outgoing-args.
7057 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7058 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7061 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7064 * cfgexpand.c (expand_asm_operands): Attach source location to
7065 ASM_INPUT rtx objects.
7066 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7068 2014-02-10 Nick Clifton <nickc@redhat.com>
7070 * config/mn10300/mn10300.c (popcount): New function.
7071 (mn10300_expand_prologue): Include saved registers in stack usage
7074 2014-02-10 Jeff Law <law@redhat.com>
7077 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7078 when changing the SET_DEST of a prior insn to avoid an input reload.
7080 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7082 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7083 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7084 -mcall-openbsd, or -mcall-linux.
7085 (CC1_ENDIAN_BIG_SPEC): Remove.
7086 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7087 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7088 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7089 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7090 and %cc1_endian_default.
7091 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7093 2014-02-10 Richard Biener <rguenther@suse.de>
7095 PR tree-optimization/60115
7096 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7097 MEM_REF handling. Properly verify that the accesses are not
7098 out of the objects bound.
7100 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7102 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7105 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7107 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7108 proper constants and fix formatting.
7109 (possible_polymorphic_call_targets): Fix formatting.
7111 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7112 Ilya Tocar <ilya.tocar@intel.com>
7114 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7115 (_mm512_loadu_epi32): Renamed into...
7116 (_mm512_loadu_si512): This.
7117 (_mm512_storeu_epi32): Renamed into...
7118 (_mm512_storeu_si512): This.
7119 (_mm512_maskz_ceil_ps): Removed.
7120 (_mm512_maskz_ceil_pd): Ditto.
7121 (_mm512_maskz_floor_ps): Ditto.
7122 (_mm512_maskz_floor_pd): Ditto.
7123 (_mm512_floor_round_ps): Ditto.
7124 (_mm512_floor_round_pd): Ditto.
7125 (_mm512_ceil_round_ps): Ditto.
7126 (_mm512_ceil_round_pd): Ditto.
7127 (_mm512_mask_floor_round_ps): Ditto.
7128 (_mm512_mask_floor_round_pd): Ditto.
7129 (_mm512_mask_ceil_round_ps): Ditto.
7130 (_mm512_mask_ceil_round_pd): Ditto.
7131 (_mm512_maskz_floor_round_ps): Ditto.
7132 (_mm512_maskz_floor_round_pd): Ditto.
7133 (_mm512_maskz_ceil_round_ps): Ditto.
7134 (_mm512_maskz_ceil_round_pd): Ditto.
7135 (_mm512_expand_pd): Ditto.
7136 (_mm512_expand_ps): Ditto.
7137 * config/i386/i386.c (ix86_builtins): Remove
7138 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7139 (bdesc_args): Ditto.
7140 * config/i386/predicates.md (const1256_operand): New.
7141 (const_1_to_2_operand): Ditto.
7142 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7143 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7144 (*avx512pf_gatherpf<mode>sf): Ditto.
7145 (avx512pf_gatherpf<mode>df): Ditto.
7146 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7147 (*avx512pf_gatherpf<mode>df): Ditto.
7148 (avx512pf_scatterpf<mode>sf): Ditto.
7149 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7150 (*avx512pf_scatterpf<mode>sf): Ditto.
7151 (avx512pf_scatterpf<mode>df): Ditto.
7152 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7153 (*avx512pf_scatterpf<mode>df): Ditto.
7154 (avx512f_expand<mode>): Removed.
7155 (<shift_insn><mode>3<mask_name>): Change predicate type.
7157 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7159 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7160 not at the end of datarefs vector use ordered_remove to avoid
7161 reordering datarefs vector.
7164 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7165 mark local addressable non-static vars as GOVD_PRIVATE
7166 instead of GOVD_LOCAL.
7167 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7168 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7172 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7173 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7174 assume_aligned or alloc_align attributes.
7175 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7176 arguments. Handle also assume_aligned and alloc_align attributes.
7177 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7178 calls to functions with assume_aligned or alloc_align attributes.
7179 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7181 2014-02-08 Terry Guo <terry.guo@arm.com>
7183 * doc/invoke.texi: Document ARM -march=armv7e-m.
7185 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7187 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7188 flag on __cilkrts_rethrow builtin.
7191 * ipa-cp.c (determine_versionability): Fail at -O0
7192 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7193 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7196 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7199 * tree-inline.c (copy_forbidden): Fail for
7200 __attribute__((optimize (0))) functions.
7202 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7204 * varpool.c: Include pointer-set.h.
7205 (varpool_remove_unreferenced_decls): Variables in other partitions
7206 will not be output; be however careful to not lose information
7209 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7211 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7212 lookup in the vtable constructor.
7214 2014-02-07 Jeff Law <law@redhat.com>
7217 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7218 define_insn_and_split.
7220 * ipa-inline.c (inline_small_functions): Fix typos.
7222 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7224 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7225 (s390_can_use_return_insn): Declare.
7226 * config/s390/s390.h (EPILOGUE_USES): Define.
7227 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7229 (s390_chunkify_start): Handle return JUMP_LABELs.
7230 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7231 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7232 (s390_can_use_return_insn): New functions.
7233 (s390_fix_long_loop_prediction): Handle conditional returns.
7234 (TARGET_SET_UP_BY_PROLOGUE): Define.
7235 * config/s390/s390.md (ANY_RETURN): New code iterator.
7236 (*creturn, *csimple_return, return, simple_return): New patterns.
7238 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7240 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7241 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7242 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7243 REG_CFA_RESTORE list when deciding not to restore a register.
7245 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7247 * config/s390/s390.c: Include tree-pass.h and context.h.
7248 (s390_early_mach): New function, split out from...
7249 (s390_emit_prologue): ...here.
7250 (pass_data_s390_early_mach): New pass structure.
7251 (pass_s390_early_mach): New class.
7252 (s390_option_override): Create and register early_mach pass.
7253 Move to end of file.
7255 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7257 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7258 to match for the exit block.
7260 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7262 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7263 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7264 Reject misaligned operands.
7266 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7268 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7270 2014-02-07 Richard Biener <rguenther@suse.de>
7273 * gimple-low.c (lower_builtin_posix_memalign): New function.
7274 (lower_stmt): Call it to lower posix_memalign in a way
7275 to make alignment info accessible.
7277 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7280 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7281 __builtin_setjmp_receiver.
7283 2014-02-07 Richard Biener <rguenther@suse.de>
7286 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7287 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7288 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7289 Handle BUILT_IN_POSIX_MEMALIGN.
7290 (find_func_clobbers): Likewise.
7291 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7292 (call_may_clobber_ref_p_1): Likewise.
7294 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7297 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7300 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7303 * lto-cgraph.c (lto_output_node): Use
7304 symtab_get_symbol_partitioning_class.
7305 (lto_output_varpool_node): likewise.
7306 (symtab_get_symbol_partitioning_class): Move here from
7308 * cgraph.h (symbol_partitioning_class): Likewise.
7309 (symtab_get_symbol_partitioning_class): Declare.
7311 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7313 * ggc.h (ggc_internal_cleared_alloc): New macro.
7314 * vec.h (vec_safe_copy): Handle memory stats.
7315 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7316 * target-globals.c (save_target_globals): Likewise.
7318 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7321 * expr.c (emit_move_resolve_push): Export; be bit more selective
7322 on when to clear alias set.
7323 * expr.h (emit_move_resolve_push): Declare.
7324 * function.h (struct function): Add tail_call_marked.
7325 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7326 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7327 * config/i386/i386.md (TImode move expander): De not call
7329 (FP push expanders): Preserve memory attributes.
7330 * config/i386/sse.md (push<mode>1): Remove.
7331 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7332 (ix86_expand_push): Remove.
7333 * config/i386/mmx.md (push<mode>1): Remove.
7335 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7337 PR rtl-optimization/60030
7338 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7339 lopart with paradoxical subreg before shifting it up by hprec.
7341 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7343 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7344 Remove extra newline at end of file.
7345 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7346 (arm_issue_rate): Handle cortexa57.
7347 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7348 (cortex-a57.cortex-a53): Likewise.
7350 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7353 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7354 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7356 (arm_expand_prologue): Adjust all callers.
7357 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7358 info, registers also at the lowest numbered registers side. Use
7359 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7363 * var-tracking.c (adjust_mems): Before adding a SET to
7364 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7366 2014-02-06 Alan Modra <amodra@gmail.com>
7369 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7370 change SDmode to DDmode when lra_in_progress.
7372 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7375 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7376 free_data_ref on the dr first, and before goto again also set dr
7377 to the next dr. For simd_lane_access, free old datarefs[i] before
7378 overwriting it. For get_vectype_for_scalar_type failure, don't
7379 free_data_ref if simd_lane_access.
7381 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7384 * tree.h (opts_for_fn): New inline function.
7385 (opt_for_fn): Define.
7386 * config/i386/i386.c (ix86_function_regparm): Use
7387 opt_for_fn (decl, optimize) instead of optimize.
7389 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7391 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7392 for SYMBOL_REF in large memory model.
7394 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7396 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7398 (cortex-a57): Likewise.
7399 (cortex-a57.cortex-a53): Likewise.
7401 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7402 Kugan Vivekanandarajah <kuganv@linaro.org>
7404 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7406 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7408 2014-02-06 Richard Biener <rguenther@suse.de>
7410 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7411 set_loop_copy and initialize_original_copy_tables.
7413 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7415 * config/aarch64/aarch64-simd.md
7416 (aarch64_ashr_simddi): Change QI to SI.
7418 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7419 Jakub Jelinek <jakub@redhat.com>
7422 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7425 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7427 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7428 CODE_FOR_altivec_vpku[hw]um to
7429 CODE_FOR_altivec_vpku[hw]um_direct.
7430 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7431 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7432 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7433 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7435 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7437 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7438 generation for -maltivec=be.
7439 (altivec_vsumsws): Simplify redundant test.
7441 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7443 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7444 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7445 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7446 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7447 gen_altivec_vpkuwum.
7448 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7450 (altivec_vpks<VI_char>ss): Likewise.
7451 (altivec_vpks<VI_char>us): Likewise.
7452 (altivec_vpku<VI_char>us): Likewise.
7453 (altivec_vpku<VI_char>um): Likewise.
7454 (altivec_vpku<VI_char>um_direct): New (copy of
7455 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7457 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7458 target is little endian and -maltivec=be is not specified.
7459 (*altivec_vupkhs<VU_char>_direct): New (copy of
7460 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7461 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7462 target is little endian and -maltivec=be is not specified.
7463 (*altivec_vupkls<VU_char>_direct): New (copy of
7464 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7465 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7466 little endian and -maltivec=be is not specified.
7467 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7468 little endian and -maltivec=be is not specified.
7470 2014-02-05 Richard Henderson <rth@redhat.com>
7473 * combine-stack-adj.c: Revert r206943.
7474 * sched-int.h (struct deps_desc): Add last_args_size.
7475 * sched-deps.c (init_deps): Initialize it.
7476 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7477 contain REG_ARGS_SIZE notes.
7479 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7481 * lto-cgraph.c (asm_nodes_output): Make global.
7482 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7483 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7484 (driver_handle_option): Handle OPT_fwpa.
7486 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7489 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7490 a comment typo and formatting issue. If odr_hash hasn't been
7491 created, return vNULL and set *completep to false.
7494 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7495 bb with no successors.
7497 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7500 * doc/invoke.texi (-march): Clarify documentation for ARM.
7504 2014-02-05 Richard Biener <rguenther@suse.de>
7506 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7507 when not vectorizing because of too many alias checks.
7508 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7509 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7511 2014-02-05 Nick Clifton <nickc@redhat.com>
7513 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7514 accept extended registers in any mode when compiling for the MN10300.
7516 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7518 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7519 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7520 sanitization attributes.
7521 (can_inline_edge_p): Likewise.
7522 (sanitize_attrs_match_for_inline_p): New function.
7524 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7526 * ipa-prop.c (detect_type_change): Shor circuit testing of
7527 type changes on THIS pointer.
7529 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7532 * config/pa/pa.c (legitimize_tls_address): Return original address
7533 if not passed a SYMBOL_REF rtx.
7534 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7536 (pa_emit_move_sequence): Simplify TLS source operands.
7537 (pa_legitimate_constant_p): Reject all TLS constants.
7538 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7539 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7541 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7543 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7544 groups when we know they are controlled by LTO.
7545 * varasm.c (default_binds_local_p_1): If object is in other partition,
7546 it will be resolved locally.
7548 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7550 * config/host-linux.c (linux_gt_pch_use_address): Don't
7551 use SSIZE_MAX because it is not always defined.
7553 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7556 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7557 threshold for pseudo splitting.
7558 (update_ebb_live_info): Process call argument hard registers and
7559 hard registers from insn definition too.
7560 (max_small_class_regs_num): New constant.
7561 (inherit_in_ebb): Update live hard regs through EBBs. Update
7562 reloads_num only for small register classes. Don't split for
7565 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7568 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7571 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7573 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7576 2014-02-04 Marek Polacek <polacek@redhat.com>
7578 * gdbinit.in (pel): Define.
7580 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7582 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7585 2014-02-04 Richard Biener <rguenther@suse.de>
7588 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7589 in function context local.
7590 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7591 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7592 similar to LTO_imported_decl_ref.
7594 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7596 PR tree-optimization/60002
7597 * cgraphclones.c (build_function_decl_skip_args): Clear
7600 PR tree-optimization/60023
7601 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7602 false to gsi_replace.
7603 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7604 has been in some EH region and vec_stmt could throw, add
7605 vec_stmt into the same EH region.
7606 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7607 has no lhs, ignore it.
7608 * internal-fn.c (expand_MASK_LOAD): Likewise.
7611 * tree-inline.c (copy_forbidden): Fail for
7612 __attribute__((optimize (0))) functions.
7615 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7616 is set, copy one less argument.
7617 (expand_simd_clones): Don't subtract clone_info->inbranch
7618 from simd_clone_struct_alloc argument.
7620 PR rtl-optimization/57915
7621 * recog.c (simplify_while_replacing): If all unary/binary/relational
7622 operation arguments are constant, attempt to simplify those.
7625 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7626 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7628 2014-02-04 Richard Biener <rguenther@suse.de>
7630 PR tree-optimization/60012
7631 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7632 TBAA disambiguation to all DDRs.
7634 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7637 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7638 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7640 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7643 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7645 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7647 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7648 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7650 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7653 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7654 to figure out targets of polymorphic calls with known decl.
7655 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7656 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7657 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7658 (get_polymorphic_call_info): ... here.
7659 (get_polymorphic_call_info_from_invariant): New function.
7661 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7663 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7664 lookup via vtable pointer; check for type consistency
7665 and turn inconsitent facts into UNREACHABLE.
7666 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7667 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7668 type inconsistent querries; return UNREACHABLE instead.
7670 2014-02-03 Richard Henderson <rth@twiddle.net>
7673 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7674 already processed this node.
7675 (normalize_one_pred_1): Pass along mark_set.
7676 (normalize_one_pred): Create and destroy a pointer_set_t.
7677 (normalize_one_pred_chain): Likewise.
7679 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7681 PR gcov-profile/58602
7682 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7684 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7687 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7688 -fno-devirtualize; try to devirtualize by the knowledge of
7689 virtual table pointer given by aggregate propagation.
7690 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7691 (ipa_print_node_jump_functions): Dump also offset that
7692 is relevant for polymorphic calls.
7693 (determine_known_aggregate_parts): Add arg_type parameter; use it
7694 instead of determining the type from pointer type.
7695 (ipa_compute_jump_functions_for_edge): Update call of
7696 determine_known_aggregate_parts.
7697 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7698 (gimple_get_virt_method_for_binfo): ... here; simplify using
7699 vtable_pointer_value_to_vtable.
7700 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7701 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7702 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7703 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7705 (vtable_pointer_value_to_binfo): ... here.
7706 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7708 2014-02-03 Teresa Johnson <tejohnson@google.com>
7710 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7711 redef of outer loop index variable.
7713 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7717 * doc/extend.texi (Function Attributes): Typo.
7719 2014-02-03 Cong Hou <congh@google.com>
7721 PR tree-optimization/60000
7722 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7723 if the vectorized statement is a store. A store statement can only
7724 appear at the end of pattern statements.
7726 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7728 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7729 (ix86_option_override_internal): Default long double to 64-bit for
7730 32-bit Bionic and to 128-bit for 64-bit Bionic.
7732 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7733 TARGET_LONG_DOUBLE_128 is true.
7734 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7736 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7737 (mlong-double-64): Negate -mlong-double-128.
7738 (mlong-double-128): New option.
7740 * config/i386/i386-c.c (ix86_target_macros): Define
7741 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7743 * doc/invoke.texi: Document -mlong-double-128.
7745 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7747 PR rtl-optimization/60024
7748 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7750 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7752 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7754 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7756 PR rtl-optimization/57662
7757 * sel-sched.c (code_motion_path_driver): Do not mark already not
7758 existing blocks in the visiting bitmap.
7760 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7762 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7763 on the insn being emitted.
7765 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7766 Will Deacon <will.deacon@arm.com>
7768 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7770 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7772 * config/arm/arm-tables.opt: Regenerate.
7774 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7776 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7777 for vector types other than V16QImode.
7778 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7779 define_expand, and call altivec_expand_vec_perm_le when producing
7780 code with little endian element order.
7781 (*altivec_vperm_<mode>_internal): New insn having previous
7782 behavior of altivec_vperm_<mode>.
7783 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7784 altivec_expand_vec_perm_le when producing code with little endian
7786 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7787 behavior of altivec_vperm_<mode>_uns.
7789 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7791 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7792 (altivec_vsumsws): Add handling for -maltivec=be with a little
7794 (altivec_vsumsws_direct): New.
7795 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7796 gen_altivec_vsumsws.
7798 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7800 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7801 vtable_pointer_value_to_binfo): New functions.
7802 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7803 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7805 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7807 * config/nios2/nios2.md (load_got_register): Initialize GOT
7808 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7809 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7811 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7813 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7814 preserverd by passthrough, do not propagate the type.
7816 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7818 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7819 (mips_atomic_assign_expand_fenv): New function.
7820 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7822 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7824 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7825 (__builtin_mips_set_fcsr): Likewise.
7826 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7827 MIPS_USI_FTYPE_VOID.
7828 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7829 (mips16_expand_set_fcsr): Likewise.
7830 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7831 (mips16_set_fcsr_stub): Likewise.
7832 (mips16_get_fcsr_one_only_stub): New class.
7833 (mips16_set_fcsr_one_only_stub): Likewise.
7834 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7835 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7836 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7837 (hard_float): New availability predicate.
7838 (mips_builtins): Add get_fcsr and set_fcsr.
7839 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7840 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7841 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7842 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7843 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7846 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7848 * config/mips/mips.c (mips_one_only_stub): New class.
7849 (mips_need_mips16_rdhwr_p): Replace with...
7850 (mips16_rdhwr_stub): ...this new variable.
7851 (mips16_stub_call_address): New function.
7852 (mips16_rdhwr_one_only_stub): New class.
7853 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7854 (mips_output_mips16_rdhwr): Delete.
7855 (mips_finish_stub): New function.
7856 (mips_code_end): Use it to handle rdhwr stubs.
7858 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7861 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7862 when calculating size of integer atomic types.
7864 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7866 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7868 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7870 PR tree-optimization/60003
7871 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7872 * profile.c (branch_prob): Use gimple_call_builtin_p
7873 to check for BUILT_IN_SETJMP_RECEIVER.
7874 * tree-inline.c (copy_bb): Call notice_special_calls.
7876 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7879 * lra-constraints.c (process_alt_operands): Update reload_sum only
7882 2014-01-31 Richard Henderson <rth@redhat.com>
7885 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7886 until after else_eh is processed.
7888 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7890 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7891 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7892 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7893 in smmintrin.h, remove them.
7894 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7895 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7896 * config/i386/i386.md (ROUND_SAE): Fix value.
7897 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7898 (const48_operand): New.
7899 * config/i386/subst.md (round), (round_expand): Use
7900 const_4_or_8_to_11_operand.
7901 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7903 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7905 * config/i386/constraints.md (Yk): Swap meaning with k.
7906 * config/i386/i386.md (movhi_internal): Change Yk to k.
7907 (movqi_internal): Ditto.
7908 (*k<logic><mode>): Ditto.
7911 (kandn<mode>): Ditto.
7912 (*<code>hi_1): Ditto.
7913 (*<code>qi_1): Ditto.
7914 (kxnor<mode>): Ditto.
7915 (kortestzhi): Ditto.
7916 (kortestchi): Ditto.
7918 (*one_cmplhi2_1): Ditto.
7919 (*one_cmplqi2_1): Ditto.
7920 * config/i386/sse.md (): Change k to Yk.
7921 (avx512f_load<mode>_mask): Ditto.
7922 (avx512f_blendm<mode>): Ditto.
7923 (avx512f_store<mode>_mask): Ditto.
7924 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7925 (avx512f_storedqu<mode>_mask): Ditto.
7926 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7928 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7929 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7930 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7931 (avx512f_maskcmp<mode>3): Ditto.
7932 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7933 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7934 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7935 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7936 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7937 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7938 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7939 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7940 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7941 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7942 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7943 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7944 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7945 (vec_extract_lo_<mode>_maskm): Ditto.
7946 (vec_extract_hi_<mode>_maskm): Ditto.
7947 (avx512f_vternlog<mode>_mask): Ditto.
7948 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7949 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7950 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7951 (avx512f_<code>v8div16qi2_mask): Ditto.
7952 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7953 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7954 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7955 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7956 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7957 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7958 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7959 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7960 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7961 (avx512cd_maskb_vec_dupv8di): Ditto.
7962 (avx512cd_maskw_vec_dupv16si): Ditto.
7963 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7964 (avx512f_vpermi2var<mode>3_mask): Ditto.
7965 (avx512f_vpermi2var<mode>3_mask): Ditto.
7966 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7967 (*avx512f_gathersi<mode>): Ditto.
7968 (*avx512f_gathersi<mode>_2): Ditto.
7969 (*avx512f_gatherdi<mode>): Ditto.
7970 (*avx512f_gatherdi<mode>_2): Ditto.
7971 (*avx512f_scattersi<mode>): Ditto.
7972 (*avx512f_scatterdi<mode>): Ditto.
7973 (avx512f_compress<mode>_mask): Ditto.
7974 (avx512f_compressstore<mode>_mask): Ditto.
7975 (avx512f_expand<mode>_mask): Ditto.
7976 * config/i386/subst.md (mask): Change k to Yk.
7977 (mask_scalar_merge): Ditto.
7980 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7982 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7984 2014-01-31 Richard Biener <rguenther@suse.de>
7987 * builtins.c (fold_builtin_memory_op): Make sure to not
7988 use a floating-point mode or a boolean or enumeral type for
7991 2014-01-30 DJ Delorie <dj@redhat.com>
7993 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7994 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7995 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7996 whenever main() has an epilogue.
7998 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8000 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8001 unused variable "field".
8002 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8003 (vsx_mergeh_<mode>): Likewise.
8004 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8005 (altivec_vmrghh): Likewise.
8006 (altivec_vmrghw): Likewise.
8007 (altivec_vmrglb): Likewise.
8008 (altivec_vmrglh): Likewise.
8009 (altivec_vmrglw): Likewise.
8010 (altivec_vspltb): Add missing uses.
8011 (altivec_vsplth): Likewise.
8012 (altivec_vspltw): Likewise.
8013 (altivec_vspltsf): Likewise.
8015 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8018 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8019 frame related instructions.
8021 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8023 PR rtl-optimization/59959
8024 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8025 any reload of register whose subreg is invalid.
8027 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8029 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8030 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8031 Add missing return type - void.
8033 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8035 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8036 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8037 remove element index adjustment for endian (now handled in vsx.md
8039 (altivec_expand_vec_perm_const): Use
8040 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8041 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8042 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8043 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8044 define_expand and a new define_insn *altivec_vspltb_internal;
8045 adjust for -maltivec=be on a little endian target.
8046 (altivec_vspltb_direct): New.
8047 (altivec_vsplth): Divide into a define_expand and a new
8048 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8049 little endian target.
8050 (altivec_vsplth_direct): New.
8051 (altivec_vspltw): Divide into a define_expand and a new
8052 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8053 little endian target.
8054 (altivec_vspltw_direct): New.
8055 (altivec_vspltsf): Divide into a define_expand and a new
8056 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8057 a little endian target.
8059 2014-01-30 Richard Biener <rguenther@suse.de>
8061 PR tree-optimization/59993
8062 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8063 can propagate form the earlier stmt and avoid the transform
8064 when the intermediate result is needed.
8066 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8068 * README.Portability: Fix typo.
8070 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8072 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8073 comparison_operator with ordered_comparison_operator.
8075 2014-01-30 Nick Clifton <nickc@redhat.com>
8077 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8078 Rename to mn10300_store_multiple_regs.
8079 * config/mn10300/mn10300.c: Likewise.
8080 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8081 store_multiple_regs.
8082 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8083 Call mn10300_store_multiple_regs.
8085 2014-01-30 Nick Clifton <nickc@redhat.com>
8086 DJ Delorie <dj@redhat.com>
8088 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8089 %fp 2 to keep registers after it properly word-aligned.
8090 (rl78_alloc_physical_registers_umul): Handle the case where both
8091 input operands are the same.
8093 2014-01-30 Richard Biener <rguenther@suse.de>
8095 PR tree-optimization/59903
8096 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8099 2014-01-30 Jason Merrill <jason@redhat.com>
8102 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8105 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8107 2014-01-30 Richard Biener <rguenther@suse.de>
8109 PR tree-optimization/59951
8110 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8112 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8115 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8116 SFmode to DFmode case.
8118 2014-01-29 DJ Delorie <dj@redhat.com>
8120 * config/msp430/msp430.opt (-minrt): New.
8121 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8123 (ENDFILE_SPEC): Likewise.
8125 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8127 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8128 (estimate_function_body_sizes): Use it.
8130 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8133 * dwarf2out.c (is_cxx_auto): New.
8134 (is_base_type): Use it.
8135 (gen_type_die_with_usage): Likewise.
8137 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8139 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8140 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8141 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8142 -maltivec=be with LE targets.
8143 (vsx_mergeh_<mode>): Likewise.
8144 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8145 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8146 (altivec_vmrghb): Replace with define_expand and new
8147 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8148 (altivec_vmrghb_direct): New define_insn.
8149 (altivec_vmrghh): Replace with define_expand and new
8150 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8151 (altivec_vmrghh_direct): New define_insn.
8152 (altivec_vmrghw): Replace with define_expand and new
8153 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8154 (altivec_vmrghw_direct): New define_insn.
8155 (*altivec_vmrghsf): Adjust for endianness.
8156 (altivec_vmrglb): Replace with define_expand and new
8157 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8158 (altivec_vmrglb_direct): New define_insn.
8159 (altivec_vmrglh): Replace with define_expand and new
8160 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8161 (altivec_vmrglh_direct): New define_insn.
8162 (altivec_vmrglw): Replace with define_expand and new
8163 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8164 (altivec_vmrglw_direct): New define_insn.
8165 (*altivec_vmrglsf): Adjust for endianness.
8166 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8167 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8168 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8169 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8170 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8171 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8172 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8173 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8175 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8177 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8178 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8181 2014-01-29 Richard Biener <rguenther@suse.de>
8183 PR tree-optimization/58742
8184 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8185 associate_pointerplus_align.
8186 (associate_pointerplus_diff): New function.
8187 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8188 and associate_pointerplus_diff.
8190 2014-01-29 Richard Biener <rguenther@suse.de>
8192 * lto-streamer.h (LTO_major_version): Bump to 3.
8193 (LTO_minor_version): Reset to 0.
8195 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8197 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8198 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8199 (arm_file_start): Generate correct asm header for armv7ve.
8200 * config/arm/bpabi.h: Add multilib support for armv7ve.
8201 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8202 and cortex-a15 to armv7ve.
8203 * config/arm/t-aprofile: Add multilib support for armv7ve.
8204 * doc/invoke.texi: Document -march=armv7ve.
8206 2014-01-29 Richard Biener <rguenther@suse.de>
8208 PR tree-optimization/58742
8209 * tree-ssa-forwprop.c (associate_plusminus): Return true
8210 if we changed sth, defer EH cleanup to ...
8211 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8212 (simplify_mult): New function.
8214 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8217 PR tree-optimization/59920
8218 * tree.c (build_common_builtin_nodes): Remove
8219 __builtin_setjmp_dispatcher initialization.
8220 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8221 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8222 instead of gsi_after_labels + manually skipping debug stmts.
8223 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8224 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8225 * tree-inline.c (copy_edges_for_bb): Remove
8226 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8227 argument. Ignore computed_goto_p stmts. Don't call
8228 make_abnormal_goto_edges. If a call might need abnormal edges
8229 for non-local gotos, see if it already has an edge to
8230 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8231 with true argument, don't do anything then, otherwise add
8232 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8233 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8235 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8236 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8237 (lower_stmt): Don't set data->calls_builtin_setjmp.
8238 (lower_builtin_setjmp): Adjust comment.
8239 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8240 * tree-cfg.c (found_computed_goto): Remove.
8241 (factor_computed_gotos): Remove.
8242 (make_goto_expr_edges): Return bool, true for computed gotos.
8243 Don't call make_abnormal_goto_edges.
8244 (build_gimple_cfg): Don't set found_computed_goto, don't call
8245 factor_computed_gotos.
8246 (computed_goto_p): No longer static.
8247 (make_blocks): Don't set found_computed_goto.
8248 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8249 (make_edges): If make_goto_expr_edges returns true, push bb
8250 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8251 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8252 vector. Record mapping between bbs and OpenMP regions if there
8253 are any, adjust make_gimple_omp_edges caller. Call
8254 handle_abnormal_edges.
8255 (make_abnormal_goto_edges): Remove.
8256 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8257 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8258 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8259 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8260 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8261 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8262 filling *region also set *region_idx to (*region)->entry->index.
8265 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8266 For REGs set ORIGINAL_REGNO.
8268 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8270 * doc/md.texi: Mention that a target shouldn't implement
8271 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8274 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8276 PR tree-optimization/59594
8277 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8278 a copy of the datarefs vector rather than the vector itself.
8280 2014-01-28 Jason Merrill <jason@redhat.com>
8283 * dwarf2out.c (auto_die): New static.
8284 (gen_type_die_with_usage): Handle C++1y 'auto'.
8285 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8288 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8291 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8292 (SPEC_X32): Likewise.
8293 (SPEC_64): Likewise.
8294 * config/i386/i386.c (ix86_option_override_internal): Turn off
8295 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8297 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8298 * config/i386/i386.h (TARGET_16BIT): New macro.
8299 (TARGET_16BIT_P): Likewise.
8300 * config/i386/i386.opt: Add m16.
8301 * doc/invoke.texi: Document -m16.
8303 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8305 PR preprocessor/59935
8306 * input.c (location_get_source_line): Bail out on when line number
8307 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8309 2014-01-28 Richard Biener <rguenther@suse.de>
8311 PR tree-optimization/58742
8312 * tree-ssa-forwprop.c (associate_plusminus): Handle
8313 pointer subtraction of the form (T)(P + A) - (T)P.
8315 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8317 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8320 2014-01-28 Richard Biener <rguenther@suse.de>
8323 2014-01-28 Richard Biener <rguenther@suse.de>
8325 PR rtl-optimization/45364
8326 PR rtl-optimization/59890
8327 * var-tracking.c (local_get_addr_clear_given_value): Handle
8328 already cleared slot.
8329 (val_reset): Handle not allocated local_get_addr_cache.
8330 (vt_find_locations): Use post-order on the inverted CFG.
8332 2014-01-28 Richard Biener <rguenther@suse.de>
8334 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8336 2014-01-28 Richard Biener <rguenther@suse.de>
8338 PR rtl-optimization/45364
8339 PR rtl-optimization/59890
8340 * var-tracking.c (local_get_addr_clear_given_value): Handle
8341 already cleared slot.
8342 (val_reset): Handle not allocated local_get_addr_cache.
8343 (vt_find_locations): Use post-order on the inverted CFG.
8345 2014-01-28 Alan Modra <amodra@gmail.com>
8347 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8348 * configure.ac <recursive call for build != host>: Define
8349 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8350 and LD_FOR_BUILD too.
8351 * configure: Regenerate.
8353 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8355 * config/i386/i386.c (get_builtin_code_for_version): Separate
8356 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8357 Broadwell from Haswell.
8359 2014-01-27 Steve Ellcey <sellcey@mips.com>
8361 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8362 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8363 * config/mips/mips.c (mips_option_override): Change setting
8365 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8366 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8367 Change from Mask to Var.
8369 2014-01-27 Jeff Law <law@redhat.com>
8371 * ipa-inline.c (inline_small_functions): Fix typo.
8373 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8375 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8376 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8377 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8378 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8379 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8380 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8381 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8382 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8383 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8384 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8385 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8386 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8387 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8388 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8389 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8390 (_mm512_storeu_epi64): Ditto.
8391 (_mm512_cmpge_epi32_mask): Ditto.
8392 (_mm512_cmpge_epu32_mask): Ditto.
8393 (_mm512_cmpge_epi64_mask): Ditto.
8394 (_mm512_cmpge_epu64_mask): Ditto.
8395 (_mm512_cmple_epi32_mask): Ditto.
8396 (_mm512_cmple_epu32_mask): Ditto.
8397 (_mm512_cmple_epi64_mask): Ditto.
8398 (_mm512_cmple_epu64_mask): Ditto.
8399 (_mm512_cmplt_epi32_mask): Ditto.
8400 (_mm512_cmplt_epu32_mask): Ditto.
8401 (_mm512_cmplt_epi64_mask): Ditto.
8402 (_mm512_cmplt_epu64_mask): Ditto.
8403 (_mm512_cmpneq_epi32_mask): Ditto.
8404 (_mm512_cmpneq_epu32_mask): Ditto.
8405 (_mm512_cmpneq_epi64_mask): Ditto.
8406 (_mm512_cmpneq_epu64_mask): Ditto.
8407 (_mm512_expand_pd): Ditto.
8408 (_mm512_expand_ps): Ditto.
8409 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8410 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8411 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8412 * config/i386/i386.c (ix86_builtins): Add
8413 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8414 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8415 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8416 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8417 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8418 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8419 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8420 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8421 IX86_BUILTIN_PMOVUSQW512_MEM.
8422 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8423 __builtin_ia32_pmovsqd512mem_mask,
8424 __builtin_ia32_pmovqd512mem_mask,
8425 __builtin_ia32_pmovusqw512mem_mask,
8426 __builtin_ia32_pmovsqw512mem_mask,
8427 __builtin_ia32_pmovqw512mem_mask,
8428 __builtin_ia32_pmovusdw512mem_mask,
8429 __builtin_ia32_pmovsdw512mem_mask,
8430 __builtin_ia32_pmovdw512mem_mask,
8431 __builtin_ia32_pmovqb512mem_mask,
8432 __builtin_ia32_pmovusqb512mem_mask,
8433 __builtin_ia32_pmovsqb512mem_mask,
8434 __builtin_ia32_pmovusdb512mem_mask,
8435 __builtin_ia32_pmovsdb512mem_mask,
8436 __builtin_ia32_pmovdb512mem_mask.
8437 (bdesc_args): Add __builtin_ia32_expanddf512,
8438 __builtin_ia32_expandsf512.
8439 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8440 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8441 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8442 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8443 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8444 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8445 (avx512f_<code>v8div16qi2_mask_store): This.
8446 (avx512f_expand<mode>): New.
8448 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8450 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8452 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8453 (_mm512_prefetch_i32scatter_pd): Ditto.
8454 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8455 (_mm512_prefetch_i64scatter_pd): Ditto.
8456 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8457 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8458 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8459 (_mm512_prefetch_i32scatter_ps): Ditto.
8460 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8461 (_mm512_prefetch_i64scatter_ps): Ditto.
8462 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8463 * config/i386/i386-builtin-types.def: Define
8464 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8465 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8466 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8467 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8468 IX86_BUILTIN_SCATTERPFQPD.
8469 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8470 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8471 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8472 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8473 __builtin_ia32_scatterpfqps.
8474 (ix86_expand_builtin): Expand new built-ins.
8475 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8476 fix memory access data type.
8477 (*avx512pf_gatherpf<mode>_mask): Ditto.
8478 (*avx512pf_gatherpf<mode>): Ditto.
8479 (avx512pf_scatterpf<mode>): Ditto.
8480 (*avx512pf_scatterpf<mode>_mask): Ditto.
8481 (*avx512pf_scatterpf<mode>): Ditto.
8482 (GATHER_SCATTER_SF_MEM_MODE): New.
8483 (avx512pf_gatherpf<mode>df): Ditto.
8484 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8485 (*avx512pf_scatterpf<mode>df): Ditto.
8487 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8490 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8491 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8494 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8496 * common/config/arm/arm-common.c
8497 (arm_rewrite_mcpu): Handle multiple names.
8499 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8501 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8503 * gimple-builder.h (create_gimple_tmp): Delete.
8505 2014-01-27 Christian Bruel <christian.bruel@st.com>
8507 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8510 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8512 * config/pa/pa.md (call): Generate indirect long calls to non-local
8513 functions when outputing 32-bit code.
8514 (call_value): Likewise except for special call to buggy powf function.
8516 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8517 portable runtime and PIC indirect calls.
8518 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8519 and PIC call sequences. Use ldo instead of blr to set return register
8520 in PIC call sequence.
8522 2014-01-25 Walter Lee <walt@tilera.com>
8524 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8525 avoid clobbering a live register.
8527 2014-01-25 Walter Lee <walt@tilera.com>
8529 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8530 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8531 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8532 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8534 2014-01-25 Walter Lee <walt@tilera.com>
8536 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8537 arguments on even registers.
8538 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8540 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8541 (BIGGEST_ALIGNMENT): Ditto.
8542 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8544 2014-01-25 Walter Lee <walt@tilera.com>
8546 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8547 insns before bundling.
8548 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8550 2014-01-25 Walter Lee <walt@tilera.com>
8552 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8553 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8554 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8556 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8558 * config/mips/constraints.md (kl): Delete.
8559 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8560 define expands, using...
8561 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8562 instructions for MIPS16.
8563 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8564 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8566 2014-01-25 Walter Lee <walt@tilera.com>
8568 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8572 2014-01-25 Walter Lee <walt@tilera.com>
8574 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8575 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8577 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8579 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8582 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8584 * print-rtl.c (in_call_function_usage): New var.
8585 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8586 EXPR_LIST mode as mode and not as reg note name.
8589 * cfgloopmanip.c (copy_loop_info): If
8590 loop->warned_aggressive_loop_optimizations, make sure
8591 the flag is set in target loop too.
8593 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8595 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8597 * builtins.def: Likewise.
8598 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8599 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8600 * ira.c (ira_setup_eliminable_regset): Likewise.
8601 * omp-low.c (gate_expand_omp): Likewise.
8602 (execute_lower_omp): Likewise.
8603 (diagnose_sb_0): Likewise.
8604 (gate_diagnose_omp_blocks): Likewise.
8605 (simd_clone_clauses_extract): Likewise.
8608 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8610 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8611 correction for little endian...
8612 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8615 2014-01-24 Jeff Law <law@redhat.com>
8617 PR tree-optimization/59919
8618 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8619 for non-returning calls.
8621 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8623 * common/config/aarch64/aarch64-common.c
8624 (aarch64_rewrite_mcpu): Handle multiple names.
8625 * config/aarch64/aarch64.h
8626 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8628 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8630 * input.c (add_file_to_cache_tab): Handle the case where fopen
8633 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8636 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8637 from push operand if code of push isn't PRE_DEC.
8639 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8642 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8643 -mquad-memory-atomic. Update -mquad-memory documentation to say
8644 it is only used for non-atomic loads/stores.
8646 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8647 -mquad-memory or -mquad-memory-atomic switches.
8649 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8650 -mquad-memory-atomic to ISA 2.07 support.
8652 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8653 to separate support of normal quad word memory operations (ldq, stq)
8654 from the atomic quad word memory operations.
8656 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8657 support to separate non-atomic quad word operations from atomic
8658 quad word operations. Disable non-atomic quad word operations in
8659 little endian mode so that we don't have to swap words after the
8660 load and before the store.
8661 (quad_load_store_p): Add comment about atomic quad word support.
8662 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8663 options printed with -mdebug=reg.
8665 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8666 -mquad-memory-atomic as the test for whether we have quad word
8667 atomic instructions.
8668 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8669 or -mp8-vector are used, allow byte/half-word atomic operations.
8671 * config/rs6000/sync.md (load_lockedti): Insure that the address
8672 is a proper indexed or indirect address for the lqarx instruction.
8673 On little endian systems, swap the hi/lo registers after the lqarx
8675 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8676 insure the address is valid for the lqarx instruction.
8677 (store_conditionalti): Insure that the address is a proper indexed
8678 or indirect address for the stqcrx. instruction. On little endian
8679 systems, swap the hi/lo registers before doing the stqcrx.
8681 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8682 insure the address is valid for the stqcrx. instruction.
8684 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8685 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8686 type of quad memory support is available.
8688 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8691 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8692 there is a danger of looping.
8694 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8696 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8697 force flag_ira_loop_pressure if set via command line.
8699 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8701 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8702 (ashr_simd): New builtin handling DI mode.
8703 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8704 (aarch64_sshr_simddi): New match pattern.
8705 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8706 (vshrd_n_s64): Likewise.
8707 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8709 2014-01-23 Nick Clifton <nickc@redhat.com>
8711 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8712 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8713 favour of mcu specific scripts.
8714 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8717 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8718 Alex Velenko <Alex.Velenko@arm.com>
8720 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8721 (vaddv_s16): Likewise.
8722 (vaddv_s32): Likewise.
8723 (vaddv_u8): Likewise.
8724 (vaddv_u16): Likewise.
8725 (vaddv_u32): Likewise.
8726 (vaddvq_s8): Likewise.
8727 (vaddvq_s16): Likewise.
8728 (vaddvq_s32): Likewise.
8729 (vaddvq_s64): Likewise.
8730 (vaddvq_u8): Likewise.
8731 (vaddvq_u16): Likewise.
8732 (vaddvq_u32): Likewise.
8733 (vaddvq_u64): Likewise.
8734 (vaddv_f32): Likewise.
8735 (vaddvq_f32): Likewise.
8736 (vaddvq_f64): Likewise.
8737 (vmaxv_f32): Likewise.
8738 (vmaxv_s8): Likewise.
8739 (vmaxv_s16): Likewise.
8740 (vmaxv_s32): Likewise.
8741 (vmaxv_u8): Likewise.
8742 (vmaxv_u16): Likewise.
8743 (vmaxv_u32): Likewise.
8744 (vmaxvq_f32): Likewise.
8745 (vmaxvq_f64): Likewise.
8746 (vmaxvq_s8): Likewise.
8747 (vmaxvq_s16): Likewise.
8748 (vmaxvq_s32): Likewise.
8749 (vmaxvq_u8): Likewise.
8750 (vmaxvq_u16): Likewise.
8751 (vmaxvq_u32): Likewise.
8752 (vmaxnmv_f32): Likewise.
8753 (vmaxnmvq_f32): Likewise.
8754 (vmaxnmvq_f64): Likewise.
8755 (vminv_f32): Likewise.
8756 (vminv_s8): Likewise.
8757 (vminv_s16): Likewise.
8758 (vminv_s32): Likewise.
8759 (vminv_u8): Likewise.
8760 (vminv_u16): Likewise.
8761 (vminv_u32): Likewise.
8762 (vminvq_f32): Likewise.
8763 (vminvq_f64): Likewise.
8764 (vminvq_s8): Likewise.
8765 (vminvq_s16): Likewise.
8766 (vminvq_s32): Likewise.
8767 (vminvq_u8): Likewise.
8768 (vminvq_u16): Likewise.
8769 (vminvq_u32): Likewise.
8770 (vminnmv_f32): Likewise.
8771 (vminnmvq_f32): Likewise.
8772 (vminnmvq_f64): Likewise.
8774 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8776 * config/aarch64/aarch64-simd.md
8777 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8778 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8779 (*aarch64_mul3_elt<mode>): Likewise.
8780 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8781 (*aarch64_mul3_elt_to_64v2df): Likewise.
8782 (*aarch64_mla_elt<mode>): Likewise.
8783 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8784 (*aarch64_mls_elt<mode>): Likewise.
8785 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8786 (*aarch64_fma4_elt<mode>): Likewise.
8787 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8788 (*aarch64_fma4_elt_to_64v2df): Likewise.
8789 (*aarch64_fnma4_elt<mode>): Likewise.
8790 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8791 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8792 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8793 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8794 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8795 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8796 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8797 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8798 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8800 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8802 * config/aarch64/aarch64-simd.md
8803 (aarch64_be_checked_get_lane<mode>): New define_expand.
8804 * config/aarch64/aarch64-simd-builtins.def
8805 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8806 New builtin definition.
8807 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8808 Use new safe be builtin.
8810 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8812 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8814 (aarch64_be_st1<mode>): Likewise.
8815 (aarch_ld1<VALL:mode>): Define_expand modified.
8816 (aarch_st1<VALL:mode>): Likewise.
8817 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8818 (UNSPEC_ST1): Likewise.
8820 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8822 * config/microblaze/microblaze.md: Add trap insn and attribute
8824 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8826 PR preprocessor/58580
8827 * input.h (location_get_source_line): Take an additional line_size
8829 (void diagnostics_file_cache_fini): Declare new function.
8830 * input.c (struct fcache): New type.
8831 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8832 New static constants.
8833 (diagnostic_file_cache_init, total_lines_num)
8834 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8835 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8836 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8837 (get_next_line, read_next_line, goto_next_line, read_line_num):
8838 New static function definitions.
8839 (diagnostic_file_cache_fini): New function.
8840 (location_get_source_line): Take an additional output line_len
8841 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8843 * diagnostic.c (diagnostic_finish): Call
8844 diagnostic_file_cache_fini.
8845 (adjust_line): Take an additional input parameter for the length
8846 of the line, rather than calculating it with strlen.
8847 (diagnostic_show_locus): Adjust the use of
8848 location_get_source_line and adjust_line with respect to their new
8849 signature. While displaying a line now, do not stop at the first
8850 null byte. Rather, display the zero byte as a space and keep
8851 going until we reach the size of the line.
8852 * Makefile.in: Add vec.o to OBJS-libcommon
8854 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8855 Ilya Tocar <ilya.tocar@intel.com>
8857 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8858 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8859 (__builtin_ia32_kmov16): Ditto.
8860 * config/i386/i386.md (UNSPEC_KMOV): New.
8863 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8865 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8866 (_mm512_storeu_si512): Ditto.
8868 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8871 * rtl.h (get_referenced_operands): Declare.
8872 * recog.c (get_referenced_operands): New function.
8873 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8874 operands have been referenced when recording LO_SUM references.
8876 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8878 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8880 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8882 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8883 Enable for generic and recent AMD targets.
8885 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8887 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8888 ARG_SIZE note when adjustment was eliminated.
8890 2014-01-22 Jeff Law <law@redhat.com>
8892 PR tree-optimization/59597
8893 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8894 in file. Accept new argument REGISTERING and use it to modify
8895 dump output appropriately.
8896 (register_jump_thread): Corresponding changes.
8897 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8898 thread paths involving joiner blocks. Add code to dump cancelled
8899 jump threading paths.
8901 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8903 PR rtl-optimization/59477
8904 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8905 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8907 2014-01-22 Tom Tromey <tromey@redhat.com>
8909 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8911 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8913 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8915 PR rtl-optimization/59896
8916 * lra-constraints.c (process_alt_operands): Check unused note for
8917 matched operands of insn with no output reloads.
8919 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8921 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8922 (mips_move_from_gpr_cost): Likewise.
8924 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8926 PR rtl-optimization/59858
8927 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8928 ira_class_hard_regs_num.
8929 (process_alt_operands): Increase reject for dying matched operand.
8931 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8934 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8935 smaller than size, perform several stores or loads and stores
8936 at dst + count - size to store or copy all of size bytes, rather
8937 than just last modesize bytes.
8939 2014-01-20 DJ Delorie <dj@redhat.com>
8941 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8942 that CLOBBERs are REGs before propogating their values.
8944 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8947 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8948 (cgraph_inline_failed_type): New function.
8949 * cgraph.h (DEFCIFCODE): Add type.
8950 (cgraph_inline_failed_type_t): New enum.
8951 (cgraph_inline_failed_type): New prototype.
8952 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8953 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8954 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8955 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8956 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8957 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8958 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8959 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8960 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8961 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8962 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8963 OPTIMIZATION_MISMATCH.
8964 * tree-inline.c (expand_call_inline): Emit errors during
8965 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8967 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8970 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8971 mode attribute in insn output.
8973 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8975 * output.h (output_constant): Delete.
8976 * varasm.c (output_constant): Make private.
8978 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8980 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8982 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8985 * tree.h (fold_builtin_strcat): New prototype.
8986 * builtins.c (fold_builtin_strcat): No longer static. Add len
8987 argument, if non-NULL, don't call c_strlen. Optimize
8988 directly into __builtin_memcpy instead of __builtin_strcpy.
8989 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8990 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8992 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8994 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8995 for SImode_address_operand operands, having only a REG argument.
8997 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8999 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9000 loader name using mbig-endian.
9001 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9003 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9005 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9009 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9011 * config/aarch64/aarch64-protos.h
9012 (aarch64_cannot_change_mode_class_ptr): Declare.
9013 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9014 aarch64_cannot_change_mode_class_ptr): New.
9015 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9016 backend hook aarch64_cannot_change_mode_class.
9018 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9020 * common/config/aarch64/aarch64-common.c
9021 (aarch64_handle_option): Don't handle any option order logic here.
9022 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9023 selected_cpu, warn on architecture version mismatch.
9024 (aarch64_override_options): Fix parsing order for option strings.
9026 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9027 Iain Sandoe <iain@codesourcery.com>
9030 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9031 warning. Amend comment to reflect current functionality.
9033 2014-01-20 Richard Biener <rguenther@suse.de>
9036 * builtins.c (fold_builtin_strcat): Remove case better handled
9037 by tree-ssa-strlen.c.
9039 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9041 * config/aarch64/aarch64.opt
9042 (mcpu, march, mtune): Make case-insensitive.
9044 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9047 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9048 if operands[1] is a REG or ZERO_EXTEND of a REG.
9050 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9052 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9054 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9056 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9057 long non-pic millicode calls.
9059 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9061 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9063 2014-01-19 Kito Cheng <kito@0xlab.org>
9065 * builtins.c (expand_movstr): Check movstr expand done or fail.
9067 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9068 H.J. Lu <hongjiu.lu@intel.com>
9071 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9072 to DImode for zero-extended addresses.
9074 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9076 PR rtl-optimization/57763
9077 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9078 on the new indirect jump_insn and increment LABEL_NUSES (label).
9080 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9084 * config.gcc (x86_archs): New variable.
9085 (x86_64_archs): Likewise.
9086 (x86_cpus): Likewise.
9087 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9088 --with-arch/--with-cpu= options.
9089 Support --with-arch=/--with-cpu={nehalem,westmere,
9090 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9092 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9094 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9095 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9097 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9099 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9101 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9104 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9105 clear cpp_get_options (parse_in)->warn_unused_macros for
9106 ix86_target_macros_internal with cpp_define.
9108 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9110 * jump.c (delete_related_insns): Keep (use (insn))s.
9111 * reorg.c (redundant_insn): Check for barriers too.
9113 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9115 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9117 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9119 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9120 call to $$dyncall when TARGET_LONG_CALLS is true.
9122 2014-01-17 Jeff Law <law@redhat.com>
9124 * ree.c (combine_set_extension): Temporarily disable test for
9125 changing number of hard registers.
9127 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9130 * ipa-inline-analysis.c (inline_free_summary):
9131 Do not free summary of aliases.
9133 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9136 * gimplify.c (gimplify_expr): Use create_tmp_var
9137 instead of create_tmp_var_raw. If cond doesn't have
9138 integral type, don't add the IFN_ANNOTATE builtin at all.
9140 2014-01-17 Martin Jambor <mjambor@suse.cz>
9143 * ipa-cp.c (prev_edge_clone): New variable.
9144 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9145 Also resize prev_edge_clone vector.
9146 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9147 (ipcp_edge_removal_hook): New function.
9148 (ipcp_driver): Register ipcp_edge_removal_hook.
9150 2014-01-17 Andrew Pinski <apinski@cavium.com>
9151 Steve Ellcey <sellcey@mips.com>
9154 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9157 2014-01-17 Jeff Law <law@redhat.com>
9160 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9161 so that pass_ccp runs first.
9163 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9165 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9166 (ix86_adjust_cost): Use !TARGET_XXX.
9167 (do_reorder_for_imul): Likewise.
9168 (swap_top_of_ready_list): Likewise.
9169 (ix86_sched_reorder): Likewise.
9171 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9173 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9174 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9175 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9176 (intel_memset): New. Duplicate slm_memset.
9177 (intel_cost): New. Duplicate slm_cost.
9178 (m_INTEL): New macro.
9179 (processor_target_table): Add "intel".
9180 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9181 with PROCESSOR_INTEL for "intel".
9182 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9183 PROCESSOR_SILVERMONT.
9184 (ix86_issue_rate): Likewise.
9185 (ix86_adjust_cost): Likewise.
9186 (ia32_multipass_dfa_lookahead): Likewise.
9187 (swap_top_of_ready_list): Likewise.
9188 (ix86_sched_reorder): Likewise.
9189 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9190 instead of TARGET_OPT_AGU.
9191 * config/i386/i386.h (TARGET_INTEL): New.
9192 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9193 (processor_type): Add PROCESSOR_INTEL.
9194 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9195 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9197 2014-01-17 Marek Polacek <polacek@redhat.com>
9200 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9203 2014-01-17 Richard Biener <rguenther@suse.de>
9205 PR tree-optimization/46590
9206 * opts.c (default_options_table): Add entries for
9207 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9208 all enabled at -O1 but not for -Og.
9209 * common.opt (fbranch-count-reg): Remove Init(1).
9210 (fmove-loop-invariants): Likewise.
9211 (ftree-pta): Likewise.
9213 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9215 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9216 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9217 decls to at least the GCC 4.8 used alignments.
9220 * tree-nested.c (convert_nonlocal_reference_stmt,
9221 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9222 of GIMPLE_BIND stmts, adjust associated decls.
9224 2014-01-17 Richard Biener <rguenther@suse.de>
9226 PR tree-optimization/46590
9227 * vec.h (vec<>::bseach): New member function implementing
9228 binary search according to C89 bsearch.
9229 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9230 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9231 bitmap pointer again. Make accesses_in_loop a flat array.
9232 (mem_ref_obstack): New global.
9233 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9234 (mark_ref_stored): Likewise.
9235 (ref_indep_loop_p_2): Likewise.
9236 (set_ref_stored_in_loop): New helper function.
9237 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9238 (memref_free): Adjust.
9239 (record_mem_ref_loc): Simplify.
9240 (gather_mem_refs_stmt): Adjust.
9241 (sort_locs_in_loop_postorder_cmp): New function.
9242 (analyze_memory_references): Sort accesses_in_loop after
9243 loop postorder number.
9244 (find_ref_loc_in_loop_cmp): New function.
9245 (for_all_locs_in_loop): Find relevant cluster of locs in
9246 accesses_in_loop and iterate without recursion.
9247 (execute_sm): Avoid uninit warning.
9248 (struct ref_always_accessed): Simplify.
9249 (ref_always_accessed::operator ()): Likewise.
9250 (ref_always_accessed_p): Likewise.
9251 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9252 loop postorder numbers here.
9253 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9256 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9259 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9260 on decls for which assemble_alias has been called.
9262 2014-01-17 Nick Clifton <nickc@redhat.com>
9264 * config/msp430/msp430.opt: (mcpu): New option.
9265 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9266 (msp430_option_override): Parse target_cpu. If the MCU name
9267 matches a generic string, clear target_mcu.
9268 (msp430_attr): Allow numeric interrupt values up to 63.
9269 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9270 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9272 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9274 * config/msp430/msp430.md (popm): Use %J rather than %I.
9275 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9276 (addhi_cy_i): Use immediate_operand for operand 2.
9277 * doc/invoke.texi: Document -mcpu option.
9279 2014-01-17 Richard Biener <rguenther@suse.de>
9281 PR rtl-optimization/38518
9282 * df.h (df_analyze_loop): Declare.
9283 * df-core.c: Include cfgloop.h.
9284 (df_analyze_1): Split out main part of df_analyze.
9285 (df_analyze): Adjust.
9286 (loop_inverted_post_order_compute): New function.
9287 (loop_post_order_compute): Likewise.
9288 (df_analyze_loop): New function avoiding whole-function
9290 * loop-invariant.c (find_defs): Use df_analyze_loop.
9291 (find_invariants): Adjust.
9292 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9294 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9296 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9297 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9299 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9301 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9302 traversal when removing references.
9304 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9307 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9309 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9312 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9313 pushing a reload for an autoinc when we had previously reloaded an
9314 inner part of the address.
9316 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9318 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9320 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9321 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9322 when not giving up or versioning for alias only because of
9324 (vect_analyze_data_ref_dependences): Set to true.
9325 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9327 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9328 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9332 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9335 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9336 operand 0 predicate for gathers, use a new pseudo as subtarget.
9338 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9341 * lra-constraints.c (process_alt_operands): Add printing debug info.
9342 Check absence of input/output reloads for matched operands too.
9344 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9346 PR rtl-optimization/59835
9347 * ira.c (ira_init_register_move_cost): Increase cost for
9350 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9352 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9354 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9357 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9358 non-register objects. Use gen_(high/low)part more consistently.
9361 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9364 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9365 endian support, remove tests for WORDS_BIG_ENDIAN.
9366 (p8_mfvsrd_3_<mode>): Likewise.
9367 (reload_gpr_from_vsx<mode>): Likewise.
9368 (reload_gpr_from_vsxsf): Likewise.
9369 (p8_mfvsrd_4_disf): Likewise.
9371 2014-01-16 Richard Biener <rguenther@suse.de>
9373 PR rtl-optimization/46590
9374 * lcm.c (compute_antinout_edge): Use postorder iteration.
9375 (compute_laterin): Use inverted postorder iteration.
9377 2014-01-16 Nick Clifton <nickc@redhat.com>
9380 * varasm.c (output_constant): Return the number of bytes actually
9382 (output_constructor_array_range): Update the field size with the
9383 number of bytes emitted by output_constant.
9384 (output_constructor_regular_field): Likewise. Also do not
9385 complain if the total number of bytes emitted is now greater
9386 than the expected fieldpos.
9387 * output.h (output_constant): Update prototype and descriptive comment.
9389 2014-01-16 Marek Polacek <polacek@redhat.com>
9392 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9393 it is error_mark_node.
9395 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9397 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9398 VALID_AVX256_REG_OR_OI_MODE.
9400 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9402 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9403 current procedure should be profiled.
9405 2014-01-15 Andrew Pinski <apinski@cavium.com>
9407 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9408 of moving from/to the STACK_REG register class.
9410 2014-01-15 Richard Henderson <rth@redhat.com>
9413 * reginfo.c (global_regs_decl): Globalize.
9414 * rtl.h (global_regs_decl): Declare.
9415 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9416 reserved via global_regs.
9418 2014-01-15 Teresa Johnson <tejohnson@google.com>
9420 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9422 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9424 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9425 and vmulosh rather than call gen_vec_widen_smult_*.
9426 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9427 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9428 (vec_widen_smult_even_v16qi): Likewise.
9429 (vec_widen_umult_even_v8hi): Likewise.
9430 (vec_widen_smult_even_v8hi): Likewise.
9431 (vec_widen_umult_odd_v16qi): Likewise.
9432 (vec_widen_smult_odd_v16qi): Likewise.
9433 (vec_widen_umult_odd_v8hi): Likewise.
9434 (vec_widen_smult_odd_v8hi): Likewise.
9435 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9436 vmuloub rather than call gen_vec_widen_umult_*.
9437 (vec_widen_umult_lo_v16qi): Likewise.
9438 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9439 vmulosb rather than call gen_vec_widen_smult_*.
9440 (vec_widen_smult_lo_v16qi): Likewise.
9441 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9442 rather than call gen_vec_widen_umult_*.
9443 (vec_widen_umult_lo_v8hi): Likewise.
9444 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9445 rather than call gen_vec_widen_smult_*.
9446 (vec_widen_smult_lo_v8hi): Likewise.
9448 2014-01-15 Jeff Law <law@redhat.com>
9450 PR tree-optimization/59747
9451 * ree.c (find_and_remove_re): Properly handle case where a second
9452 eliminated extension requires widening a copy created for elimination
9453 of a prior extension.
9454 (combine_set_extension): Ensure that the number of hard regs needed
9455 for a destination register does not change when we widen it.
9457 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9459 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9460 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9461 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9462 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9463 (avr-*-rtems*): Likewise.
9464 (bfin*-rtems*): Likewise.
9465 (moxie-*-rtems*): Likewise.
9466 (h8300-*-rtems*): Likewise.
9467 (i[34567]86-*-rtems*): Likewise.
9468 (lm32-*-rtems*): Likewise.
9469 (m32r-*-rtems*): Likewise.
9470 (m68k-*-rtems*): Likewise.
9471 (microblaze*-*-rtems*): Likewise.
9472 (mips*-*-rtems*): Likewise.
9473 (powerpc-*-rtems*): Likewise.
9474 (sh-*-rtems*): Likewise.
9475 (sparc-*-rtems*): Likewise.
9476 (sparc64-*-rtems*): Likewise.
9477 (v850-*-rtems*): Likewise.
9478 (m32c-*-rtems*): Likewise.
9480 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9482 PR rtl-optimization/59511
9483 * ira.c (ira_init_register_move_cost): Use memory costs for some
9484 cases of register move cost calculations.
9485 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9486 instead of BB frequency.
9487 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9488 * lra-assigns.c (find_hard_regno_for): Ditto.
9490 2014-01-15 Richard Biener <rguenther@suse.de>
9492 PR tree-optimization/59822
9493 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9494 (vectorizable_load): Use it to hoist defs of uses of invariant
9495 loads out of the loop.
9497 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9498 Kugan Vivekanandarajah <kuganv@linaro.org>
9501 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9504 2014-01-15 Richard Biener <rguenther@suse.de>
9506 PR rtl-optimization/59802
9507 * lcm.c (compute_available): Use inverted postorder to seed
9508 the initial worklist.
9510 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9513 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9514 ADDR_REGS for invalid symrefs in non-PIC code.
9516 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9519 * builtins.c (determine_block_size): Initialize *probable_max_size
9520 even if len_rtx is CONST_INT.
9522 2014-01-14 Andrew Pinski <apinski@cavium.com>
9524 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9525 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9526 (cortexa53_tunings): Likewise.
9527 (aarch64_sched_issue_rate): New function.
9528 (TARGET_SCHED_ISSUE_RATE): Define.
9530 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9532 * ira-costs.c (find_costs_and_classes): Add missed
9533 ira_init_register_move_cost_if_necessary.
9535 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9538 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9540 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9543 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9544 to indicate if type is used for function return value. Warn ABI
9545 change if the vector mode isn't available for function return value.
9546 (ix86_function_arg_advance): Pass false to type_natural_mode.
9547 (ix86_function_arg): Likewise.
9548 (ix86_gimplify_va_arg): Likewise.
9549 (function_arg_32): Don't warn ABI change.
9550 (ix86_function_value): Pass true to type_natural_mode.
9551 (ix86_return_in_memory): Likewise.
9552 (ix86_struct_value_rtx): Removed.
9553 (TARGET_STRUCT_VALUE_RTX): Likewise.
9555 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9557 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9558 converting a conditional jump into a conditional return.
9560 2014-01-14 Richard Biener <rguenther@suse.de>
9562 PR tree-optimization/58921
9563 PR tree-optimization/59006
9564 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9565 hoisting invariant stmts.
9566 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9567 invariant loads on the preheader edge if possible.
9569 2014-01-14 Joey Ye <joey.ye@arm.com>
9571 * doc/plugin.texi (Building GCC plugins): Update to C++.
9573 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9575 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9576 (_mm_rcp28_round_ss): Ditto.
9577 (_mm_rsqrt28_round_sd): Ditto.
9578 (_mm_rsqrt28_round_ss): Ditto.
9579 (_mm_rcp28_sd): Ditto.
9580 (_mm_rcp28_ss): Ditto.
9581 (_mm_rsqrt28_sd): Ditto.
9582 (_mm_rsqrt28_ss): Ditto.
9583 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9584 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9585 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9586 (IX86_BUILTIN_RCP28SD): Ditto.
9587 (IX86_BUILTIN_RCP28SS): Ditto.
9588 (IX86_BUILTIN_RSQRT28SD): Ditto.
9589 (IX86_BUILTIN_RSQRT28SS): Ditto.
9590 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9591 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9592 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9593 (ix86_expand_special_args_builtin): Expand new FTYPE.
9594 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9595 (srcp14<mode>): Make insn unary.
9596 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9597 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9598 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9599 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9600 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9601 Fix rounding: make it SAE only.
9602 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9604 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9606 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9607 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9608 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9609 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9610 (round_saeonly_mask_scalar_operand4): Ditto.
9611 (round_saeonly_mask_scalar_op3): Ditto.
9612 (round_saeonly_mask_scalar_op4): Ditto.
9614 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9616 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9617 Implement -maltivec=be for vec_insert and vec_extract.
9619 2014-01-10 DJ Delorie <dj@redhat.com>
9621 * config/msp430/msp430.md (call_internal): Don't allow memory
9622 references with SP as the base register.
9623 (call_value_internal): Likewise.
9624 * config/msp430/constraints.md (Yc): New. For memory references
9625 that don't use SP as a base register.
9627 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9628 "an integer without a # prefix"
9629 * config/msp430/msp430.md (epilogue_helper): Use it.
9631 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9634 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9635 AVX512F gather builtins.
9636 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9637 on gather decls with INTEGER_TYPE masktype.
9638 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9639 directly into the builtin rather than hoisting it before loop.
9641 PR tree-optimization/59387
9642 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9643 (scev_const_prop): If folded_casts and type has undefined overflow,
9644 use force_gimple_operand instead of force_gimple_operand_gsi and
9645 for each added stmt if it is assign with
9646 arith_code_with_undefined_signed_overflow, call
9647 rewrite_to_defined_overflow.
9648 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9649 gimple-fold.h instead.
9650 (arith_code_with_undefined_signed_overflow,
9651 rewrite_to_defined_overflow): Moved to ...
9652 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9653 rewrite_to_defined_overflow): ... here. No longer static.
9654 Include gimplify-me.h.
9655 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9656 rewrite_to_defined_overflow): New prototypes.
9658 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9660 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9662 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9664 * builtins.c (get_object_alignment_2): Minor tweak.
9665 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9667 2014-01-13 Christian Bruel <christian.bruel@st.com>
9669 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9670 optimized non constant lengths.
9672 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9675 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9676 load as __atomic_load_N if possible.
9678 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9680 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9682 (rs6000_expand_builtin): Adjust call.
9684 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9687 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9688 * config/rs6000/rs6000.c: Include target-globals.h.
9689 (rs6000_set_current_function): Instead of doing target_reinit
9690 unconditionally, use save_target_globals_default_opts and
9691 restore_target_globals.
9693 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9695 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9696 (rs6000_expand_builtin): Handle mffs and mtfsf.
9697 (rs6000_init_builtins): Define mffs and mtfsf.
9698 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9699 (rs6000_mffs): New pattern.
9700 (rs6000_mtfsf): New pattern.
9702 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9704 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9705 Start narrowing with START. Apply candidate-use pair
9706 and check overall cost in narrowing.
9707 (iv_ca_prune): Pass new argument.
9709 2014-01-10 Jeff Law <law@redhat.com>
9712 * ree.c (combine_reaching_defs): Ensure the defining statement
9713 occurs before the extension when optimizing extensions with
9714 different source and destination hard registers.
9716 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9719 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9720 vtables into the type inheritance graph.
9722 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9724 PR rtl-optimization/59754
9725 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9726 modes in the REGNO != REGNO case.
9728 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9730 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9732 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9734 PR tree-optimization/59745
9735 * tree-predcom.c (tree_predictive_commoning_loop): Call
9736 free_affine_expand_cache if giving up because components is NULL.
9738 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9739 GC in payload of target_globals struct instead of allocating them on
9740 the heap and the larger structs separately using GC.
9741 * target-globals.h (struct target_globals): Make regs, hard_regs,
9742 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9743 of GTY((skip)) and change type to void *.
9744 (reset_target_globals): Cast loads from those fields to corresponding
9747 2014-01-10 Steve Ellcey <sellcey@mips.com>
9750 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9751 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9752 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9754 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9757 * aarch64-modes.def (CC_Zmode): New flags mode.
9758 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9759 represents an equality.
9760 (aarch64_get_condition_code): Handle CC_Zmode.
9761 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9763 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9765 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9766 extraction in good case.
9768 2014-01-10 Richard Biener <rguenther@suse.de>
9770 PR tree-optimization/59374
9771 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9772 checking after SLP discovery. Mark stmts not participating
9773 in any SLP instance properly.
9775 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9777 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9778 when handling a SET rtx.
9780 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9782 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9783 (cortex-a57): Likewise.
9784 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9786 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9788 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9789 non-iwmmxt builtins.
9791 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9795 * ipa-devirt.c record_target_from_binfo): Take as argument
9796 stack of binfos and lookup matching one for virtual inheritance.
9797 (possible_polymorphic_call_targets_1): Update.
9799 2014-01-10 Huacai Chen <chenhc@lemote.com>
9801 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9802 kernel strings for Loongson-2E/2F/3A.
9804 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9807 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9808 is_gimple_call before calling gimple_call_internal_p.
9810 2014-01-09 Steve Ellcey <sellcey@mips.com>
9812 * Makefile.in (TREE_FLOW_H): Remove.
9813 (TREE_SSA_H): Add file names from tree-flow.h.
9814 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9815 * tree.h: Remove tree-flow.h reference.
9816 * hash-table.h: Remove tree-flow.h reference.
9817 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9818 reference with tree-ssa-loop.h.
9820 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9822 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9823 default element-order behavior for -maltivec.
9824 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9825 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9826 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9827 when targeting big endian, at least for now.
9828 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9830 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9833 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9834 var satisfies use_register_for_decl, just take into account type
9835 alignment, rather than decl alignment.
9837 PR tree-optimization/59622
9838 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9839 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9840 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9841 Don't devirtualize for inplace at all. For targets.length () == 1,
9842 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9844 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9846 * config/i386/i386.md (cpu): Remove the unused btver1.
9848 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9850 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9852 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9855 * tree-core.h (struct target_globals): New forward declaration.
9856 (struct tree_target_option): Add globals field.
9857 * tree.h (TREE_TARGET_GLOBALS): Define.
9858 (prepare_target_option_nodes_for_pch): New prototype.
9859 * target-globals.h (struct target_globals): Define even if
9861 * tree.c (prepare_target_option_node_for_pch,
9862 prepare_target_option_nodes_for_pch): New functions.
9863 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9864 * config/i386/i386.c: Include target-globals.h.
9865 (ix86_set_current_function): Instead of doing target_reinit
9866 unconditionally, use save_target_globals_default_opts and
9867 restore_target_globals.
9869 2014-01-09 Richard Biener <rguenther@suse.de>
9871 PR tree-optimization/59715
9872 * tree-cfg.h (split_critical_edges): Declare.
9873 * tree-cfg.c (split_critical_edges): Export.
9874 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9876 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9878 * cfgexpand.c (expand_stack_vars): Optionally disable
9879 asan stack protection.
9880 (expand_used_vars): Likewise.
9881 (partition_stack_vars): Likewise.
9882 * asan.c (asan_emit_stack_protection): Optionally disable
9883 after return stack usage.
9884 (instrument_derefs): Optionally disable memory access instrumentation.
9885 (instrument_builtin_call): Likewise.
9886 (instrument_strlen_call): Likewise.
9887 (asan_protect_global): Optionally disable global variables protection.
9888 * doc/invoke.texi: Added doc for new options.
9889 * params.def: Added new options.
9890 * params.h: Likewise.
9892 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9894 PR rtl-optimization/59724
9895 * ifcvt.c (cond_exec_process_if_block): Don't call
9896 flow_find_head_matching_sequence with 0 longest_match.
9897 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9898 non-active insns if !stop_after.
9899 (try_head_merge_bb): Revert 2014-01-07 changes.
9901 2014-01-08 Jeff Law <law@redhat.com>
9903 * ree.c (get_sub_rtx): New function, extracted from...
9904 (merge_def_and_ext): Here.
9905 (combine_reaching_defs): Use get_sub_rtx.
9907 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9909 * cgraph.h (varpool_variable_node): Do not choke on null node.
9911 2014-01-08 Catherine Moore <clm@codesourcery.com>
9913 * config/mips/mips.md (simple_return): Attempt to use JRC
9915 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9917 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9919 PR rtl-optimization/59137
9920 * reorg.c (steal_delay_list_from_target): Call update_block for
9922 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9924 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9926 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9927 two duplicate entries.
9929 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9932 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9934 * config/mips/mips.c (mips_truncated_op_cost): New function.
9935 (mips_rtx_costs): Adjust test for BADDU.
9936 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9938 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9940 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9941 (*baddu_si): ...this new pattern.
9943 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9946 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9948 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9951 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9953 (expand_expr, expand_normal): Adjust.
9954 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9955 inner_reference_p. Use inner_reference_p to expand inner references.
9956 (store_expr): Adjust.
9957 * cfgexpand.c (expand_call_stmt): Adjust.
9959 2014-01-08 Rong Xu <xur@google.com>
9961 * gcov-io.c (gcov_var): Move from gcov-io.h.
9962 (gcov_position): Ditto.
9963 (gcov_is_error): Ditto.
9964 (gcov_rewrite): Ditto.
9965 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9966 only part to libgcc/libgcov.h.
9968 2014-01-08 Marek Polacek <polacek@redhat.com>
9971 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9973 2014-01-08 Marek Polacek <polacek@redhat.com>
9976 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9978 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9980 PR rtl-optimization/59649
9981 * stor-layout.c (get_mode_bounds): For BImode return
9982 0 and STORE_FLAG_VALUE.
9984 2014-01-08 Richard Biener <rguenther@suse.de>
9987 * gimple.h (is_gimple_builtin_call): Remove.
9988 (gimple_builtin_call_types_compatible_p): New.
9989 (gimple_call_builtin_p): New overload.
9990 * gimple.c (is_gimple_builtin_call): Remove.
9991 (validate_call): Rename to ...
9992 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9994 (validate_type): New static function.
9995 (gimple_call_builtin_p): New overload and adjust.
9996 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9997 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9998 (gimple_fold_stmt_to_constant_1): Likewise.
9999 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10001 2014-01-08 Richard Biener <rguenther@suse.de>
10003 PR middle-end/59471
10004 * gimplify.c (gimplify_expr): Gimplify register-register type
10005 VIEW_CONVERT_EXPRs to separate stmts.
10007 2014-01-07 Jeff Law <law@redhat.com>
10009 PR middle-end/53623
10010 * ree.c (combine_set_extension): Handle case where source
10011 and destination registers in an extension insn are different.
10012 (combine_reaching_defs): Allow source and destination registers
10013 in extension to be different under limited circumstances.
10014 (add_removable_extension): Remove restriction that the
10015 source and destination registers in the extension are the same.
10016 (find_and_remove_re): Emit a copy from the extension's
10017 destination to its source after the defining insn if
10018 the source and destination registers are different.
10020 PR middle-end/59285
10021 * ifcvt.c (merge_if_block): If we are merging a block with more than
10022 one successor with a block with no successors, remove any BARRIER
10023 after the second block.
10025 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10027 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10029 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10032 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10033 for 14-bit register offsets when INT14_OK_STRICT is false.
10035 2014-01-07 Roland Stigge <stigge@antcom.de>
10036 Michael Meissner <meissner@linux.vnet.ibm.com>
10039 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10040 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10042 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10044 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10047 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10049 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10050 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10051 rtx is const0_rtx or not.
10053 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10056 * target-globals.c (save_target_globals): Remove this_fn_optab
10058 * toplev.c: Include optabs.h.
10059 (target_reinit): Temporarily restore the global options if another
10060 set of options are in force.
10062 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10064 PR rtl-optimization/58668
10065 * cfgcleanup.c (flow_find_cross_jump): Don't count
10066 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10067 to determine what is counted.
10068 (flow_find_head_matching_sequence): Use active_insn_p to determine
10070 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10072 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10073 determine what is counted.
10075 PR tree-optimization/59643
10076 * tree-predcom.c (split_data_refs_to_components): If one dr is
10077 read and one write, determine_offset fails and the write isn't
10078 in the bad component, just put the read into the bad component.
10080 2014-01-07 Mike Stump <mikestump@comcast.net>
10081 Jakub Jelinek <jakub@redhat.com>
10084 * tree-core.h (struct tree_optimization_option): Change optabs
10085 type from unsigned char * to void *.
10086 * optabs.c (init_tree_optimization_optabs): Adjust
10087 TREE_OPTIMIZATION_OPTABS initialization.
10089 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10092 * config/i386/i386.h (struct machine_function): Add
10093 no_drap_save_restore field.
10094 * config/i386/i386.c (ix86_save_reg): Use
10095 !cfun->machine->no_drap_save_restore instead of
10096 crtl->stack_realign_needed.
10097 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10098 this function clears frame_pointer_needed. Set
10099 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10100 and DRAP reg is needed.
10102 2014-01-06 Marek Polacek <polacek@redhat.com>
10105 * doc/implement-c.texi: Mention that other integer types are
10106 permitted as bit-field types in strictly conforming mode.
10108 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10110 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10111 is newly allocated.
10113 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10115 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10117 2014-01-06 Martin Jambor <mjambor@suse.cz>
10120 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10122 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10124 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10128 * var-tracking.c (add_stores): Preserve the value of the source even if
10129 we don't record the store.
10131 2014-01-06 Terry Guo <terry.guo@arm.com>
10133 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10135 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10138 * config/darwin.c (darwin_function_section): Adjust return values to
10139 correspond to optimisation changes made in r206070.
10141 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10143 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10144 from prefetch_block tune setting.
10145 (nocona_cost): Correct size of prefetch block to 64.
10147 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10149 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10150 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10151 used to save the static chain register in the computation of the offset
10152 from which the FP registers need to be restored.
10154 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10156 PR tree-optimization/59519
10157 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10158 ICE if get_current_def (current_new_name) is already non-NULL, as long
10159 as it is a phi result of some other phi in *new_exit_bb that has
10162 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10163 or vmovdqu* for misaligned_operand.
10164 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10165 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10166 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10167 aligned_mem for AVX512F masked aligned load and store builtins and for
10168 non-temporal moves.
10170 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10172 PR tree-optimization/59651
10173 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10174 Address range for negative step should be added by TYPE_SIZE_UNIT.
10176 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10178 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10179 overlapping registers also for registers other than the stack pointer.
10181 2014-01-03 Marek Polacek <polacek@redhat.com>
10184 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10187 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10190 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10193 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10194 (push splitter): Use <P:MODE_SIZE> instead of
10195 GET_MODE_SIZE (<P:MODE>mode).
10196 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10197 (mov -1, reg peephole2): Likewise.
10198 * config/i386/sse.md (*mov<mode>_internal,
10199 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10200 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10201 *<code><mode>3, *andnot<mode>3<mask_name>,
10202 <mask_codefor><code><mode>3<mask_name>): Likewise.
10203 * config/i386/subst.md (mask_mode512bit_condition,
10204 sd_mask_mode512bit_condition): Likewise.
10206 2014-01-02 Xinliang David Li <davidxl@google.com>
10208 PR tree-optimization/59303
10209 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10210 (dump_predicates): Better output format.
10211 (pred_equal_p): New function.
10212 (is_neq_relop_p): Ditto.
10213 (is_neq_zero_form_p): Ditto.
10214 (pred_expr_equal_p): Ditto.
10215 (pred_neg_p): Ditto.
10216 (simplify_pred): Ditto.
10217 (simplify_preds_2): Ditto.
10218 (simplify_preds_3): Ditto.
10219 (simplify_preds_4): Ditto.
10220 (simplify_preds): Ditto.
10221 (push_pred): Ditto.
10222 (push_to_worklist): Ditto.
10223 (get_pred_info_from_cmp): Ditto.
10224 (is_degenerated_phi): Ditto.
10225 (normalize_one_pred_1): Ditto.
10226 (normalize_one_pred): Ditto.
10227 (normalize_one_pred_chain): Ditto.
10228 (normalize_preds): Ditto.
10229 (normalize_cond_1): Remove function.
10230 (normalize_cond): Ditto.
10231 (is_gcond_subset_of): Ditto.
10232 (is_subset_of_any): Ditto.
10233 (is_or_set_subset_of): Ditto.
10234 (is_and_set_subset_of): Ditto.
10235 (is_norm_cond_subset_of): Ditto.
10236 (pred_chain_length_cmp): Ditto.
10237 (convert_control_dep_chain_into_preds): Type change.
10238 (find_predicates): Ditto.
10239 (find_def_preds): Ditto.
10240 (destroy_predicates_vecs): Ditto.
10241 (find_matching_predicates_in_rest_chains): Ditto.
10242 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10243 (is_pred_expr_subset): Ditto.
10244 (is_pred_chain_subset_of): Ditto.
10245 (is_included_in): Ditto.
10246 (is_superset_of): Ditto.
10248 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10250 Update copyright years.
10252 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10254 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10255 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10256 config/arc/arc.md, config/arc/arc.opt,
10257 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10258 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10259 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10260 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10261 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10262 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10263 vtable-verify.c, vtable-verify.h: Use the standard form for the
10266 2014-01-02 Tobias Burnus <burnus@net-b.de>
10268 * gcc.c (process_command): Update copyright notice dates.
10269 * gcov-dump.c: Ditto.
10271 * doc/cpp.texi: Bump @copying's copyright year.
10272 * doc/cppinternals.texi: Ditto.
10273 * doc/gcc.texi: Ditto.
10274 * doc/gccint.texi: Ditto.
10275 * doc/gcov.texi: Ditto.
10276 * doc/install.texi: Ditto.
10277 * doc/invoke.texi: Ditto.
10279 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10281 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10283 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10285 * config/i386/sse.md (*mov<mode>_internal): Guard
10286 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10288 PR rtl-optimization/59647
10289 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10290 new_rtx into UNSIGNED_FLOAT rtxes.
10292 Copyright (C) 2014 Free Software Foundation, Inc.
10294 Copying and distribution of this file, with or without modification,
10295 are permitted in any medium without royalty provided the copyright
10296 notice and this notice are preserved.
projects / gcc.git / blob