gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove pointer propagation special...
[gcc.git] / gcc / ChangeLog
1 2014-05-08 Richard Biener <rguenther@suse.de>
2
3 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
4 pointer propagation special-case.
5
6 2014-05-08 Bin Cheng <bin.cheng@arm.com>
7
8 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
9 core part of address expressions.
10
11 2014-05-08 Alan Modra <amodra@gmail.com>
12
13 PR target/60737
14 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
15 loads and stores when -mno-strict-align at any alignment.
16 (expand_block_clear): Similarly. Also correct calculation of
17 instruction count.
18
19 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
20
21 PR middle-end/39246
22 * tree-complex.c (expand_complex_move): Keep line info when expanding
23 complex move.
24 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
25 of complex expression. Use new argument to display correct location
26 for values coming from phi statement.
27 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
28 (warn_uninitialized_phi): Pass location of phi argument to
29 warn_uninit.
30 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
31 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
32
33 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
34
35 * config/rs6000/predicates.md (indexed_address_mem): New.
36 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
37 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
38 fpstore_ux, fpstore_u.
39 (sign_extend, indexed, update): New.
40 (cell_micro): Adjust.
41 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
42 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
43 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
44 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
45 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
46 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
47 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
48 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
49 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
50 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
51 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
52 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
53 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
54 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
55 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
56
57 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
58 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
59 *vsx_extract_<mode>_store): Adjust.
60 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
61 is_cracked_insn, insn_must_be_first_in_group,
62 insn_must_be_last_in_group): Adjust.
63
64 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
65 Adjust.
66 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
67 ppc440-fpstore): Adjust.
68 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
69 ppc476-fpstore): Adjust.
70 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
71 ppc601-fpstore): Adjust.
72 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
73 Adjust.
74 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
75 Adjust.
76 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
77 ppc7450-fpstore): Adjust.
78 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
79 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
80 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
81 Adjust.
82 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
83 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
84 cell-fpstore, cell-fpstore-update): Adjust.
85 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
86 ppce300c3_store, ppce300c3_fpstore): Adjust.
87 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
88 e500mc_fpstore): Adjust.
89 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
90 e500mc64_store, e500mc64_fpstore): Adjust.
91 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
92 e5500_fpstore): Adjust.
93 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
94 e6500_fpstore): Adjust.
95 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
96 Adjust.
97 * config/rs6000/power4.md (power4-load, power4-load-ext,
98 power4-load-ext-update, power4-load-ext-update-indexed,
99 power4-load-update-indexed, power4-load-update, power4-fpload,
100 power4-fpload-update, power4-store, power4-store-update,
101 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
102 Adjust.
103 * config/rs6000/power5.md (power5-load, power5-load-ext,
104 power5-load-ext-update, power5-load-ext-update-indexed,
105 power5-load-update-indexed, power5-load-update, power5-fpload,
106 power5-fpload-update, power5-store, power5-store-update,
107 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
108 Adjust.
109 * config/rs6000/power6.md (power6-load, power6-load-ext,
110 power6-load-update, power6-load-update-indexed,
111 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
112 power6-fpload-update, power6-store, power6-store-update,
113 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
114 Adjust.
115 * config/rs6000/power7.md (power7-load, power7-load-ext,
116 power7-load-update, power7-load-update-indexed,
117 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
118 power7-fpload-update, power7-store, power7-store-update,
119 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
120 Adjust.
121 * config/rs6000/power8.md (power8-load, power8-load-update,
122 power8-load-ext, power8-load-ext-update, power8-fpload,
123 power8-fpload-update, power8-store, power8-store-update-indexed,
124 power8-fpstore, power8-fpstore-update): Adjust.
125 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
126 Adjust.
127 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
128 titan_lsu_store, titan_lsu_fpstore): Adjust.
129 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
130
131 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
132
133 PR target/60884
134 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
135 unrolled byte insns. Emit address increments after move insns.
136
137 2014-05-07 David Malcolm <dmalcolm@redhat.com>
138
139 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
140 const_gimple, rather than a gimple.
141 (gimple_call_builtin_p): Likewise, for the three variants.
142
143 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
144 (gimple_call_builtin_p): Likewise, for the three variants.
145
146 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
147
148 PR tree-optimization/61095
149 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
150
151 2014-05-07 Richard Biener <rguenther@suse.de>
152
153 PR tree-optimization/61034
154 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
155 (maybe_skip_until): Use translate to take into account
156 lattices when trying to do disambiguations.
157 (get_continuation_for_phi_1): Likewise.
158 (get_continuation_for_phi): Adjust for added translate
159 arguments.
160 (walk_non_aliased_vuses): Likewise.
161 * tree-ssa-alias.h (get_continuation_for_phi): Adjust
162 prototype.
163 (walk_non_aliased_vuses): Likewise.
164 (call_may_clobber_ref_p_1): Declare.
165 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also
166 disambiguate against calls. Stop early if we are
167 only supposed to disambiguate.
168 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
169
170 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
171
172 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
173 Emit an error when the function has arguments.
174
175 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
176
177 * cfgloop.h (unswitch_loops): Remove.
178 * doc/passes.texi: Remove references to loop-unswitch.c
179 * timevar.def (TV_LOOP_UNSWITCH): Remove.
180
181 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
182
183 * tree-vect-data-refs.c (vect_grouped_load_supported): New
184 check for loads group of length 3.
185 (vect_permute_load_chain): New permutations for loads group of
186 length 3.
187 * tree-vect-stmts.c (vect_model_load_cost): Change cost
188 of vec_perm_shuffle for the new permutations.
189
190 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
191
192 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
193 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
194 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
195 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
196 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
197 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
198 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
199 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
200
201 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
202
203 * loop-unswitch.c: Delete.
204
205 2014-05-07 Richard Biener <rguenther@suse.de>
206
207 * config.gcc: Always set need_64bit_hwint to yes.
208
209 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
210
211 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
212 of using optimize_size.
213
214 2014-05-06 Mike Stump <mikestump@comcast.net>
215
216 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
217
218 2014-05-06 Joseph Myers <joseph@codesourcery.com>
219
220 * config/i386/sse.md (*mov<mode>_internal)
221 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
222 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
223 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
224 (*<code><mode>3, *andnot<mode>3<mask_name>)
225 (<mask_codefor><code><mode>3<mask_name>): Only consider
226 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
227
228 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
229
230 Revert:
231 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
232
233 * lra-constraints.c (valid_address_p): Move earlier in file.
234 Add a constraint argument to the address_info version.
235 (satisfies_memory_constraint_p): New function.
236 (satisfies_address_constraint_p): Likewise.
237 (process_alt_operands, curr_insn_transform): Use them.
238 (process_address): Pass the constraint to valid_address_p when
239 checking address operands.
240
241 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
242
243 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
244 to their respective blocks. Fix inadvertent use of "node".
245
246 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
247
248 * emit-rtl.c (init_derived_machine_modes): New functionm, split
249 out from...
250 (init_emit_once): ...here.
251 * rtl.h (init_derived_machine_modes): Declare.
252 * toplev.c (do_compile): Call it even if no_backend.
253
254 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
255 Mike Stump <mikestump@comcast.net>
256 Richard Sandiford <rdsandiford@googlemail.com>
257 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
258
259 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
260 (rtx_equal_for_memref_p): Update comment.
261 (adjust_offset_for_component_ref): Use wide-int interfaces.
262 * builtins.c (get_object_alignment_2): Likewise.
263 (c_readstr): Likewise.
264 (target_char_cast): Add comment.
265 (determine_block_size): Use wide-int interfaces.
266 (expand_builtin_signbit): Likewise.
267 (fold_builtin_int_roundingfn): Likewise.
268 (fold_builtin_bitop): Likewise.
269 (fold_builtin_bswap): Likewise.
270 (fold_builtin_logarithm): Use signop.
271 (fold_builtin_pow): Likewise.
272 (fold_builtin_memory_op): Use wide-int interfaces.
273 (fold_builtin_object_size): Likewise.
274 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
275 nb_iterations_estimate.
276 (record_niter_bound): Use wide-int interfaces.
277 (get_estimated_loop_iterations_int): Likewise.
278 (get_estimated_loop_iterations): Likewise.
279 (get_max_loop_iterations): Likewise.
280 * cfgloop.h: Include wide-int.h.
281 (struct nb_iter_bound): Change bound to widest_int.
282 (struct loop): Change nb_iterations_upper_bound and
283 nb_iterations_estimate to widest_int.
284 (record_niter_bound): Switch to use widest_int.
285 (get_estimated_loop_iterations): Likewise.
286 (get_max_loop_iterations): Likewise.
287 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
288 update for wide-int.
289 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
290 * combine.c (try_combine): Likewise.
291 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
292 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
293 interfaces.
294 (aarch64_float_const_representable_p): Likewise.
295 * config/arc/arc.c: Include wide-int.h.
296 (arc_can_use_doloop_p): Use wide-int interfaces.
297 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
298 (vfp3_const_double_index): Likewise.
299 * config/avr/avr.c (avr_out_round): Likewise.
300 (avr_fold_builtin): Likewise.
301 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
302 (bfin_can_use_doloop_p): Likewise.
303 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
304 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
305 * config/i386/i386.c: Include wide-int.h.
306 (ix86_data_alignment): Use wide-int interfaces.
307 (ix86_local_alignment): Likewise.
308 (ix86_emit_swsqrtsf): Update real_from_integer.
309 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
310 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
311 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
312 (zero_constant): Likewise.
313 (input_operand): Likewise.
314 (splat_input_operand): Likewise.
315 (non_logical_cint_operand): Change const_double to const_wide_int.
316 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
317 (easy_altivec_constant): Remove comment.
318 (paired_expand_vector_init): Use CONSTANT_P.
319 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
320 (rs6000_emit_move): Update checks.
321 (rs6000_aggregate_candidate): Use wide-int interfaces.
322 (rs6000_expand_ternop_builtin): Likewise.
323 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
324 (rs6000_assemble_integer): Likewise.
325 (rs6000_hash_constant): Likewise.
326 (output_toc): Likewise.
327 (rs6000_rtx_costs): Likewise.
328 (rs6000_emit_swrsqrt); Update call to real_from_integer.
329 * config/rs6000/rs6000-c.c: Include wide-int.h.
330 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
331 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
332 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
333 Handle CONST_WIDE_INT.
334 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
335 Use tree_fits_uhwi_p.
336 * config/sparc/sparc.c: Include wide-int.h.
337 (sparc_fold_builtin): Use wide-int interfaces.
338 * config/vax/vax.c: Include wide-int.h.
339 (vax_float_literal): Use real_from_integer.
340 * coretypes.h (struct hwivec_def): New.
341 (hwivec): New.
342 (const_hwivec): New.
343 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
344 (equiv_constant): Handle CONST_WIDE_INT.
345 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
346 (cselib_hash_rtx): Handle CONST_WIDE_INT.
347 * dbxout.c (stabstr_U): Use wide-int interfaces.
348 (dbxout_type): Update to use cst_fits_shwi_p.
349 * defaults.h (LOG2_BITS_PER_UNIT): Define.
350 (TARGET_SUPPORTS_WIDE_INT): Add default.
351 * dfp.c: Include wide-int.h.
352 (decimal_real_to_integer2): Use wide-int interfaces and rename to
353 decimal_real_to_integer.
354 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
355 decimal_real_to_integer.
356 * doc/generic.texi (Constant expressions): Update for wide_int.
357 * doc/rtl.texi (const_double): Likewise.
358 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
359 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
360 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
361 (REAL_VALUE_FROM_INT): Remove.
362 (TARGET_SUPPORTS_WIDE_INT): New.
363 * doc/tm.texi: Regenerate.
364 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
365 * double-int.h: Include wide-int.h.
366 (struct wi::int_traits): New.
367 * dwarf2out.c (get_full_len): New.
368 (dw_val_equal_p): Add case dw_val_class_wide_int.
369 (size_of_loc_descr): Likewise.
370 (output_loc_operands): Likewise.
371 (insert_double): Remove.
372 (insert_wide_int): New.
373 (add_AT_wide): New.
374 (print_die): Add case dw_val_class_wide_int.
375 (attr_checksum): Likewise.
376 (attr_checksum_ordered): Likewise.
377 (same_dw_val_p): Likewise.
378 (size_of_die): Likewise.
379 (value_format): Likewise.
380 (output_die): Likewise.
381 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
382 Use wide-int.
383 (clz_loc_descriptor): Use wide-int interfaces.
384 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
385 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
386 (round_up_to_align): Use wide-int interfaces.
387 (field_byte_offset): Likewise.
388 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
389 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
390 CONST_DOUBLE handling. Use wide-int interfaces.
391 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
392 (gen_enumeration_type_die): Use add_AT_wide.
393 (hash_loc_operands): Add case dw_val_class_wide_int.
394 (compare_loc_operands): Likewise.
395 * dwarf2out.h: Include wide-int.h.
396 (wide_int_ptr): New.
397 (enum dw_val_class): Add dw_val_class_wide_int.
398 (struct dw_val_struct): Add val_wide.
399 * emit-rtl.c (const_wide_int_htab): New.
400 (const_wide_int_htab_hash): New.
401 (const_wide_int_htab_eq): New.
402 (lookup_const_wide_int): New.
403 (const_double_htab_hash): Use wide-int interfaces.
404 (const_double_htab_eq): Likewise.
405 (rtx_to_double_int): Conditionally compile for wide-int.
406 (immed_double_int_const): Rename to immed_wide_int_const and
407 update for wide-int.
408 (immed_double_const): Conditionally compile for wide-int.
409 (init_emit_once): Use wide-int interfaces.
410 * explow.c (plus_constant): Likewise.
411 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
412 (lshift_value): Use wide-int interfaces.
413 (expand_mult): Likewise.
414 (choose_multiplier): Likewise.
415 (expand_smod_pow2): Likewise.
416 (make_tree): Likewise.
417 * expr.c (convert_modes): Consolidate handling of constants.
418 Use wide-int interfaces.
419 (emit_group_load_1): Add note.
420 (store_expr): Update comment.
421 (get_inner_reference): Use wide-int interfaces.
422 (expand_constructor): Update comment.
423 (expand_expr_real_2): Use wide-int interfaces.
424 (expand_expr_real_1): Likewise.
425 (reduce_to_bit_field_precision): Likewise.
426 (const_vector_from_tree): Likewise.
427 * final.c: Include wide-int-print.h.
428 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
429 * fixed-value.c: Include wide-int.h.
430 (fixed_from_string): Use wide-int interfaces.
431 (fixed_to_decimal): Likewise.
432 (fixed_convert_from_real): Likewise.
433 (real_convert_from_fixed): Likewise.
434 * fold-const.h (mem_ref_offset): Return an offset_int.
435 (div_if_zero_remainder): Remove code parameter.
436 * fold-const.c (div_if_zero_remainder): Remove code parameter.
437 Use wide-int interfaces.
438 (may_negate_without_overflow_p): Use wide-int interfaces.
439 (negate_expr_p): Likewise.
440 (fold_negate_expr): Likewise.
441 (int_const_binop_1): Likewise.
442 (const_binop): Likewise.
443 (fold_convert_const_int_from_int): Likewise.
444 (fold_convert_const_int_from_real): Likewise.
445 (fold_convert_const_int_from_fixed): Likewise.
446 (fold_convert_const_fixed_from_int): Likewise.
447 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
448 (sign_bit_p): Use wide-int interfaces.
449 (make_range_step): Likewise.
450 (build_range_check): Likewise. Pass an integer of the correct type
451 instead of using integer_one_node.
452 (range_predecessor): Pass an integer of the correct type instead
453 of using integer_one_node.
454 (range_successor): Likewise.
455 (merge_ranges): Likewise.
456 (unextend): Use wide-int interfaces.
457 (extract_muldiv_1): Likewise.
458 (fold_div_compare): Likewise.
459 (fold_single_bit_test): Likewise.
460 (fold_sign_changed_comparison): Likewise.
461 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
462 (fold_plusminus_mult_expr): Use wide-int interfaces.
463 (native_encode_int): Likewise.
464 (native_interpret_int): Likewise.
465 (fold_unary_loc): Likewise.
466 (pointer_may_wrap_p): Likewise.
467 (size_low_cst): Likewise.
468 (mask_with_tz): Likewise.
469 (fold_binary_loc): Likewise.
470 (fold_ternary_loc): Likewise.
471 (multiple_of_p): Likewise.
472 (tree_call_nonnegative_warnv_p): Update calls to
473 tree_int_cst_min_precision and real_from_integer.
474 (fold_negate_const): Use wide-int interfaces.
475 (fold_abs_const): Likewise.
476 (fold_relational_const): Use tree_int_cst_lt.
477 (round_up_loc): Use wide-int interfaces.
478 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
479 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
480 * gengtype.c: Remove include of double-int.h.
481 (do_typedef): Use wide-int interfaces.
482 (open_base_files): Add wide-int.h.
483 (main): Add offset_int and widest_int typedefs.
484 * gengtype-lex.l: Handle "^".
485 (CXX_KEYWORD): Add "static".
486 * gengtype-parse.c (require3): New.
487 (require_template_declaration): Handle constant template arguments
488 and nested templates.
489 * gengtype-state.c: Don't include "double-int.h".
490 * genpreds.c (write_one_predicate_function): Update comment.
491 (write_tm_constrs_h): Add check for hval and lval use in
492 CONST_WIDE_INT.
493 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
494 (add_to_sequence): Likewise.
495 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
496 and const_double_operand.
497 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
498 interfaces.
499 * gimple-fold.c (get_base_constructor): Likewise.
500 (fold_array_ctor_reference): Likewise.
501 (fold_nonarray_ctor_reference): Likewise.
502 (fold_const_aggregate_ref_1): Likewise.
503 (gimple_val_nonnegative_real_p): Likewise.
504 (gimple_fold_indirect_ref): Likewise.
505 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
506 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
507 (struct slsr_cand_d): Change index to be widest_int.
508 (struct incr_info_d): Change incr to be widest_int.
509 (alloc_cand_and_find_basis): Use wide-int interfaces.
510 (slsr_process_phi): Likewise.
511 (backtrace_base_for_ref): Likewise. Return a widest_int.
512 (restructure_reference): Take a widest_int instead of a double_int.
513 (slsr_process_ref): Use wide-int interfaces.
514 (create_mul_ssa_cand): Likewise.
515 (create_mul_imm_cand): Likewise.
516 (create_add_ssa_cand): Likewise.
517 (create_add_imm_cand): Take a widest_int instead of a double_int.
518 (slsr_process_add): Use wide-int interfaces.
519 (slsr_process_cast): Likewise.
520 (slsr_process_copy): Likewise.
521 (dump_candidate): Likewise.
522 (dump_incr_vec): Likewise.
523 (replace_ref): Likewise.
524 (cand_increment): Likewise. Return a widest_int.
525 (cand_abs_increment): Likewise.
526 (replace_mult_candidate): Take a widest_int instead of a double_int.
527 (replace_unconditional_candidate): Use wide-int interfaces.
528 (incr_vec_index): Take a widest_int instead of a double_int.
529 (create_add_on_incoming_edge): Likewise.
530 (create_phi_basis): Use wide-int interfaces.
531 (replace_conditional_candidate): Likewise.
532 (record_increment): Take a widest_int instead of a double_int.
533 (record_phi_increments): Use wide-int interfaces.
534 (phi_incr_cost): Take a widest_int instead of a double_int.
535 (lowest_cost_path): Likewise.
536 (total_savings): Likewise.
537 (analyze_increments): Use wide-int interfaces.
538 (ncd_with_phi): Take a widest_int instead of a double_int.
539 (ncd_of_cand_and_phis): Likewise.
540 (nearest_common_dominator_for_cands): Likewise.
541 (insert_initializers): Use wide-int interfaces.
542 (all_phi_incrs_profitable): Likewise.
543 (replace_one_candidate): Likewise.
544 (replace_profitable_candidates): Likewise.
545 * godump.c: Include wide-int-print.h.
546 (go_output_typedef): Use wide-int interfaces.
547 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
548 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
549 (build_loop_iteration_domains): Likewise.
550 * hooks.h: Include wide-int.h rather than double-int.h.
551 (hook_bool_dint_dint_uint_bool_true): Delete.
552 (hook_bool_wint_wint_uint_bool_true): Declare.
553 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
554 (hook_bool_wint_wint_uint_bool_true): New.
555 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
556 interfaces.
557 (ubsan_expand_si_overflow_mul_check): Likewise.
558 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
559 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
560 (get_ancestor_addr_info): Likewise.
561 (ipa_modify_call_arguments): Likewise.
562 * loop-doloop.c (doloop_modify): Likewise.
563 (doloop_optimize): Likewise.
564 * loop-iv.c (iv_number_of_iterations): Likewise.
565 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
566 (unroll_loop_constant_iterations): Likewise.
567 (decide_unroll_runtime_iterations): Likewise.
568 (unroll_loop_runtime_iterations): Likewise.
569 (decide_peel_simple): Likewise.
570 (decide_unroll_stupid): Likewise.
571 * lto-streamer-in.c (streamer_read_wi): Add.
572 (input_cfg): Use wide-int interfaces.
573 (lto_input_tree_1): Likewise.
574 * lto-streamer-out.c (streamer_write_wi): Add.
575 (hash_tree): Use wide-int interfaces.
576 (output_cfg): Likewise.
577 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
578 (GTFILES): Add wide-int.h and signop.h.
579 (TAGS): Look for .cc files too.
580 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
581 * optabs.c (expand_subword_shift): Likewise.
582 (expand_doubleword_shift): Likewise.
583 (expand_absneg_bit): Likewise.
584 (expand_copysign_absneg): Likewise.
585 (expand_copysign_bit): Likewise.
586 * postreload.c (reload_cse_simplify_set): Likewise.
587 * predict.c (predict_iv_comparison): Likewise.
588 * pretty-print.h: Include wide-int-print.h.
589 (pp_wide_int) New.
590 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
591 * print-tree.c: Include wide-int-print.h.
592 (print_node_brief): Use wide-int interfaces.
593 (print_node): Likewise.
594 * read-rtl.c (validate_const_wide_int): New.
595 (read_rtx_code): Add CONST_WIDE_INT case.
596 * real.c: Include wide-int.h.
597 (real_to_integer2): Delete.
598 (real_to_integer): New function, returning a wide_int.
599 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
600 (ten_to_ptwo): Update call to real_from_integer.
601 (real_digit): Likewise.
602 * real.h: Include signop.h, wide-int.h and insn-modes.h.
603 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
604 (REAL_VALUE_TO_INT): Delete.
605 (real_to_integer): Declare a wide-int form.
606 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
607 * recog.c (const_int_operand): Improve comment.
608 (const_scalar_int_operand): New.
609 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
610 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
611 (split_double): Likewise.
612 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
613 (rtx_size): Likewise.
614 (rtx_alloc_stat_v): New.
615 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
616 (cwi_output_hex): New.
617 (iterative_hash_rtx): Handle CONST_WIDE_INT.
618 (cwi_check_failed_bounds): New.
619 * rtl.def (CONST_WIDE_INT): New.
620 * rtl.h: Include <utility> and wide-int.h.
621 (struct hwivec_def): New.
622 (CWI_GET_NUM_ELEM): New.
623 (CWI_PUT_NUM_ELEM): New.
624 (struct rtx_def): Add num_elem and hwiv.
625 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
626 (CASE_CONST_UNIQUE): Likewise.
627 (CASE_CONST_ANY): Likewise.
628 (CONST_SCALAR_INT_P): Likewise.
629 (CONST_WIDE_INT_P): New.
630 (CWI_ELT): New.
631 (HWIVEC_CHECK): New.
632 (cwi_check_failed_bounds): New.
633 (CWI_ELT): New.
634 (HWIVEC_CHECK): New.
635 (CONST_WIDE_INT_VEC) New.
636 (CONST_WIDE_INT_NUNITS) New.
637 (CONST_WIDE_INT_ELT) New.
638 (rtx_mode_t): New type.
639 (wi::int_traits <rtx_mode_t>): New.
640 (wi::shwi): New.
641 (wi::min_value): New.
642 (wi::max_value): New.
643 (rtx_alloc_v) New.
644 (const_wide_int_alloc): New.
645 (immed_wide_int_const): New.
646 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
647 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
648 * signop.h: New file.
649 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
650 (simplify_const_unary_operation): Use wide-int interfaces.
651 (simplify_binary_operation_1): Likewise.
652 (simplify_const_binary_operation): Likewise.
653 (simplify_const_relational_operation): Likewise.
654 (simplify_immed_subreg): Likewise.
655 * stmt.c (expand_case): Likewise.
656 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
657 signop rather than a bool.
658 * stor-layout.c (layout_type): Use wide-int interfaces.
659 (initialize_sizetypes): Update calls to
660 set_min_and_max_values_for_integral_type.
661 (set_min_and_max_values_for_integral_type): Take a signop rather
662 than a bool. Use wide-int interfaces.
663 (fixup_signed_type): Update accordingly. Remove
664 HOST_BITS_PER_DOUBLE_INT limit.
665 (fixup_unsigned_type): Likewise.
666 * system.h (STATIC_CONSTANT_P): New.
667 (STATIC_ASSERT): New.
668 * target.def (can_use_doloop_p): Take widest_ints rather than
669 double_ints.
670 * target.h: Include wide-int.h rather than double-int.h.
671 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
672 than double_ints.
673 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
674 rather than INT_CST_LT_UNSIGNED.
675 (can_use_doloop_if_innermost): Take widest_ints rather than
676 double_ints.
677 * tree-affine.c: Include wide-int-print.h.
678 (double_int_ext_for_comb): Delete.
679 (wide_int_ext_for_comb): New.
680 (aff_combination_zero): Use wide-int interfaces.
681 (aff_combination_const): Take a widest_int instead of a double_int.
682 (aff_combination_elt): Use wide-int interfaces.
683 (aff_combination_scale): Take a widest_int instead of a double_int.
684 (aff_combination_add_elt): Likewise.
685 (aff_combination_add_cst): Likewise.
686 (aff_combination_add): Use wide-int interfaces.
687 (aff_combination_convert): Likewise.
688 (tree_to_aff_combination): Likewise.
689 (add_elt_to_tree): Take a widest_int instead of a double_int.
690 (aff_combination_to_tree): Use wide-int interfaces.
691 (aff_combination_remove_elt): Likewise.
692 (aff_combination_add_product): Take a widest_int instead of
693 a double_int.
694 (aff_combination_mult): Use wide-int interfaces.
695 (aff_combination_expand): Likewise.
696 (double_int_constant_multiple_p): Delete.
697 (wide_int_constant_multiple_p): New.
698 (aff_combination_constant_multiple_p): Take a widest_int pointer
699 instead of a double_int pointer.
700 (print_aff): Use wide-int interfaces.
701 (get_inner_reference_aff): Take a widest_int pointer
702 instead of a double_int pointer.
703 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
704 * tree-affine.h: Include wide-int.h.
705 (struct aff_comb_elt): Change type of coef to widest_int.
706 (struct affine_tree_combination): Change type of offset to widest_int.
707 (double_int_ext_for_comb): Delete.
708 (wide_int_ext_for_comb): New.
709 (aff_combination_const): Use widest_int instead of double_int.
710 (aff_combination_scale): Likewise.
711 (aff_combination_add_elt): Likewise.
712 (aff_combination_constant_multiple_p): Likewise.
713 (get_inner_reference_aff): Likewise.
714 (aff_comb_cannot_overlap_p): Likewise.
715 (aff_combination_zero_p): Use wide-int interfaces.
716 * tree.c: Include tree.h.
717 (init_ttree): Use make_int_cst.
718 (tree_code_size): Removed code for INTEGER_CST case.
719 (tree_size): Add INTEGER_CST case.
720 (make_node_stat): Update comment.
721 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
722 (build_int_cst_type): Use wide-int interfaces.
723 (double_int_to_tree): Likewise.
724 (double_int_fits_to_tree_p): Delete.
725 (force_fit_type_double): Delete.
726 (force_fit_type): New.
727 (int_cst_hash_hash): Use wide-int interfaces.
728 (int_cst_hash_eq): Likewise.
729 (build_int_cst_wide): Delete.
730 (wide_int_to_tree): New.
731 (cache_integer_cst): Use wide-int interfaces.
732 (build_low_bits_mask): Likewise.
733 (cst_and_fits_in_hwi): Likewise.
734 (real_value_from_int_cst): Likewise.
735 (make_int_cst_stat): New.
736 (integer_zerop): Use wide_int interfaces.
737 (integer_onep): Likewise.
738 (integer_all_onesp): Likewise.
739 (integer_pow2p): Likewise.
740 (integer_nonzerop): Likewise.
741 (tree_log2): Likewise.
742 (tree_floor_log2): Likewise.
743 (tree_ctz): Likewise.
744 (int_size_in_bytes): Likewise.
745 (mem_ref_offset): Return an offset_int rather than a double_int.
746 (build_type_attribute_qual_variant): Use wide_int interfaces.
747 (type_hash_eq): Likewise
748 (tree_int_cst_equal): Likewise.
749 (tree_int_cst_lt): Delete.
750 (tree_int_cst_compare): Likewise.
751 (tree_fits_shwi_p): Use wide_int interfaces.
752 (tree_fits_uhwi_p): Likewise.
753 (tree_int_cst_sign_bit): Likewise.
754 (tree_int_cst_sgn): Likewise.
755 (tree_int_cst_min_precision): Take a signop rather than a bool.
756 (simple_cst_equal): Use wide_int interfaces.
757 (compare_tree_int): Likewise.
758 (iterative_hash_expr): Likewise.
759 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
760 INT_CST_LT.
761 (get_type_static_bounds): Use wide_int interfaces.
762 (tree_int_cst_elt_check_failed): New.
763 (build_common_tree_nodes): Reordered to set prec before filling in
764 value.
765 (int_cst_value): Check cst_and_fits_in_hwi.
766 (widest_int_cst_value): Use wide_int interfaces.
767 (upper_bound_in_type): Likewise.
768 (lower_bound_in_type): Likewise.
769 (num_ending_zeros): Likewise.
770 (drop_tree_overflow): Likewise.
771 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
772 (gen_conditions_for_pow_cst_base): Likewise.
773 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
774 (group_case_labels_stmt): Use wide-int interfaces.
775 (verify_gimple_assign_binary): Likewise.
776 (print_loop): Likewise.
777 * tree-chrec.c (tree_fold_binomial): Likewise.
778 * tree-core.h (struct tree_base): Add int_length.
779 (struct tree_int_cst): Change rep of value.
780 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
781 (dr_may_alias_p): Likewise.
782 (max_stmt_executions_tree): Likewise.
783 * tree.def (INTEGER_CST): Update comment.
784 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
785 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
786 * tree-dump.c: Include wide-int.h and wide-int-print.h.
787 (dequeue_and_dump): Use wide-int interfaces.
788 * tree.h: Include wide-int.h.
789 (NULL_TREE): Moved to earlier loc in file.
790 (TREE_INT_CST_ELT_CHECK): New.
791 (tree_int_cst_elt_check_failed): New.
792 (TYPE_SIGN): New.
793 (TREE_INT_CST): Delete.
794 (TREE_INT_CST_LOW): Use wide-int interfaces.
795 (TREE_INT_CST_HIGH): Delete.
796 (TREE_INT_CST_NUNITS): New.
797 (TREE_INT_CST_EXT_NUNITS): Likewise.
798 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
799 (TREE_INT_CST_ELT): Likewise.
800 (INT_CST_LT): Delete.
801 (tree_int_cst_elt_check): New (two forms).
802 (type_code_size): Update comment.
803 (make_int_cst_stat, make_int_cst): New.
804 (tree_to_double_int): Delete.
805 (double_int_fits_to_tree_p): Delete.
806 (force_fit_type_double): Delete.
807 (build_int_cstu): Replace with out-of-line function.
808 (build_int_cst_wide): Delete.
809 (tree_int_cst_lt): Define inline.
810 (tree_int_cst_le): New.
811 (tree_int_cst_compare): Define inline.
812 (tree_int_cst_min_precision): Take a signop rather than a bool.
813 (wi::int_traits <const_tree>): New.
814 (wi::int_traits <tree>): New.
815 (wi::extended_tree): New.
816 (wi::int_traits <wi::extended_tree>): New.
817 (wi::to_widest): New.
818 (wi::to_offset): New.
819 (wi::fits_to_tree_p): New.
820 (wi::min_value): New.
821 (wi::max_value): New.
822 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
823 (copy_tree_body_r): Likewise.
824 * tree-object-size.c (compute_object_offset): Likewise.
825 (addr_object_size): Likewise.
826 * tree-predcom.c: Include wide-int-print.h.
827 (struct dref_d): Change type of offset to widest_int.
828 (dump_dref): Call wide-int printer.
829 (aff_combination_dr_offset): Use wide-int interfaces.
830 (determine_offset): Take a widest_int pointer rather than a
831 double_int pointer.
832 (split_data_refs_to_components): Use wide-int interfaces.
833 (suitable_component_p): Likewise.
834 (order_drefs): Likewise.
835 (add_ref_to_chain): Likewise.
836 (valid_initializer_p): Likewise.
837 (determine_roots_comp): Likewise.
838 * tree-pretty-print.c: Include wide-int-print.h.
839 (dump_generic_node): Use wide-int interfaces.
840 * tree-sra.c (sra_ipa_modify_expr): Likewise.
841 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
842 (move_fixed_address_to_symbol): Likewise.
843 (move_hint_to_base): Likewise.
844 (move_pointer_to_base): Likewise.
845 (move_variant_to_index): Likewise.
846 (most_expensive_mult_to_index): Likewise.
847 (addr_to_parts): Likewise.
848 (copy_ref_info): Likewise.
849 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
850 (indirect_refs_may_alias_p): Likewise.
851 (stmt_kills_ref_p_1): Likewise.
852 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
853 * tree-ssa-ccp.c: Update comment at top of file. Include
854 wide-int-print.h.
855 (struct prop_value_d): Change type of mask to widest_int.
856 (extend_mask): New function.
857 (dump_lattice_value): Use wide-int interfaces.
858 (get_default_value): Likewise.
859 (set_constant_value): Likewise.
860 (set_value_varying): Likewise.
861 (valid_lattice_transition): Likewise.
862 (set_lattice_value): Likewise.
863 (value_to_double_int): Delete.
864 (value_to_wide_int): New.
865 (get_value_from_alignment): Use wide-int interfaces.
866 (get_value_for_expr): Likewise.
867 (do_dbg_cnt): Likewise.
868 (ccp_finalize): Likewise.
869 (ccp_lattice_meet): Likewise.
870 (bit_value_unop_1): Use widest_ints rather than double_ints.
871 (bit_value_binop_1): Likewise.
872 (bit_value_unop): Use wide-int interfaces.
873 (bit_value_binop): Likewise.
874 (bit_value_assume_aligned): Likewise.
875 (evaluate_stmt): Likewise.
876 (ccp_fold_stmt): Likewise.
877 (visit_cond_stmt): Likewise.
878 (ccp_visit_stmt): Likewise.
879 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
880 (constant_pointer_difference): Likewise.
881 (associate_pointerplus): Likewise.
882 (combine_conversions): Likewise.
883 * tree-ssa-loop.h: Include wide-int.h.
884 (struct tree_niter_desc): Change type of max to widest_int.
885 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
886 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
887 (remove_redundant_iv_tests): Likewise.
888 (canonicalize_loop_induction_variables): Likewise.
889 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
890 (constant_multiple_of): Take a widest_int pointer instead of
891 a double_int pointer.
892 (get_computation_aff): Use wide-int interfaces.
893 (ptr_difference_cost): Likewise.
894 (difference_cost): Likewise.
895 (get_loop_invariant_expr_id): Likewise.
896 (get_computation_cost_at): Likewise.
897 (iv_elimination_compare_lt): Likewise.
898 (may_eliminate_iv): Likewise.
899 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
900 instead of double_int.
901 (max_loop_iterations): Likewise.
902 (max_stmt_executions): Likewise.
903 (estimated_stmt_executions): Likewise.
904 * tree-ssa-loop-niter.c: Include wide-int-print.h.
905 (split_to_var_and_offset): Use wide-int interfaces.
906 (determine_value_range): Likewise.
907 (bound_difference_of_offsetted_base): Likewise.
908 (bounds_add): Take a widest_int instead of a double_int.
909 (number_of_iterations_ne_max): Use wide-int interfaces.
910 (number_of_iterations_ne): Likewise.
911 (number_of_iterations_lt_to_ne): Likewise.
912 (assert_loop_rolls_lt): Likewise.
913 (number_of_iterations_lt): Likewise.
914 (number_of_iterations_le): Likewise.
915 (number_of_iterations_cond): Likewise.
916 (number_of_iterations_exit): Likewise.
917 (finite_loop_p): Likewise.
918 (derive_constant_upper_bound_assign): Likewise.
919 (derive_constant_upper_bound): Return a widest_int.
920 (derive_constant_upper_bound_ops): Likewise.
921 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
922 (record_estimate): Take a widest_int rather than a double_int.
923 (record_nonwrapping_iv): Use wide-int interfaces.
924 (double_int_cmp): Delete.
925 (wide_int_cmp): New.
926 (bound_index): Take a widest_int rather than a double_int.
927 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
928 (maybe_lower_iteration_bound): Likewise.
929 (estimate_numbers_of_iterations_loop): Likewise.
930 (estimated_loop_iterations): Take a widest_int pointer than than
931 a double_int pointer.
932 (estimated_loop_iterations_int): Use wide-int interfaces.
933 (max_loop_iterations): Take a widest_int pointer than than
934 a double_int pointer.
935 (max_loop_iterations_int): Use wide-int interfaces.
936 (max_stmt_executions): Take a widest_int pointer than than
937 a double_int pointer.
938 (estimated_stmt_executions): Likewise.
939 (n_of_executions_at_most): Use wide-int interfaces.
940 (scev_probably_wraps_p): Likewise.
941 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
942 to real_to_integer.
943 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
944 interfaces.
945 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
946 double_ints. Adjust for trailing_wide_ints <3> representation.
947 (set_nonzero_bits): Likewise.
948 (get_range_info): Return wide_ints rather than double_ints.
949 Adjust for trailing_wide_ints <3> representation.
950 (get_nonzero_bits): Likewise.
951 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
952 representation.
953 * tree-ssanames.h (struct range_info_def): Replace min, max and
954 nonzero_bits with a trailing_wide_ints <3>.
955 (set_range_info): Use wide_int_refs rather than double_ints.
956 (set_nonzero_bits): Likewise.
957 (get_range_info): Return wide_ints rather than double_ints.
958 (get_nonzero_bits): Likewise.
959 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
960 * tree-ssa-pre.c (phi_translate_1): Likewise.
961 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
962 (acceptable_pow_call): Likewise.
963 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
964 interfaces.
965 (vn_reference_fold_indirect): Likewise.
966 (vn_reference_maybe_forwprop_address): Likewise.
967 (valueize_refs_1): Likewise.
968 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
969 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
970 tree_int_cst_lt and tree_int_cst_le.
971 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
972 interfaces.
973 (streamer_alloc_tree): Likewise.
974 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
975 (streamer_write_tree_header): Likewise.
976 (streamer_write_integer_cst): Likewise.
977 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
978 (build_constructors): Likewise.
979 (array_value_type): Likewise.
980 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
981 (vect_check_gather): Likewise.
982 * tree-vect-generic.c (build_replicated_const): Likewise.
983 (expand_vector_divmod): Likewise.
984 * tree-vect-loop.c (vect_transform_loop): Likewise.
985 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
986 (vect_do_peeling_for_alignment): Likewise.
987 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
988 * tree-vrp.c: Include wide-int.h.
989 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
990 (extract_range_from_assert): Use wide-int interfaces.
991 (vrp_int_const_binop): Likewise.
992 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
993 double_int pointers.
994 (ranges_from_anti_range): Use wide-int interfaces.
995 (quad_int_cmp): Delete.
996 (quad_int_pair_sort): Likewise.
997 (extract_range_from_binary_expr_1): Use wide-int interfaces.
998 (extract_range_from_unary_expr_1): Likewise.
999 (adjust_range_with_scev): Likewise.
1000 (masked_increment): Take and return wide_ints rather than double_ints.
1001 (register_edge_assert_for_2): Use wide-int interfaces.
1002 (check_array_ref): Likewise.
1003 (search_for_addr_array): Likewise.
1004 (maybe_set_nonzero_bits): Likewise.
1005 (union_ranges): Pass an integer of the correct type instead of
1006 using integer_one_node.
1007 (intersect_ranges): Likewise.
1008 (simplify_truth_ops_using_ranges): Likewise.
1009 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1010 (range_fits_type_p): Likewise.
1011 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1012 a bool.
1013 (simplify_conversion_using_ranges): Use wide-int interfaces.
1014 (simplify_float_conversion_using_ranges): Likewise.
1015 (vrp_finalize): Likewise.
1016 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1017 (gimple_stringops_transform): Likewise.
1018 * varasm.c (decode_addr_const): Likewise.
1019 (const_hash_1): Likewise.
1020 (const_rtx_hash_1): Likewise
1021 (output_constant): Likewise.
1022 (array_size_for_constructor): Likewise.
1023 (output_constructor_regular_field): Likewise.
1024 (output_constructor_bitfield): Likewise.
1025 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1026 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1027 GENERATOR_FILEs.
1028 * gencheck.c: Define BITS_PER_UNIT.
1029 * wide-int.cc: New.
1030 * wide-int.h: New.
1031 * wide-int-print.cc: New.
1032 * wide-int-print.h: New.
1033
1034 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1035
1036 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1037
1038 2014-05-06 Richard Biener <rguenther@suse.de>
1039
1040 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1041 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1042 (TODO_verify_all): Adjust.
1043 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1044 TODO_verify_stmts and TODO_verify_rtl_sharing.
1045 * bb-reorder.c: Likewise.
1046 * cfgexpand.c: Likewise.
1047 * cprop.c: Likewise.
1048 * cse.c: Likewise.
1049 * function.c: Likewise.
1050 * fwprop.c: Likewise.
1051 * gcse.c: Likewise.
1052 * gimple-ssa-isolate-paths.c: Likewise.
1053 * gimple-ssa-strength-reduction.c: Likewise.
1054 * ipa-split.c: Likewise.
1055 * loop-init.c: Likewise.
1056 * loop-unroll.c: Likewise.
1057 * lower-subreg.c: Likewise.
1058 * modulo-sched.c: Likewise.
1059 * postreload-gcse.c: Likewise.
1060 * predict.c: Likewise.
1061 * recog.c: Likewise.
1062 * sched-rgn.c: Likewise.
1063 * store-motion.c: Likewise.
1064 * tracer.c: Likewise.
1065 * trans-mem.c: Likewise.
1066 * tree-call-cdce.c: Likewise.
1067 * tree-cfg.c: Likewise.
1068 * tree-cfgcleanup.c: Likewise.
1069 * tree-complex.c: Likewise.
1070 * tree-eh.c: Likewise.
1071 * tree-emutls.c: Likewise.
1072 * tree-if-conv.c: Likewise.
1073 * tree-into-ssa.c: Likewise.
1074 * tree-loop-distribution.c: Likewise.
1075 * tree-object-size.c: Likewise.
1076 * tree-parloops.c: Likewise.
1077 * tree-pass.h: Likewise.
1078 * tree-sra.c: Likewise.
1079 * tree-ssa-ccp.c: Likewise.
1080 * tree-ssa-copy.c: Likewise.
1081 * tree-ssa-copyrename.c: Likewise.
1082 * tree-ssa-dce.c: Likewise.
1083 * tree-ssa-dom.c: Likewise.
1084 * tree-ssa-dse.c: Likewise.
1085 * tree-ssa-forwprop.c: Likewise.
1086 * tree-ssa-ifcombine.c: Likewise.
1087 * tree-ssa-loop-ch.c: Likewise.
1088 * tree-ssa-loop-ivcanon.c: Likewise.
1089 * tree-ssa-loop.c: Likewise.
1090 * tree-ssa-math-opts.c: Likewise.
1091 * tree-ssa-phiopt.c: Likewise.
1092 * tree-ssa-phiprop.c: Likewise.
1093 * tree-ssa-pre.c: Likewise.
1094 * tree-ssa-reassoc.c: Likewise.
1095 * tree-ssa-sink.c: Likewise.
1096 * tree-ssa-strlen.c: Likewise.
1097 * tree-ssa-tail-merge.c: Likewise.
1098 * tree-ssa-uncprop.c: Likewise.
1099 * tree-switch-conversion.c: Likewise.
1100 * tree-tailcall.c: Likewise.
1101 * tree-vect-generic.c: Likewise.
1102 * tree-vectorizer.c: Likewise.
1103 * tree-vrp.c: Likewise.
1104 * tsan.c: Likewise.
1105 * var-tracking.c: Likewise.
1106 * bt-load.c: Likewise.
1107 * cfgcleanup.c: Likewise.
1108 * combine-stack-adj.c: Likewise.
1109 * combine.c: Likewise.
1110 * compare-elim.c: Likewise.
1111 * config/epiphany/resolve-sw-modes.c: Likewise.
1112 * config/i386/i386.c: Likewise.
1113 * config/mips/mips.c: Likewise.
1114 * config/s390/s390.c: Likewise.
1115 * config/sh/sh_treg_combine.cc: Likewise.
1116 * config/sparc/sparc.c: Likewise.
1117 * dce.c: Likewise.
1118 * dse.c: Likewise.
1119 * final.c: Likewise.
1120 * ifcvt.c: Likewise.
1121 * mode-switching.c: Likewise.
1122 * passes.c: Likewise.
1123 * postreload.c: Likewise.
1124 * ree.c: Likewise.
1125 * reg-stack.c: Likewise.
1126 * regcprop.c: Likewise.
1127 * regrename.c: Likewise.
1128 * web.c: Likewise.
1129
1130 2014-05-06 Richard Biener <rguenther@suse.de>
1131
1132 PR middle-end/61070
1133 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1134 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1135
1136 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1137
1138 PR ipa/60965
1139 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1140
1141 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1142 Tom de Vries <tom@codesourcery.com>
1143
1144 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
1145 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1146 Hooks to @menu.
1147 (@node Miscellaneous Register Hooks): New node.
1148 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1149 * doc/tm.texi: Regenerate.
1150
1151 2014-05-05 Marek Polacek <polacek@redhat.com>
1152
1153 PR driver/61065
1154 * opts.c (common_handle_option): Call error_at instead of warning_at.
1155
1156 2014-05-05 Richard Biener <rguenther@suse.de>
1157
1158 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1159 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1160 under the TODO_verify_il umbrella.
1161
1162 2014-05-05 Richard Biener <rguenther@suse.de>
1163
1164 * passes.c (execute_function_todo): Move TODO_verify_flow under
1165 the TODO_verify_ul umbrella.
1166
1167 2014-05-05 Richard Biener <rguenther@suse.de>
1168
1169 PR middle-end/61010
1170 * fold-const.c (fold_binary_loc): Consistently avoid
1171 canonicalizing X & CST away from a CST that is the mask
1172 of a mode.
1173
1174 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1175
1176 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1177 int argument to enum machine_mode.
1178 (picochip_class_max_nregs): Ditto.
1179 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1180 (picochip_class_max_nregs): Ditto.
1181
1182 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1183
1184 * target.def: Add new target hook.
1185 * doc/tm.texi: Regenerate.
1186 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1187 * targhooks.c (default_keep_leaf_when_profiled): New function.
1188
1189 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1190 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1191
1192 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1193
1194 PR tree-optimization/60363
1195 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1196 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1197 (update_destination_phis): New parameter.
1198 (create_edge_and_update_destination_phis): Ditto.
1199 (ssa_fix_duplicate_block_edges): Pass new arguments.
1200 (thread_single_edge): Ditto.
1201
1202 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1203
1204 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1205 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1206 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1207 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1208 Use RS6000_BTM_HARD_FLOAT.
1209 (BU_MISC_2): Likewise.
1210 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1211 RS6000_BTM_HARD_FLOAT.
1212 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1213 is explicitly used.
1214 (rs6000_invalid_builtin): Add hard floating builtin support.
1215 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1216 hard float builtins.
1217 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1218
1219 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1220
1221 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1222 Add missing function* argument.
1223
1224 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1225
1226 * lra-constraints.c (valid_address_p): Move earlier in file.
1227 Add a constraint argument to the address_info version.
1228 (satisfies_memory_constraint_p): New function.
1229 (satisfies_address_constraint_p): Likewise.
1230 (process_alt_operands, curr_insn_transform): Use them.
1231 (process_address): Pass the constraint to valid_address_p when
1232 checking address operands.
1233
1234 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1235
1236 * config/mips/mips.c (mips_isa_rev): New variable.
1237 (mips_set_architecture): Set it.
1238 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1239 from mips_isa_rev.
1240 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1241 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1242 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1243 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1244 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1245 conditions in terms of mips_isa_rev.
1246 (mips_isa_rev): Declare.
1247
1248 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1249
1250 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1251 (prob_unlikely, prob_likely): Make variables const.
1252
1253 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1254
1255 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1256
1257 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1258
1259 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1260
1261 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1262
1263 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1264 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1265 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1266 functions.
1267 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1268 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1269 sh_pass_in_reg_p.
1270 Replace usage of ROUND_REG with sh_round_reg.
1271 Use CEIL instead of ROUND_ADVANCE.
1272
1273 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1274
1275 PR target/61026
1276 * config/sh/sh.c: Include stdlib headers before everything else.
1277
1278 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1279
1280 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1281 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1282 (gimplify_adjust_omp_clauses): Simd region is never
1283 directly nested in combined parallel. Instead, for linear
1284 with copyin/copyout, if in combined for simd loop, make decl
1285 firstprivate/lastprivate on OMP_FOR.
1286 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1287 expand_omp_for_static_chunk): When setting endvar, also set
1288 fd->loop.v to the same value.
1289
1290 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1291
1292 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1293
1294 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1295
1296 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1297 expression.
1298
1299 2014-05-02 Marek Polacek <polacek@redhat.com>
1300
1301 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1302
1303 2014-05-02 Kito Cheng <kito@0xlab.org>
1304
1305 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1306 to a C expression marco.
1307 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1308 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1309 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1310 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1311 HONOR_REG_ALLOC_ORDER.
1312 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1313
1314 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1315
1316 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1317
1318 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1319
1320 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1321
1322 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1323
1324 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1325 (convert_scalar_cond_reduction): Likewise.
1326 (predicate_scalar_phi): Add recognition and transformation
1327 of simple conditioanl reduction to be vectorizable.
1328
1329 2014-05-01 Marek Polacek <polacek@redhat.com>
1330
1331 PR c/43245
1332 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1333
1334 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1335
1336 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1337 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1338 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1339 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1340 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1341 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1342 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1343 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1344
1345 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1346
1347 * config/arc/arc.opt (mlra): Move comment above option name
1348 to avoid mis-parsing as language options.
1349
1350 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1351
1352 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1353 * config/sol2.h: ... here.
1354 * config/sol2-10.h: Remove.
1355
1356 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1357 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1358 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1359 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1360 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1361 * config/sol2.h: ... here.
1362 (SECTION_NAME_FORMAT): Don't redefine.
1363 (STARTFILE_ARCH32_SPEC): Rename to ...
1364 (STARTFILE_ARCH_SPEC): ... this.
1365 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1366 * config/sparc/sol2.h: ... here.
1367 (SECTION_NAME_FORMAT): Don't undef.
1368 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1369 (SUBTARGET_EXTRA_SPECS): Remove.
1370 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1371
1372 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1373 (MD_STARTFILE_PREFIX): Remove.
1374 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1375 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1376 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1377 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1378 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1379 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1380 * config/i386/sol2.h: ... here.
1381 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1382 * config/i386/sol2-bi.h: Remove.
1383 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1384 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1385
1386 * config/i386/t-sol2-64: Rename to ...
1387 * config/i386/t-sol2: ... this.
1388 * config/sparc/t-sol2-64: Rename to ...
1389 * config/sparc/t-sol2: ... this.
1390
1391 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1392 sol2_tm_file_head, sol2_tm_file_tail.
1393 Include ${cpu_type}/sol2.h before sol2.h.
1394 Remove sol2-10.h.
1395 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1396 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1397 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1398 Reflect i386/t-sol2-64 renaming.
1399 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1400 Reflect sparc/t-sol2-64 renaming.
1401
1402 2014-04-30 Richard Biener <rguenther@suse.de>
1403
1404 * passes.c (execute_function_todo): Move TODO_verify_stmts
1405 and TODO_verify_ssa under the TODO_verify_il umbrella.
1406 * tree-ssa.h (verify_ssa): Adjust prototype.
1407 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1408 we should verify SSA operands.
1409 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1410 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1411 whether we should verify whether not throwing stmts have EH info.
1412 * graphite-scop-detection.c (create_sese_edges): Adjust.
1413 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1414 * tree-eh.c (lower_try_finally_switch): Do not add the
1415 default case label twice.
1416
1417 2014-04-30 Marek Polacek <polacek@redhat.com>
1418
1419 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1420 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1421 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1422 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1423
1424 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1425
1426 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1427 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1428 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1429 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1430 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1431 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1432 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1433 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1434
1435 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1436
1437 * tree-cfg.c (dump_function_to_file): Dump the return type of
1438 functions, in a line to itself before the function body, mimicking
1439 the layout of a C function.
1440
1441 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1442
1443 PR tree-optimization/60971
1444 * tree-tailcall.c (process_assignment): Reject conversions which
1445 reduce precision.
1446
1447 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1448
1449 * calls.c (initialize_argument_information): Always treat
1450 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1451 (expand_call): Likewise.
1452 (emit_library_call_calue_1): Likewise.
1453 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1454 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1455 code accordingly.
1456
1457 2014-04-29 Nick Clifton <nickc@redhat.com>
1458
1459 * config/msp430/msp430.md (umulsidi): Fix typo.
1460 (mulhisi3): Enable even inside interrupt handlers.
1461 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1462 bigger return address pushed in large mode.
1463
1464 2014-04-29 Nick Clifton <nickc@redhat.com>
1465
1466 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1467 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1468 available modes.
1469 * config/m32r/m32r.c (init_reg_tables): Likewise.
1470 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1471 enum to hold the modes.
1472
1473 2014-04-29 Richard Biener <rguenther@suse.de>
1474
1475 * dominance.c (free_dominance_info): Add overload with
1476 function parameter.
1477 (dom_info_state): Likewise.
1478 (dom_info_available_p): Likewise.
1479 * basic-block.h (free_dominance_info, dom_info_state,
1480 dom_info_available_p): Declare overloads.
1481 * passes.c (execute_function_todo): Verify that verifiers
1482 don't change dominator info state. Drop dominator info
1483 for IPA pass invocations.
1484 * cgraph.c (release_function_body): Restore asserts that
1485 dominator information is released.
1486
1487 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1488
1489 * doc/invoke.texi: Fix typo.
1490 * tree-vrp.c: Fix typos.
1491 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1492 condition.
1493
1494 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1495
1496 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1497
1498 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1499
1500 * config/aarch64/aarch64-builtins.c
1501 (aarch64_types_storestruct_lane_qualifiers): New.
1502 (TYPES_STORESTRUCT_LANE): Likewise.
1503 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1504 (st3_lane): Likewise.
1505 (st4_lane): Likewise.
1506 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1507 (vec_store_lanesci_lane<mode>): Likewise.
1508 (vec_store_lanesxi_lane<mode>): Likewise.
1509 (aarch64_st2_lane<VQ:mode>): Likewise.
1510 (aarch64_st3_lane<VQ:mode>): Likewise.
1511 (aarch64_st4_lane<VQ:mode>): Likewise.
1512 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1513 * config/aarch64/arm_neon.h
1514 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1515 use new macro arguments.
1516 (__ST3_LANE_FUNC): Likewise.
1517 (__ST4_LANE_FUNC): Likewise.
1518 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1519 (V_THREE_ELEM): Likewise.
1520 (V_FOUR_ELEM): Likewise.
1521
1522 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1523
1524 * doc/gimple.texi: Replace the description of the now-defunct
1525 union gimple_statement_d with a diagram showing the
1526 gimple_statement_base class hierarchy and its relationships to
1527 the GSS_ and GIMPLE_ enums.
1528
1529 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1530
1531 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1532 * config/aarch64/aarch64.c
1533 (aarch64_cannot_change_mode_class): Weaken conditions.
1534 (aarch64_modes_tieable_p): New.
1535 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1536
1537 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1538
1539 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1540 (loadsync_<mode>): Change mode.
1541 (load_quadpti, store_quadpti): New.
1542 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1543 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1544
1545 2014-04-28 Martin Jambor <mjambor@suse.cz>
1546
1547 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1548 same alias type as the original statement.
1549 (subreplacement_assignment_data): New type.
1550 (handle_unscalarized_data_in_subtree): New type of parameter,
1551 generate new memory accesses with same alias type as the original
1552 statement.
1553 (load_assign_lhs_subreplacements): Likewise.
1554 (sra_modify_constructor_assign): Generate new memory accesses with
1555 same alias type as the original statement.
1556
1557 2014-04-28 Richard Biener <rguenther@suse.de>
1558
1559 * tree-pass.h (TODO_verify_il): Define.
1560 (TODO_verify_all): Complete properly.
1561 * passes.c (execute_function_todo): Move existing loop-closed
1562 SSA verification under TODO_verify_il.
1563 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1564 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1565 Fix tree sharing issue.
1566
1567 2014-04-28 Richard Biener <rguenther@suse.de>
1568
1569 PR middle-end/60092
1570 * builtins.def (DEF_C11_BUILTIN): Add.
1571 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1572 * coretypes.h (enum function_class): Add function_c11_misc.
1573 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1574 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1575 (call_may_clobber_ref_p_1): Likewise.
1576 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1577 (mark_all_reaching_defs_necessary_1): Likewise.
1578 (propagate_necessity): Likewise.
1579 (eliminate_unnecessary_stmts): Likewise.
1580 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1581
1582 2014-04-28 Richard Biener <rguenther@suse.de>
1583
1584 * tree-vrp.c (vrp_var_may_overflow): Remove.
1585 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1586 with overflow immediately bump to one before that value and
1587 let iteration figure out overflow status.
1588
1589 2014-04-28 Richard Biener <rguenther@suse.de>
1590
1591 * configure.ac: Do valgrind header checks unconditionally.
1592 Add --enable-valgrind-annotations.
1593 * system.h: Guard valgrind header inclusion with
1594 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1595 * alloc-pool.c (pool_alloc, pool_free): Use
1596 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1597 to guard possibly dead code.
1598 * config.in: Regenerated.
1599 * configure: Likewise.
1600
1601 2014-04-28 Jeff Law <law@redhat.com>
1602
1603 PR tree-optimization/60902
1604 * tree-ssa-threadedge.c
1605 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1606 over real defs when invalidating outputs from statements that do not
1607 produce useful outputs for threading.
1608
1609 2014-04-28 Richard Biener <rguenther@suse.de>
1610
1611 PR tree-optimization/60979
1612 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1613 SCOPs that end in a block with a successor with abnormal
1614 predecessors.
1615
1616 2014-04-28 Richard Biener <rguenther@suse.de>
1617
1618 * tree-pass.h (execute_pass_list): Adjust prototype.
1619 * passes.c (pass_manager::execute_early_local_passes):
1620 Adjust.
1621 (do_per_function): Change callback signature, push all actual
1622 work to the callbals.
1623 (do_per_function_toporder): Likewise.
1624 (execute_function_dump): Adjust.
1625 (execute_function_todo): Likewise.
1626 (clear_last_verified): Likewise.
1627 (verify_curr_properties): Likewise.
1628 (update_properties_after_pass): Likewise.
1629 (execute_pass_list_1): Split out from ...
1630 (execute_pass_list): ... here. Adjust.
1631 (execute_ipa_pass_list): Likewise.
1632 * cgraphunit.c (cgraph_add_new_function): Adjust.
1633 (analyze_function): Likewise.
1634 (expand_function): Likewise.
1635 * cgraph.c (release_function_body): Free dominance info
1636 here instead of asserting it was magically freed elsewhere.
1637
1638 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1639
1640 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1641 * configure: Regenerate.
1642 * config/sparc/sparc.opt (muser-mode): New option.
1643 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1644 for LEON3.
1645 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1646 * doc/invoke.texi (SPARC options): Document -muser-mode.
1647
1648 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1649
1650 * cselib.c (find_slot_memmode): Delete.
1651 (cselib_hasher): Change compare_type to a struct.
1652 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1653 constants.
1654 (preserve_constants_and_equivs): Adjust for new compare_type.
1655 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1656 (wrap_constant): Delete.
1657 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1658
1659 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1660
1661 * doc/install.texi (Building with profile feedback): Remove
1662 outdated sentence.
1663
1664 2014-04-26 Tom de Vries <tom@codesourcery.com>
1665
1666 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1667 array accesses.
1668
1669 2014-04-25 Cary Coutant <ccoutant@google.com>
1670
1671 PR debug/60929
1672 * dwarf2out.c (should_move_die_to_comdat): A type definition
1673 can contain a subprogram definition, but don't move it to a
1674 comdat unit.
1675 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1676 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1677 from original DIE.
1678 (clone_tree_hash): Rename to...
1679 (clone_tree_partial): ...this; change callers. Copy
1680 DW_TAG_subprogram DIEs as declarations.
1681 (copy_decls_walk): Don't copy children of a declaration into a
1682 type unit.
1683
1684 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1685
1686 PR target/60969
1687 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1688 alternative 12.
1689
1690 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1691
1692 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1693 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1694 reg for long_call.
1695 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1696 restriction.
1697
1698 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1699
1700 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1701 T16-related fields.
1702
1703 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1704
1705 PR tree-optimization/60930
1706 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1707 creating a multiply candidate by folding two constant
1708 multiplicands when the result overflows.
1709
1710 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1711
1712 PR tree-optimization/60960
1713 * tree-vect-generic.c (expand_vector_operation): Only call
1714 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1715
1716 2014-04-25 Tom de Vries <tom@codesourcery.com>
1717
1718 * expr.c (clobber_reg_mode): New function.
1719 * expr.h (clobber_reg): New function.
1720
1721 2014-04-25 Tom de Vries <tom@codesourcery.com>
1722
1723 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1724 clobbers.
1725
1726 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1727 Tom de Vries <tom@codesourcery.com>
1728
1729 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1730 handle.
1731 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1732 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1733 new argument to find_all_hard_reg_sets call.
1734
1735 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1736
1737 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1738 Use HOST_WIDE_INT_C for mask literal.
1739 (aarch_rev16_shleft_mask_imm_p): Likewise.
1740
1741 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1742
1743 PR target/60941
1744 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1745
1746 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1747
1748 PR preprocessor/56540
1749 * config/i386/i386-c.c (ix86_target_macros): Define
1750 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1751
1752 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1753
1754 * configure.ac (tga_func): Remove.
1755 (LIB_TLS_SPEC): Remove.
1756 * configure: Regenerate.
1757 * config.in: Regenerate.
1758 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1759
1760 2014-04-25 Richard Biener <rguenther@suse.de>
1761
1762 PR ipa/60912
1763 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1764 call stmt use/clobber sets during stmt walk instead of
1765 walking the possibly incomplete set of caller edges.
1766
1767 2014-04-25 Richard Biener <rguenther@suse.de>
1768
1769 PR ipa/60911
1770 * passes.c (apply_ipa_transforms): Inline into only caller ...
1771 (execute_one_pass): ... here. Properly bring in function
1772 bodies for nodes we want to apply IPA transforms to.
1773
1774 2014-04-24 Cong Hou <congh@google.com>
1775
1776 PR tree-optimization/60896
1777 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1778 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1779 (vect_mark_pattern_stmts): Set the def type of all statements in
1780 PATTERN_DEF_SEQ as vect_internal_def.
1781
1782 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1783
1784 * doc/extend.texi (PowerPC Built-in Functions): Document new
1785 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1786 functions.
1787 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1788
1789 * config/rs6000/predicates.md (const_0_to_3_operand): New
1790 predicate to match 0..3 integer constants.
1791
1792 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1793 to support adding miscellaneous builtin functions.
1794 (BU_DFP_MISC_2): Likewise.
1795 (BU_P7_MISC_1): Likewise.
1796 (BU_P7_MISC_2): Likewise.
1797 (BU_P8V_MISC_3): Likewise.
1798 (BU_MISC_1): Likewise.
1799 (BU_MISC_2): Likewise.
1800 (DIVWE): Add extended divide builtin functions.
1801 (DIVWEO): Likewise.
1802 (DIVWEU): Likewise.
1803 (DIVWEUO): Likewise.
1804 (DIVDE): Likewise.
1805 (DIVDEO): Likewise.
1806 (DIVDEU): Likewise.
1807 (DIVDEUO): Likewise.
1808 (DXEX): Add decimal floating-point builtin functions.
1809 (DXEXQ): Likewise.
1810 (DDEDPD): Likewise.
1811 (DDEDPDQ): Likewise.
1812 (DENBCD): Likewise.
1813 (DENBCDQ): Likewise.
1814 (DIEX): Likewise.
1815 (DIEXQ): Likewise.
1816 (DSCLI): Likewise.
1817 (DSCLIQ): Likewise.
1818 (DSCRI): Likewise.
1819 (DSCRIQ): Likewise.
1820 (CDTBCD): Add new BCD builtin functions.
1821 (CBCDTD): Likewise.
1822 (ADDG6S): Likewise.
1823 (BCDADD): Likewise.
1824 (BCDADD_LT): Likewise.
1825 (BCDADD_EQ): Likewise.
1826 (BCDADD_GT): Likewise.
1827 (BCDADD_OV): Likewise.
1828 (BCDSUB): Likewise.
1829 (BCDSUB_LT): Likewise.
1830 (BCDSUB_EQ): Likewise.
1831 (BCDSUB_GT): Likewise.
1832 (BCDSUB_OV): Likewise.
1833 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1834 (UNPACK_TD): Likewise.
1835 (PACK_TF): Likewise.
1836 (UNPACK_TF): Likewise.
1837 (UNPACK_TF_0): Likewise.
1838 (UNPACK_TF_1): Likewise.
1839 (PACK_V1TI): Likewise.
1840 (UNPACK_V1TI): Likewise.
1841
1842 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1843 support for decimal floating point builtin functions.
1844 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1845 functions that take constant arguments.
1846 (rs6000_invalid_builtin): Add decimal floating point builtin
1847 support.
1848 (rs6000_init_builtins): Setup long double, _Decimal64, and
1849 _Decimal128 types for new builtin functions.
1850 (builtin_function_type): Set the unsigned flags appropriately for
1851 the new builtin functions.
1852 (rs6000_opt_masks): Add support for decimal floating point builtin
1853 functions.
1854
1855 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1856 floating point builtin functions.
1857 (RS6000_BTM_COMMON): Likewise.
1858 (RS6000_BTI_long_double): Likewise.
1859 (RS6000_BTI_dfloat64): Likewise.
1860 (RS6000_BTI_dfloat128): Likewise.
1861 (long_double_type_internal_node): Likewise.
1862 (dfloat64_type_internal_node): Likewise.
1863 (dfloat128_type_internal_node): Likewise.
1864
1865 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1866 2.07 bcd arithmetic instructions.
1867 (UNSPEC_BCDSUB): Likewise.
1868 (UNSPEC_BCD_OVERFLOW): Likewise.
1869 (UNSPEC_BCD_ADD_SUB): Likewise.
1870 (bcd_add_sub): Likewise.
1871 (BCD_TEST): Likewise.
1872 (bcd<bcd_add_sub>): Likewise.
1873 (bcd<bcd_add_sub>_test): Likewise.
1874 (bcd<bcd_add_sub>_test2): Likewise.
1875 (bcd<bcd_add_sub>_<code>): Likewise.
1876 (peephole2 for combined bcd ops): Likewise.
1877
1878 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1879 decimal floating point builtin functions.
1880 (UNSPEC_DENBCD): Likewise.
1881 (UNSPEC_DXEX): Likewise.
1882 (UNSPEC_DIEX): Likewise.
1883 (UNSPEC_DSCLI): Likewise.
1884 (UNSPEC_DSCRI): Likewise.
1885 (D64_D128): Likewise.
1886 (dfp_suffix): Likewise.
1887 (dfp_ddedpd_<mode>): Likewise.
1888 (dfp_denbcd_<mode>): Likewise.
1889 (dfp_dxex_<mode>): Likewise.
1890 (dfp_diex_<mode>): Likewise.
1891 (dfp_dscli_<mode>): Likewise.
1892 (dfp_dscri_<mode>): Likewise.
1893
1894 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1895 builtin functions.
1896 (UNSPEC_CDTBCD): Likewise.
1897 (UNSPEC_CBCDTD): Likewise.
1898 (UNSPEC_DIVE): Add support for new extended divide builtin
1899 functions.
1900 (UNSPEC_DIVEO): Likewise.
1901 (UNSPEC_DIVEU): Likewise.
1902 (UNSPEC_DIVEUO): Likewise.
1903 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1904 pack/unpack 128-bit types.
1905 (UNSPEC_PACK_128BIT): Likewise.
1906 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1907 (udiv<mode>3): Use idiv_ldiv mode attribute.
1908 (div<mode>3): Likewise.
1909 (addg6s): Add new BCD builtin functions.
1910 (cdtbcd): Likewise.
1911 (cbcdtd): Likewise.
1912 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1913 instructions.
1914 (div_extend): Likewise.
1915 (div<div_extend>_<mode>"): Likewise.
1916 (FP128_64): Add support for new builtin functions to pack/unpack
1917 128-bit types.
1918 (unpack<mode>): Likewise.
1919 (unpacktf_0): Likewise.
1920 (unpacktf_1): Likewise.
1921 (unpack<mode>_dm): Likewise.
1922 (unpack<mode>_nodm): Likewise.
1923 (pack<mode>): Likewise.
1924 (unpackv1ti): Likewise.
1925 (packv1ti): Likewise.
1926
1927 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1928
1929 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1930 is disabled.
1931
1932 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1933
1934 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1935 * gimplify.c (omp_is_private): Change last argument's type to int.
1936 Only diagnose lastprivate if the simd argument is 1, only diagnose
1937 linear if the simd argument is 2.
1938 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1939 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1940 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1941 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1942 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1943 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1944 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1945 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1946 * tree-nested.c (convert_nonlocal_omp_clauses,
1947 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1948
1949 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1950
1951 PR target/60822
1952 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1953 operand 1.
1954
1955 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1956
1957 * flag-types.h (enum ivar_visibility): Add.
1958
1959 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1960
1961 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1962 function * argument.
1963
1964 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1965
1966 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1967
1968 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1969 Tom de Vries <tom@codesourcery.com>
1970
1971 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1972 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1973 reg-note.
1974 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1975 * emit-rtl.c (try_split): Same.
1976
1977 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1978 Tom de Vries <tom@codesourcery.com>
1979
1980 * common.opt (fuse-caller-save): New option.
1981
1982 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1983
1984 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1985 for big-endian.
1986
1987 2014-04-24 Richard Biener <rguenther@suse.de>
1988
1989 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1990 during TER and instead use the sepops interface for expanding
1991 non-GIMPLE_SINGLE_RHS.
1992
1993 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1994
1995 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1996 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1997
1998 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1999
2000 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2001 assembler 64-bit option.
2002 * configure: Regenerate.
2003
2004 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2005
2006 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2007 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2008 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2009 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2010 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2011
2012 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2013
2014 * config/aarch64/aarch64-builtins.c
2015 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2016 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2017 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2018 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2019 builtins.
2020 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2021 (Vrevsuff): New mode attribute.
2022
2023 2014-04-24 Terry Guo <terry.guo@arm.com>
2024
2025 * config/arm/arm.h (machine_function): Define variable
2026 after_arm_reorg here.
2027 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2028 (arm_split_constant): Update the way to access variable
2029 after_arm_reorg.
2030 (arm_reorg): Ditto.
2031 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2032
2033 2014-04-23 Tom de Vries <tom@codesourcery.com>
2034
2035 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2036
2037 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2038
2039 * is-a.h: Update comments to reflect the following changes to the
2040 "pointerness" of the API, making the template parameter match the
2041 return type, allowing use of is-a.h with typedefs of pointers.
2042 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2043 that the return type matches the parameter to the is_a_helper.
2044 (as_a): Likewise.
2045 (dyn_cast): Likewise.
2046
2047 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2048 pointer from the is-a.h API.
2049
2050 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2051 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2052 is-a.h API.
2053 (is_a_helper <varpool_node>::test): Likewise, convert to...
2054 (is_a_helper <varpool_node *>::test): ...this.
2055
2056 (varpool_first_variable): Update for removal of implicit pointer
2057 from the is-a.h API.
2058 (varpool_next_variable): Likewise.
2059 (varpool_first_static_initializer): Likewise.
2060 (varpool_next_static_initializer): Likewise.
2061 (varpool_first_defined_variable): Likewise.
2062 (varpool_next_defined_variable): Likewise.
2063 (cgraph_first_defined_function): Likewise.
2064 (cgraph_next_defined_function): Likewise.
2065 (cgraph_first_function): Likewise.
2066 (cgraph_next_function): Likewise.
2067 (cgraph_first_function_with_gimple_body): Likewise.
2068 (cgraph_next_function_with_gimple_body): Likewise.
2069 (cgraph_alias_target): Likewise.
2070 (varpool_alias_target): Likewise.
2071 (cgraph_function_or_thunk_node): Likewise.
2072 (varpool_variable_node): Likewise.
2073 (symtab_real_symbol_p): Likewise.
2074 * cgraphunit.c (referred_to_p): Likewise.
2075 (analyze_functions): Likewise.
2076 (handle_alias_pairs): Likewise.
2077 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2078 * gimple-ssa.h (gimple_vuse_op): Likewise.
2079 (gimple_vdef_op): Likewise.
2080 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2081 * gimple.c (gimple_build_asm_1): Likewise.
2082 (gimple_build_try): Likewise.
2083 (gimple_build_resx): Likewise.
2084 (gimple_build_eh_dispatch): Likewise.
2085 (gimple_build_omp_for): Likewise.
2086 (gimple_omp_for_set_clauses): Likewise.
2087
2088 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2089 (is_a_helper <gimple_statement_asm *>::test): ...this.
2090 (is_a_helper <gimple_statement_bind>::test): Convert to...
2091 (is_a_helper <gimple_statement_bind *>::test): ...this.
2092 (is_a_helper <gimple_statement_call>::test): Convert to...
2093 (is_a_helper <gimple_statement_call *>::test): ...this.
2094 (is_a_helper <gimple_statement_catch>::test): Convert to...
2095 (is_a_helper <gimple_statement_catch *>::test): ...this.
2096 (is_a_helper <gimple_statement_resx>::test): Convert to...
2097 (is_a_helper <gimple_statement_resx *>::test): ...this.
2098 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2099 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2100 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2101 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2102 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2103 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2104 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2105 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2106 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2107 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2108 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2109 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2110 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2111 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2112 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2113 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2114 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2115 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2116 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2117 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2118 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2119 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2120 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2121 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2122 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2123 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2124 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2125 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2126 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2127 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2128 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2129 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2130 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2131 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2132 (is_a_helper <gimple_statement_phi>::test): Convert to...
2133 (is_a_helper <gimple_statement_phi *>::test): ...this.
2134 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2135 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2136 (is_a_helper <gimple_statement_try>::test): Convert to...
2137 (is_a_helper <gimple_statement_try *>::test): ...this.
2138 (is_a_helper <gimple_statement_wce>::test): Convert to...
2139 (is_a_helper <gimple_statement_wce *>::test): ...this.
2140 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2141 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2142 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2143 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2144 (is_a_helper <const gimple_statement_call>::test): Convert to...
2145 (is_a_helper <const gimple_statement_call *>::test): ...this.
2146 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2147 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2148 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2149 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2150 (is_a_helper <const gimple_statement_eh_dispatch>::test):
2151 Convert to...
2152 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
2153 ...this.
2154 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
2155 to...
2156 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2157 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2158 Convert to...
2159 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2160 ...this.
2161 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2162 Convert to...
2163 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2164 ...this.
2165 (is_a_helper <const gimple_statement_omp_return>::test): Convert
2166 to...
2167 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2168 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2169 to...
2170 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2171 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2172 to...
2173 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2174 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2175 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2176 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2177 to...
2178 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2179 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2180 to...
2181 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2182 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2183 to...
2184 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2185 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2186 to...
2187 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2188 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2189 to...
2190 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2191 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2192 to...
2193 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2194 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2195 to...
2196 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2197 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2198 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2199 (is_a_helper <const gimple_statement_transaction>::test): Convert
2200 to...
2201 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2202 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2203 to...
2204 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2205 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2206 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2207 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2208 Convert to...
2209 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2210 ...this.
2211 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2212 to...
2213 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2214
2215 (gimple_use_ops): Update for removal of implicit pointer from the
2216 is-a.h API.
2217 (gimple_set_use_ops): Likewise.
2218 (gimple_vuse): Likewise.
2219 (gimple_vdef): Likewise.
2220 (gimple_vuse_ptr): Likewise.
2221 (gimple_vdef_ptr): Likewise.
2222 (gimple_set_vuse): Likewise.
2223 (gimple_set_vdef): Likewise.
2224 (gimple_omp_return_set_lhs): Likewise.
2225 (gimple_omp_return_lhs): Likewise.
2226 (gimple_omp_return_lhs_ptr): Likewise.
2227 (gimple_call_fntype): Likewise.
2228 (gimple_call_set_fntype): Likewise.
2229 (gimple_call_set_internal_fn): Likewise.
2230 (gimple_call_use_set): Likewise.
2231 (gimple_call_clobber_set): Likewise.
2232 (gimple_bind_vars): Likewise.
2233 (gimple_bind_set_vars): Likewise.
2234 (gimple_bind_body_ptr): Likewise.
2235 (gimple_bind_set_body): Likewise.
2236 (gimple_bind_add_stmt): Likewise.
2237 (gimple_bind_block): Likewise.
2238 (gimple_bind_set_block): Likewise.
2239 (gimple_asm_ninputs): Likewise.
2240 (gimple_asm_noutputs): Likewise.
2241 (gimple_asm_nclobbers): Likewise.
2242 (gimple_asm_nlabels): Likewise.
2243 (gimple_asm_input_op): Likewise.
2244 (gimple_asm_input_op_ptr): Likewise.
2245 (gimple_asm_output_op): Likewise.
2246 (gimple_asm_output_op_ptr): Likewise.
2247 (gimple_asm_set_output_op): Likewise.
2248 (gimple_asm_clobber_op): Likewise.
2249 (gimple_asm_set_clobber_op): Likewise.
2250 (gimple_asm_label_op): Likewise.
2251 (gimple_asm_set_label_op): Likewise.
2252 (gimple_asm_string): Likewise.
2253 (gimple_catch_types): Likewise.
2254 (gimple_catch_types_ptr): Likewise.
2255 (gimple_catch_handler_ptr): Likewise.
2256 (gimple_catch_set_types): Likewise.
2257 (gimple_catch_set_handler): Likewise.
2258 (gimple_eh_filter_types): Likewise.
2259 (gimple_eh_filter_types_ptr): Likewise.
2260 (gimple_eh_filter_failure_ptr): Likewise.
2261 (gimple_eh_filter_set_types): Likewise.
2262 (gimple_eh_filter_set_failure): Likewise.
2263 (gimple_eh_must_not_throw_fndecl): Likewise.
2264 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2265 (gimple_eh_else_n_body_ptr): Likewise.
2266 (gimple_eh_else_e_body_ptr): Likewise.
2267 (gimple_eh_else_set_n_body): Likewise.
2268 (gimple_eh_else_set_e_body): Likewise.
2269 (gimple_try_eval_ptr): Likewise.
2270 (gimple_try_cleanup_ptr): Likewise.
2271 (gimple_try_set_eval): Likewise.
2272 (gimple_try_set_cleanup): Likewise.
2273 (gimple_wce_cleanup_ptr): Likewise.
2274 (gimple_wce_set_cleanup): Likewise.
2275 (gimple_phi_capacity): Likewise.
2276 (gimple_phi_num_args): Likewise.
2277 (gimple_phi_result): Likewise.
2278 (gimple_phi_result_ptr): Likewise.
2279 (gimple_phi_set_result): Likewise.
2280 (gimple_phi_arg): Likewise.
2281 (gimple_phi_set_arg): Likewise.
2282 (gimple_resx_region): Likewise.
2283 (gimple_resx_set_region): Likewise.
2284 (gimple_eh_dispatch_region): Likewise.
2285 (gimple_eh_dispatch_set_region): Likewise.
2286 (gimple_omp_critical_name): Likewise.
2287 (gimple_omp_critical_name_ptr): Likewise.
2288 (gimple_omp_critical_set_name): Likewise.
2289 (gimple_omp_for_clauses): Likewise.
2290 (gimple_omp_for_clauses_ptr): Likewise.
2291 (gimple_omp_for_set_clauses): Likewise.
2292 (gimple_omp_for_collapse): Likewise.
2293 (gimple_omp_for_index): Likewise.
2294 (gimple_omp_for_index_ptr): Likewise.
2295 (gimple_omp_for_set_index): Likewise.
2296 (gimple_omp_for_initial): Likewise.
2297 (gimple_omp_for_initial_ptr): Likewise.
2298 (gimple_omp_for_set_initial): Likewise.
2299 (gimple_omp_for_final): Likewise.
2300 (gimple_omp_for_final_ptr): Likewise.
2301 (gimple_omp_for_set_final): Likewise.
2302 (gimple_omp_for_incr): Likewise.
2303 (gimple_omp_for_incr_ptr): Likewise.
2304 (gimple_omp_for_set_incr): Likewise.
2305 (gimple_omp_for_pre_body_ptr): Likewise.
2306 (gimple_omp_for_set_pre_body): Likewise.
2307 (gimple_omp_parallel_clauses): Likewise.
2308 (gimple_omp_parallel_clauses_ptr): Likewise.
2309 (gimple_omp_parallel_set_clauses): Likewise.
2310 (gimple_omp_parallel_child_fn): Likewise.
2311 (gimple_omp_parallel_child_fn_ptr): Likewise.
2312 (gimple_omp_parallel_set_child_fn): Likewise.
2313 (gimple_omp_parallel_data_arg): Likewise.
2314 (gimple_omp_parallel_data_arg_ptr): Likewise.
2315 (gimple_omp_parallel_set_data_arg): Likewise.
2316 (gimple_omp_task_clauses): Likewise.
2317 (gimple_omp_task_clauses_ptr): Likewise.
2318 (gimple_omp_task_set_clauses): Likewise.
2319 (gimple_omp_task_child_fn): Likewise.
2320 (gimple_omp_task_child_fn_ptr): Likewise.
2321 (gimple_omp_task_set_child_fn): Likewise.
2322 (gimple_omp_task_data_arg): Likewise.
2323 (gimple_omp_task_data_arg_ptr): Likewise.
2324 (gimple_omp_task_set_data_arg): Likewise.
2325 (gimple_omp_taskreg_clauses): Likewise.
2326 (gimple_omp_taskreg_clauses_ptr): Likewise.
2327 (gimple_omp_taskreg_set_clauses): Likewise.
2328 (gimple_omp_taskreg_child_fn): Likewise.
2329 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2330 (gimple_omp_taskreg_set_child_fn): Likewise.
2331 (gimple_omp_taskreg_data_arg): Likewise.
2332 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2333 (gimple_omp_taskreg_set_data_arg): Likewise.
2334 (gimple_omp_task_copy_fn): Likewise.
2335 (gimple_omp_task_copy_fn_ptr): Likewise.
2336 (gimple_omp_task_set_copy_fn): Likewise.
2337 (gimple_omp_task_arg_size): Likewise.
2338 (gimple_omp_task_arg_size_ptr): Likewise.
2339 (gimple_omp_task_set_arg_size): Likewise.
2340 (gimple_omp_task_arg_align): Likewise.
2341 (gimple_omp_task_arg_align_ptr): Likewise.
2342 (gimple_omp_task_set_arg_align): Likewise.
2343 (gimple_omp_single_clauses): Likewise.
2344 (gimple_omp_single_clauses_ptr): Likewise.
2345 (gimple_omp_single_set_clauses): Likewise.
2346 (gimple_omp_target_clauses): Likewise.
2347 (gimple_omp_target_clauses_ptr): Likewise.
2348 (gimple_omp_target_set_clauses): Likewise.
2349 (gimple_omp_target_child_fn): Likewise.
2350 (gimple_omp_target_child_fn_ptr): Likewise.
2351 (gimple_omp_target_set_child_fn): Likewise.
2352 (gimple_omp_target_data_arg): Likewise.
2353 (gimple_omp_target_data_arg_ptr): Likewise.
2354 (gimple_omp_target_set_data_arg): Likewise.
2355 (gimple_omp_teams_clauses): Likewise.
2356 (gimple_omp_teams_clauses_ptr): Likewise.
2357 (gimple_omp_teams_set_clauses): Likewise.
2358 (gimple_omp_sections_clauses): Likewise.
2359 (gimple_omp_sections_clauses_ptr): Likewise.
2360 (gimple_omp_sections_set_clauses): Likewise.
2361 (gimple_omp_sections_control): Likewise.
2362 (gimple_omp_sections_control_ptr): Likewise.
2363 (gimple_omp_sections_set_control): Likewise.
2364 (gimple_omp_for_set_cond): Likewise.
2365 (gimple_omp_for_cond): Likewise.
2366 (gimple_omp_atomic_store_set_val): Likewise.
2367 (gimple_omp_atomic_store_val): Likewise.
2368 (gimple_omp_atomic_store_val_ptr): Likewise.
2369 (gimple_omp_atomic_load_set_lhs): Likewise.
2370 (gimple_omp_atomic_load_lhs): Likewise.
2371 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2372 (gimple_omp_atomic_load_set_rhs): Likewise.
2373 (gimple_omp_atomic_load_rhs): Likewise.
2374 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2375 (gimple_omp_continue_control_def): Likewise.
2376 (gimple_omp_continue_control_def_ptr): Likewise.
2377 (gimple_omp_continue_set_control_def): Likewise.
2378 (gimple_omp_continue_control_use): Likewise.
2379 (gimple_omp_continue_control_use_ptr): Likewise.
2380 (gimple_omp_continue_set_control_use): Likewise.
2381 (gimple_transaction_body_ptr): Likewise.
2382 (gimple_transaction_label): Likewise.
2383 (gimple_transaction_label_ptr): Likewise.
2384 (gimple_transaction_set_body): Likewise.
2385 (gimple_transaction_set_label): Likewise.
2386
2387 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2388 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2389 * ipa-ref.c (ipa_record_reference): Likewise.
2390 * ipa-reference.c (analyze_function): Likewise.
2391 (ipa_reference_write_optimization_summary): Likewise.
2392 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2393 (address_taken_from_non_vtable_p): Likewise.
2394 (comdat_can_be_unshared_p_1): Likewise.
2395 * lto-cgraph.c (lto_output_ref): Likewise.
2396 (add_references): Likewise.
2397 (compute_ltrans_boundary): Likewise.
2398 (output_symtab): Likewise.
2399 (input_ref): Likewise.
2400 (input_cgraph_1): Likewise.
2401 (output_cgraph_opt_summary): Likewise.
2402 * lto-streamer-out.c (lto_output): Likewise.
2403 (output_symbol_p): Likewise.
2404 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2405 (lsei_start_function_in_partition): Likewise.
2406 (lsei_next_variable_in_partition): Likewise.
2407 (lsei_start_variable_in_partition): Likewise.
2408 * symtab.c (insert_to_assembler_name_hash): Likewise.
2409 (unlink_from_assembler_name_hash): Likewise.
2410 (symtab_unregister_node): Likewise.
2411 (symtab_remove_node): Likewise.
2412 (dump_symtab_node): Likewise.
2413 (verify_symtab_base): Likewise.
2414 (verify_symtab_node): Likewise.
2415 (symtab_make_decl_local): Likewise.
2416 (symtab_alias_ultimate_target): Likewise.
2417 (symtab_resolve_alias): Likewise.
2418 (symtab_get_symbol_partitioning_class): Likewise.
2419 * tree-phinodes.c (allocate_phi_node): Likewise.
2420 (reserve_phi_args_for_new_edge): Likewise.
2421 (remove_phi_args): Likewise.
2422 * varpool.c (varpool_node_for_asm): Likewise.
2423 (varpool_remove_unreferenced_decls): Likewise.
2424
2425 2014-04-23 Jeff Law <law@redhat.com>
2426
2427 PR tree-optimization/60902
2428 * tree-ssa-threadedge.c
2429 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2430 invalidate outputs from statements that do not produce useful
2431 outputs for threading.
2432
2433 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2434
2435 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2436 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2437 machine descriptions for Stack Smashing Protector.
2438
2439 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2440
2441 * aarch64.md (<optab>_rol<mode>3): New pattern.
2442 (<optab>_rolsi3_uxtw): Likewise.
2443 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2444
2445 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2446
2447 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2448 (arm_cortex_a12_tune): Likewise.
2449
2450 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2451
2452 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2453
2454 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2455
2456 * config/arm/arm.md (arm_rev16si2): New pattern.
2457 (arm_rev16si2_alt): Likewise.
2458 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2459
2460 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2461
2462 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2463 (rev16<mode>2_alt): Likewise.
2464 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2465 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2466 (aarch_rev16_shleft_mask_imm_p): Likewise.
2467 (aarch_rev16_p_1): Likewise.
2468 (aarch_rev16_p): Likewise.
2469 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2470 (aarch_rev16_shright_mask_imm_p): Likewise.
2471 (aarch_rev16_shleft_mask_imm_p): Likewise.
2472
2473 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2474
2475 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2476 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2477 rev cost.
2478 (cortex_a53_extra_costs): Likewise.
2479 (cortex_a57_extra_costs): Likewise.
2480 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2481 (cortexa7_extra_costs): Likewise.
2482 (cortexa8_extra_costs): Likewise.
2483 (cortexa12_extra_costs): Likewise.
2484 (cortexa15_extra_costs): Likewise.
2485 (v7m_extra_costs): Likewise.
2486 (arm_new_rtx_costs): Handle BSWAP.
2487
2488 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2489
2490 * config/arm/arm.c (cortexa8_extra_costs): New table.
2491 (arm_cortex_a8_tune): New tuning struct.
2492 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2493
2494 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2495
2496 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2497
2498 2014-04-23 Richard Biener <rguenther@suse.de>
2499
2500 * Makefile.in (OBJS): Remove loop-unswitch.o.
2501 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2502 * passes.def (pass_rtl_unswitch): Likewise.
2503 * loop-init.c (gate_rtl_unswitch): Likewise.
2504 (rtl_unswitch): Likewise.
2505 (pass_data_rtl_unswitch): Likewise.
2506 (pass_rtl_unswitch): Likewise.
2507 (make_pass_rtl_unswitch): Likewise.
2508 * rtl.h (reversed_condition): Likewise.
2509 (compare_and_jump_seq): Likewise.
2510 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2511 and make static.
2512 * loop-unroll.c (compare_and_jump_seq): Likewise.
2513
2514 2014-04-23 Richard Biener <rguenther@suse.de>
2515
2516 PR tree-optimization/60903
2517 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2518 commented code block.
2519 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2520 loop flags to newly created BBs and edges.
2521
2522 2014-04-23 Nick Clifton <nickc@redhat.com>
2523
2524 * config/msp430/msp430.c (msp430_handle_option): Move function
2525 to msp430-common.c
2526 (msp430_option_override): Simplify mcu and mcpu option handling.
2527 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2528 support for -mhwmult command line option.
2529 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2530 -mhwmult command line option.
2531 (msp430_hwmult_enabled): Delete.
2532 (msp43o_output_labelref): Add support for -mhwmult command line
2533 option.
2534 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2535 (umulsidi3): Likewise.
2536 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2537 (mcpu, mlarge, msmall): Likewise.
2538 (mhwmult): New option.
2539 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2540 prototype.
2541 (msp430_is_f5_mcu): Remove prototype.
2542 (msp430_use_f5_series_hwmult): Add prototype.
2543 * config/msp430/msp430-opts.h: New file.
2544 * common/config/msp430: New directory.
2545 * common/config/msp430/msp430-common.c: New file.
2546 * config.gcc (msp430): Remove target_has_targetm_common.
2547 * doc/invoke.texi: Document -mhwmult command line option.
2548
2549 2014-04-23 Nick Clifton <nickc@redhat.com>
2550
2551 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2552 default-manifest.o if it can be found in the search path.
2553 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2554
2555 2014-04-23 Terry Guo <terry.guo@arm.com>
2556
2557 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2558
2559 2014-04-23 Richard Biener <rguenther@suse.de>
2560
2561 PR middle-end/60895
2562 * tree-inline.c (declare_return_variable): Use mark_addressable.
2563
2564 2014-04-23 Richard Biener <rguenther@suse.de>
2565
2566 PR middle-end/60891
2567 * loop-init.c (loop_optimizer_init): Make sure to apply
2568 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2569
2570 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2571
2572 PR sanitizer/60275
2573 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2574 New options.
2575 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2576 if flag_sanitize_undefined_trap_on_error.
2577 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2578 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2579 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2580 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2581 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2582 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2583 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2584 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2585 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2586 * ubsan.c (ubsan_instrument_unreachable): Return
2587 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2588 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2589 if flag_sanitize_undefined_trap_on_error and
2590 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2591 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2592 instrument_bool_enum_load): Emit __builtin_trap () if
2593 flag_sanitize_undefined_trap_on_error and
2594 __builtin_handle_*_abort () if !flag_sanitize_recover.
2595 * doc/invoke.texi (-fsanitize-recover,
2596 -fsanitize-undefined-trap-on-error): Document.
2597
2598 2014-04-22 Christian Bruel <christian.bruel@st.com>
2599
2600 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2601 Force immediates to SImode.
2602
2603 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2604
2605 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2606 (lroundsfsi2): New.
2607 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2608 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2609 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2610 (nios2_fpu_insn): Add entry for round.
2611 (N2FPU_NO_ERRNO_P): Define.
2612 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2613 flag_errno_math.
2614 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2615
2616 2014-04-22 Richard Henderson <rth@redhat.com>
2617
2618 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2619 (add<GPI>3_compare0): Remove leading * from name.
2620 (add<GPI>3_carryin): Likewise.
2621 (sub<GPI>3_compare0): Likewise.
2622 (sub<GPI>3_carryin): Likewise.
2623 (<su_optab>mulditi3): New expander.
2624 (multi3): New expander.
2625 (madd<GPI>): Remove leading * from name.
2626
2627 2014-04-22 Martin Jambor <mjambor@suse.cz>
2628
2629 * cgraphclones.c (cgraph_function_versioning): Copy
2630 ipa_transforms_to_apply instead of asserting it is empty.
2631
2632 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2633
2634 PR target/60868
2635 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2636 on count_exp to get mode.
2637
2638 2014-04-22 Andrew Pinski <apinski@cavium.com>
2639
2640 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2641 Handle TLS for ILP32.
2642 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2643 (tlsie_small_<mode>): this and handle PTR.
2644 (tlsie_small_sidi): New pattern.
2645 (tlsle_small): Change to an expand to handle ILP32.
2646 (tlsle_small_<mode>): New pattern.
2647 (tlsdesc_small): Rename to ...
2648 (tlsdesc_small_<mode>): this and handle PTR.
2649
2650 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2651
2652 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2653
2654 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2655
2656 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2657 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2658 (aarch64_types_signed_poly_qualifiers): Likewise.
2659 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2660 (aarch64_types_poly_signed_qualifiers): Likewise.
2661 (TYPES_REINTERP_SS): Type macro added.
2662 (TYPES_REINTERP_SU): Likewise.
2663 (TYPES_REINTERP_SP): Likewise.
2664 (TYPES_REINTERP_US): Likewise.
2665 (TYPES_REINTERP_PS): Likewise.
2666 (aarch64_fold_builtin): New expression folding added.
2667 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2668 Declarations removed.
2669 (REINTERP_SS): Declarations added.
2670 (REINTERP_US): Likewise.
2671 (REINTERP_PS): Likewise.
2672 (REINTERP_SU): Likewise.
2673 (REINTERP_SP): Likewise.
2674 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2675 (vreinterpretq_p8_f64): Likewise.
2676 (vreinterpret_p16_f64): Likewise.
2677 (vreinterpretq_p16_f64): Likewise.
2678 (vreinterpret_f32_f64): Likewise.
2679 (vreinterpretq_f32_f64): Likewise.
2680 (vreinterpret_f64_f32): Likewise.
2681 (vreinterpret_f64_p8): Likewise.
2682 (vreinterpret_f64_p16): Likewise.
2683 (vreinterpret_f64_s8): Likewise.
2684 (vreinterpret_f64_s16): Likewise.
2685 (vreinterpret_f64_s32): Likewise.
2686 (vreinterpret_f64_s64): Likewise.
2687 (vreinterpret_f64_u8): Likewise.
2688 (vreinterpret_f64_u16): Likewise.
2689 (vreinterpret_f64_u32): Likewise.
2690 (vreinterpret_f64_u64): Likewise.
2691 (vreinterpretq_f64_f32): Likewise.
2692 (vreinterpretq_f64_p8): Likewise.
2693 (vreinterpretq_f64_p16): Likewise.
2694 (vreinterpretq_f64_s8): Likewise.
2695 (vreinterpretq_f64_s16): Likewise.
2696 (vreinterpretq_f64_s32): Likewise.
2697 (vreinterpretq_f64_s64): Likewise.
2698 (vreinterpretq_f64_u8): Likewise.
2699 (vreinterpretq_f64_u16): Likewise.
2700 (vreinterpretq_f64_u32): Likewise.
2701 (vreinterpretq_f64_u64): Likewise.
2702 (vreinterpret_s64_f64): Likewise.
2703 (vreinterpretq_s64_f64): Likewise.
2704 (vreinterpret_u64_f64): Likewise.
2705 (vreinterpretq_u64_f64): Likewise.
2706 (vreinterpret_s8_f64): Likewise.
2707 (vreinterpretq_s8_f64): Likewise.
2708 (vreinterpret_s16_f64): Likewise.
2709 (vreinterpretq_s16_f64): Likewise.
2710 (vreinterpret_s32_f64): Likewise.
2711 (vreinterpretq_s32_f64): Likewise.
2712 (vreinterpret_u8_f64): Likewise.
2713 (vreinterpretq_u8_f64): Likewise.
2714 (vreinterpret_u16_f64): Likewise.
2715 (vreinterpretq_u16_f64): Likewise.
2716 (vreinterpret_u32_f64): Likewise.
2717 (vreinterpretq_u32_f64): Likewise.
2718
2719 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2720
2721 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2722 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2723 (vreinterpret_p8_s8): Likewise.
2724 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2725 (vreinterpret_p8_s16): Likewise.
2726 (vreinterpret_p8_s32): Likewise.
2727 (vreinterpret_p8_s64): Likewise.
2728 (vreinterpret_p8_f32): Likewise.
2729 (vreinterpret_p8_u8): Likewise.
2730 (vreinterpret_p8_u16): Likewise.
2731 (vreinterpret_p8_u32): Likewise.
2732 (vreinterpret_p8_u64): Likewise.
2733 (vreinterpret_p8_p16): Likewise.
2734 (vreinterpretq_p8_s8): Likewise.
2735 (vreinterpretq_p8_s16): Likewise.
2736 (vreinterpretq_p8_s32): Likewise.
2737 (vreinterpretq_p8_s64): Likewise.
2738 (vreinterpretq_p8_f32): Likewise.
2739 (vreinterpretq_p8_u8): Likewise.
2740 (vreinterpretq_p8_u16): Likewise.
2741 (vreinterpretq_p8_u32): Likewise.
2742 (vreinterpretq_p8_u64): Likewise.
2743 (vreinterpretq_p8_p16): Likewise.
2744 (vreinterpret_p16_s8): Likewise.
2745 (vreinterpret_p16_s16): Likewise.
2746 (vreinterpret_p16_s32): Likewise.
2747 (vreinterpret_p16_s64): Likewise.
2748 (vreinterpret_p16_f32): Likewise.
2749 (vreinterpret_p16_u8): Likewise.
2750 (vreinterpret_p16_u16): Likewise.
2751 (vreinterpret_p16_u32): Likewise.
2752 (vreinterpret_p16_u64): Likewise.
2753 (vreinterpret_p16_p8): Likewise.
2754 (vreinterpretq_p16_s8): Likewise.
2755 (vreinterpretq_p16_s16): Likewise.
2756 (vreinterpretq_p16_s32): Likewise.
2757 (vreinterpretq_p16_s64): Likewise.
2758 (vreinterpretq_p16_f32): Likewise.
2759 (vreinterpretq_p16_u8): Likewise.
2760 (vreinterpretq_p16_u16): Likewise.
2761 (vreinterpretq_p16_u32): Likewise.
2762 (vreinterpretq_p16_u64): Likewise.
2763 (vreinterpretq_p16_p8): Likewise.
2764 (vreinterpret_f32_s8): Likewise.
2765 (vreinterpret_f32_s16): Likewise.
2766 (vreinterpret_f32_s32): Likewise.
2767 (vreinterpret_f32_s64): Likewise.
2768 (vreinterpret_f32_u8): Likewise.
2769 (vreinterpret_f32_u16): Likewise.
2770 (vreinterpret_f32_u32): Likewise.
2771 (vreinterpret_f32_u64): Likewise.
2772 (vreinterpret_f32_p8): Likewise.
2773 (vreinterpret_f32_p16): Likewise.
2774 (vreinterpretq_f32_s8): Likewise.
2775 (vreinterpretq_f32_s16): Likewise.
2776 (vreinterpretq_f32_s32): Likewise.
2777 (vreinterpretq_f32_s64): Likewise.
2778 (vreinterpretq_f32_u8): Likewise.
2779 (vreinterpretq_f32_u16): Likewise.
2780 (vreinterpretq_f32_u32): Likewise.
2781 (vreinterpretq_f32_u64): Likewise.
2782 (vreinterpretq_f32_p8): Likewise.
2783 (vreinterpretq_f32_p16): Likewise.
2784 (vreinterpret_s64_s8): Likewise.
2785 (vreinterpret_s64_s16): Likewise.
2786 (vreinterpret_s64_s32): Likewise.
2787 (vreinterpret_s64_f32): Likewise.
2788 (vreinterpret_s64_u8): Likewise.
2789 (vreinterpret_s64_u16): Likewise.
2790 (vreinterpret_s64_u32): Likewise.
2791 (vreinterpret_s64_u64): Likewise.
2792 (vreinterpret_s64_p8): Likewise.
2793 (vreinterpret_s64_p16): Likewise.
2794 (vreinterpretq_s64_s8): Likewise.
2795 (vreinterpretq_s64_s16): Likewise.
2796 (vreinterpretq_s64_s32): Likewise.
2797 (vreinterpretq_s64_f32): Likewise.
2798 (vreinterpretq_s64_u8): Likewise.
2799 (vreinterpretq_s64_u16): Likewise.
2800 (vreinterpretq_s64_u32): Likewise.
2801 (vreinterpretq_s64_u64): Likewise.
2802 (vreinterpretq_s64_p8): Likewise.
2803 (vreinterpretq_s64_p16): Likewise.
2804 (vreinterpret_u64_s8): Likewise.
2805 (vreinterpret_u64_s16): Likewise.
2806 (vreinterpret_u64_s32): Likewise.
2807 (vreinterpret_u64_s64): Likewise.
2808 (vreinterpret_u64_f32): Likewise.
2809 (vreinterpret_u64_u8): Likewise.
2810 (vreinterpret_u64_u16): Likewise.
2811 (vreinterpret_u64_u32): Likewise.
2812 (vreinterpret_u64_p8): Likewise.
2813 (vreinterpret_u64_p16): Likewise.
2814 (vreinterpretq_u64_s8): Likewise.
2815 (vreinterpretq_u64_s16): Likewise.
2816 (vreinterpretq_u64_s32): Likewise.
2817 (vreinterpretq_u64_s64): Likewise.
2818 (vreinterpretq_u64_f32): Likewise.
2819 (vreinterpretq_u64_u8): Likewise.
2820 (vreinterpretq_u64_u16): Likewise.
2821 (vreinterpretq_u64_u32): Likewise.
2822 (vreinterpretq_u64_p8): Likewise.
2823 (vreinterpretq_u64_p16): Likewise.
2824 (vreinterpret_s8_s16): Likewise.
2825 (vreinterpret_s8_s32): Likewise.
2826 (vreinterpret_s8_s64): Likewise.
2827 (vreinterpret_s8_f32): Likewise.
2828 (vreinterpret_s8_u8): Likewise.
2829 (vreinterpret_s8_u16): Likewise.
2830 (vreinterpret_s8_u32): Likewise.
2831 (vreinterpret_s8_u64): Likewise.
2832 (vreinterpret_s8_p8): Likewise.
2833 (vreinterpret_s8_p16): Likewise.
2834 (vreinterpretq_s8_s16): Likewise.
2835 (vreinterpretq_s8_s32): Likewise.
2836 (vreinterpretq_s8_s64): Likewise.
2837 (vreinterpretq_s8_f32): Likewise.
2838 (vreinterpretq_s8_u8): Likewise.
2839 (vreinterpretq_s8_u16): Likewise.
2840 (vreinterpretq_s8_u32): Likewise.
2841 (vreinterpretq_s8_u64): Likewise.
2842 (vreinterpretq_s8_p8): Likewise.
2843 (vreinterpretq_s8_p16): Likewise.
2844 (vreinterpret_s16_s8): Likewise.
2845 (vreinterpret_s16_s32): Likewise.
2846 (vreinterpret_s16_s64): Likewise.
2847 (vreinterpret_s16_f32): Likewise.
2848 (vreinterpret_s16_u8): Likewise.
2849 (vreinterpret_s16_u16): Likewise.
2850 (vreinterpret_s16_u32): Likewise.
2851 (vreinterpret_s16_u64): Likewise.
2852 (vreinterpret_s16_p8): Likewise.
2853 (vreinterpret_s16_p16): Likewise.
2854 (vreinterpretq_s16_s8): Likewise.
2855 (vreinterpretq_s16_s32): Likewise.
2856 (vreinterpretq_s16_s64): Likewise.
2857 (vreinterpretq_s16_f32): Likewise.
2858 (vreinterpretq_s16_u8): Likewise.
2859 (vreinterpretq_s16_u16): Likewise.
2860 (vreinterpretq_s16_u32): Likewise.
2861 (vreinterpretq_s16_u64): Likewise.
2862 (vreinterpretq_s16_p8): Likewise.
2863 (vreinterpretq_s16_p16): Likewise.
2864 (vreinterpret_s32_s8): Likewise.
2865 (vreinterpret_s32_s16): Likewise.
2866 (vreinterpret_s32_s64): Likewise.
2867 (vreinterpret_s32_f32): Likewise.
2868 (vreinterpret_s32_u8): Likewise.
2869 (vreinterpret_s32_u16): Likewise.
2870 (vreinterpret_s32_u32): Likewise.
2871 (vreinterpret_s32_u64): Likewise.
2872 (vreinterpret_s32_p8): Likewise.
2873 (vreinterpret_s32_p16): Likewise.
2874 (vreinterpretq_s32_s8): Likewise.
2875 (vreinterpretq_s32_s16): Likewise.
2876 (vreinterpretq_s32_s64): Likewise.
2877 (vreinterpretq_s32_f32): Likewise.
2878 (vreinterpretq_s32_u8): Likewise.
2879 (vreinterpretq_s32_u16): Likewise.
2880 (vreinterpretq_s32_u32): Likewise.
2881 (vreinterpretq_s32_u64): Likewise.
2882 (vreinterpretq_s32_p8): Likewise.
2883 (vreinterpretq_s32_p16): Likewise.
2884 (vreinterpret_u8_s8): Likewise.
2885 (vreinterpret_u8_s16): Likewise.
2886 (vreinterpret_u8_s32): Likewise.
2887 (vreinterpret_u8_s64): Likewise.
2888 (vreinterpret_u8_f32): Likewise.
2889 (vreinterpret_u8_u16): Likewise.
2890 (vreinterpret_u8_u32): Likewise.
2891 (vreinterpret_u8_u64): Likewise.
2892 (vreinterpret_u8_p8): Likewise.
2893 (vreinterpret_u8_p16): Likewise.
2894 (vreinterpretq_u8_s8): Likewise.
2895 (vreinterpretq_u8_s16): Likewise.
2896 (vreinterpretq_u8_s32): Likewise.
2897 (vreinterpretq_u8_s64): Likewise.
2898 (vreinterpretq_u8_f32): Likewise.
2899 (vreinterpretq_u8_u16): Likewise.
2900 (vreinterpretq_u8_u32): Likewise.
2901 (vreinterpretq_u8_u64): Likewise.
2902 (vreinterpretq_u8_p8): Likewise.
2903 (vreinterpretq_u8_p16): Likewise.
2904 (vreinterpret_u16_s8): Likewise.
2905 (vreinterpret_u16_s16): Likewise.
2906 (vreinterpret_u16_s32): Likewise.
2907 (vreinterpret_u16_s64): Likewise.
2908 (vreinterpret_u16_f32): Likewise.
2909 (vreinterpret_u16_u8): Likewise.
2910 (vreinterpret_u16_u32): Likewise.
2911 (vreinterpret_u16_u64): Likewise.
2912 (vreinterpret_u16_p8): Likewise.
2913 (vreinterpret_u16_p16): Likewise.
2914 (vreinterpretq_u16_s8): Likewise.
2915 (vreinterpretq_u16_s16): Likewise.
2916 (vreinterpretq_u16_s32): Likewise.
2917 (vreinterpretq_u16_s64): Likewise.
2918 (vreinterpretq_u16_f32): Likewise.
2919 (vreinterpretq_u16_u8): Likewise.
2920 (vreinterpretq_u16_u32): Likewise.
2921 (vreinterpretq_u16_u64): Likewise.
2922 (vreinterpretq_u16_p8): Likewise.
2923 (vreinterpretq_u16_p16): Likewise.
2924 (vreinterpret_u32_s8): Likewise.
2925 (vreinterpret_u32_s16): Likewise.
2926 (vreinterpret_u32_s32): Likewise.
2927 (vreinterpret_u32_s64): Likewise.
2928 (vreinterpret_u32_f32): Likewise.
2929 (vreinterpret_u32_u8): Likewise.
2930 (vreinterpret_u32_u16): Likewise.
2931 (vreinterpret_u32_u64): Likewise.
2932 (vreinterpret_u32_p8): Likewise.
2933 (vreinterpret_u32_p16): Likewise.
2934 (vreinterpretq_u32_s8): Likewise.
2935 (vreinterpretq_u32_s16): Likewise.
2936 (vreinterpretq_u32_s32): Likewise.
2937 (vreinterpretq_u32_s64): Likewise.
2938 (vreinterpretq_u32_f32): Likewise.
2939 (vreinterpretq_u32_u8): Likewise.
2940 (vreinterpretq_u32_u16): Likewise.
2941 (vreinterpretq_u32_u64): Likewise.
2942 (vreinterpretq_u32_p8): Likewise.
2943 (vreinterpretq_u32_p16): Likewise.
2944
2945 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2946
2947 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2948 Pattern extended.
2949 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2950 extended.
2951 (sqabs): Likewise.
2952 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2953 (vqnegd_s64): Likewise.
2954 (vqabs_s64): Likewise.
2955 (vqabsd_s64): Likewise.
2956
2957 2014-04-22 Richard Henderson <rth@redhat.com>
2958
2959 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2960 computation to the top of the loop.
2961
2962 2014-04-22 Renlin <renlin.li@arm.com>
2963 Jiong Wang <jiong.wang@arm.com>
2964
2965 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2966 * config/aarch64/aarch64.c (aarch64_layout_frame)
2967 (aarch64_initial_elimination_offset): Likewise.
2968
2969 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2970
2971 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2972 Fix indentation.
2973
2974 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2975
2976 * machmode.h (bitwise_mode_for_mode): Declare.
2977 * stor-layout.h (bitwise_type_for_mode): Likewise.
2978 * stor-layout.c (bitwise_mode_for_mode): New function.
2979 (bitwise_type_for_mode): Likewise.
2980 * builtins.c (fold_builtin_memory_op): Use it instead of
2981 int_mode_for_mode and build_nonstandard_integer_type.
2982
2983 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2984
2985 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2986 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2987 (*-*-solaris2*): Simplify.
2988 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2989 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2990 *-*-solaris2.9* handling.
2991
2992 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2993 as bug.
2994 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2995 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2996 handling, simplify.
2997 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2998 * configure: Regenerate.
2999
3000 * config/i386/sol2-9.h: Remove.
3001
3002 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3003 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3004 Remove Solaris 9 references.
3005
3006 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3007
3008 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3009 (floatuns<GPI:mode><GPF:mode>2): Remove.
3010 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3011 and floatuns conversions.
3012 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3013 and floatuns conversions.
3014 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3015 (w1,w2): New mode attributes for inequal width conversions.
3016
3017 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3018
3019 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3020 the output asm format.
3021
3022 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3023
3024 * config/aarch64/aarch64-simd.md
3025 (aarch64_cm<optab>di): Always split.
3026 (*aarch64_cm<optab>di): New.
3027 (aarch64_cmtstdi): Always split.
3028 (*aarch64_cmtstdi): New.
3029
3030 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3031
3032 PR tree-optimization/60823
3033 * omp-low.c (ipa_simd_modify_function_body): Go through
3034 all SSA_NAMEs and for those refering to vector arguments
3035 which are going to be replaced adjust SSA_NAME_VAR and,
3036 if it is a default definition, change it into a non-default
3037 definition assigned at the beginning of function from new_decl.
3038 (ipa_simd_modify_stmt_ops): Rewritten.
3039 * tree-dfa.c (set_ssa_default_def): When removing default def,
3040 check for NULL loc instead of NULL *loc.
3041
3042 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3043
3044 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3045 restrictions on core registers for DImode values in Thumb2.
3046
3047 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3048
3049 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3050 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3051
3052 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3053
3054 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3055 (*iordi_notzesidi_di): Likewise.
3056 (*iordi_notsesidi_di): Likewise.
3057
3058 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3059
3060 * config/arm/arm-protos.h (tune_params): New struct members.
3061 * config/arm/arm.c: Initialise tune_params per processor.
3062 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3063 for speed, based on new tune_params.
3064
3065 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3066
3067 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
3068 added.
3069 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
3070 macro.
3071 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
3072 corrected.
3073 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3074 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3075 (vrnda_f64): Likewise.
3076 (vrndi_f64): Likewise.
3077 (vrndm_f64): Likewise.
3078 (vrndn_f64): Likewise.
3079 (vrndp_f64): Likewise.
3080 (vrndx_f64): Likewise.
3081
3082 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3083
3084 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3085 GET_MODE_SIZE argument is enum machine_mode.
3086
3087 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3088
3089 PR target/60910
3090 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3091 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3092
3093 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3094
3095 PR middle-end/60281
3096 * asan.c (asan_emit_stack_protection): Force the base to align to
3097 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3098 appropriate bits if STRICT_ALIGNMENT.
3099 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3100 when asan is on.
3101 (expand_used_vars): Leave a space in the stack frame for alignment
3102 if STRICT_ALIGNMENT.
3103
3104 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3105
3106 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3107 than a gimple.
3108 (gimple_store_p): Likewise.
3109 (gimple_assign_load_p): Likewise.
3110 (gimple_assign_cast_p): Likewise.
3111 (gimple_clobber_p): Likewise.
3112
3113 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3114 rather than a gimple.
3115 (gimple_assign_cast_p): Likewise.
3116
3117 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3118
3119 PR target/60735
3120 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3121 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3122
3123 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3124 more debug information for E500 if -mdebug=reg.
3125
3126 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3127
3128 PR target/60909
3129 * config/i386/i386.c (ix86_expand_builtin)
3130 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3131 register for target RTX.
3132 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3133
3134 2014-04-18 Cong Hou <congh@google.com>
3135
3136 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3137 the widen-mult pattern by handling two operands with different sizes,
3138 and operands whose size is smaller than half of the result type.
3139
3140 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3141
3142 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3143 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3144 (do_estimate_edge_time): Compute it.
3145 * ipa-inline.c (want_inline_small_function_p): Bypass
3146 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3147
3148 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3149
3150 * ipa-inline.c (spec_rem): New static variable.
3151 (dump_overall_stats): New function.
3152 (dump_inline_stats): New function.
3153
3154 2014-04-18 Richard Henderson <rth@redhat.com>
3155
3156 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3157 to GET_MODE_SIZE, not a reg_class_t.
3158
3159 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3160
3161 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3162 (vsx_xxmrglw_<mode>): Likewise.
3163
3164 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3165
3166 PR target/60876
3167 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3168 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3169 (rs6000_init_hard_regno_mode_ok): Likewise.
3170
3171 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3172
3173 * ipa-inline.c (inline_small_functions): Account only non-cold
3174 functions.
3175 * doc/invoke.texi (inline-unit-growth): Update documentation.
3176
3177 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3178
3179 * config/rs6000/rs6000.md (addti3, subti3): New.
3180
3181 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3182
3183 PR target/60863
3184 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3185 comment. Check optimize_insn_for_size_p instead of
3186 optimize_insn_for_speed_p.
3187
3188 2014-04-17 Martin Jambor <mjambor@suse.cz>
3189
3190 * gimple-iterator.c (gsi_start_edge): New function.
3191 * gimple-iterator.h (gsi_start_edge): Declare.
3192 * tree-sra.c (single_non_eh_succ): New function.
3193 (disqualify_ops_if_throwing_stmt): Renamed to
3194 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3195 having one non-EH successor BB.
3196 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3197 generate loads into replacements.
3198 (sra_modify_assign): Likewise and and also use the simple path for
3199 such statements.
3200 (sra_modify_function_body): Commit statements on edges.
3201
3202 2014-04-17 Richard Biener <rguenther@suse.de>
3203
3204 PR middle-end/60849
3205 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3206 comparison results and add clarifying comment.
3207
3208 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3209
3210 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3211 (blank_mode): Initialize it.
3212 (emit_mode_size_inline, emit_mode_nunits_inline,
3213 emit_mode_inner_inline): New functions.
3214 (emit_insn_modes_h): Call them and surround their output with
3215 #if GCC_VERSION >= 4001 ... #endif.
3216 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3217 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3218 mode_* arrays if the argument is __builtin_constant_p.
3219 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3220 is enum machine_mode.
3221
3222 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3223
3224 * passes.c (opt_pass::execute): Adjust.
3225 (pass_manager::execute_pass_mode_switching): Likewise.
3226 (early_local_passes::execute): Likewise.
3227 (execute_one_pass): Pass cfun to the pass's execute method.
3228 * tree-pass.h (opt_pass::execute): Add function * argument.
3229 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3230 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3231 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3232 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3233 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3234 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3235 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3236 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3237 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3238 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3239 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3240 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3241 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3242 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3243 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3244 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3245 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3246 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3247 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3248 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3249 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3250 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3251 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3252 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3253 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3254 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3255 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3256 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3257 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3258 Adjust.
3259
3260 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3261
3262 * passes.c (opt_pass::gate): Take function * argument.
3263 (gate_all_early_local_passes): Merge into
3264 (early_local_passes::gate): this.
3265 (gate_all_early_optimizations): Merge into
3266 (all_early_optimizations::gate): this.
3267 (gate_all_optimizations): Mege into
3268 (all_optimizations::gate): this.
3269 (gate_all_optimizations_g): Merge into
3270 (all_optimizations_g::gate): this.
3271 (gate_rest_of_compilation): Mege into
3272 (rest_of_compilation::gate): this.
3273 (gate_postreload): Merge into
3274 (postreload::gate): this.
3275 (dump_one_pass): Pass cfun to the pass's gate method.
3276 (execute_ipa_summary_passes): Likewise.
3277 (execute_one_pass): Likewise.
3278 (ipa_write_summaries_2): Likewise.
3279 (ipa_write_optimization_summaries_1): Likewise.
3280 (ipa_read_summaries_1): Likewise.
3281 (ipa_read_optimization_summaries_1): Likewise.
3282 (execute_ipa_stmt_fixups): Likewise.
3283 * tree-pass.h (opt_pass::gate): Add function * argument.
3284 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3285 combine-stack-adj.c, combine.c, compare-elim.c,
3286 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3287 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3288 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3289 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3290 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3291 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3292 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3293 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3294 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3295 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3296 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3297 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3298 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3299 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3300 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3301 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3302 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3303 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3304 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3305 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3306 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3307 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3308 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3309 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3310 var-tracking.c, vtable-verify.c, web.c: Adjust.
3311
3312 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3313
3314 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3315 * configure: Regenerate.
3316
3317 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3318
3319 * passes.c (dump_one_pass): don't check pass->has_gate.
3320 (execute_ipa_summary_passes): Likewise.
3321 (execute_one_pass): Likewise.
3322 (ipa_write_summaries_2): Likewise.
3323 (ipa_write_optimization_summaries_1): Likewise.
3324 (ipa_read_optimization_summaries_1): Likewise.
3325 (execute_ipa_stmt_fixups): Likewise.
3326 * tree-pass.h (pass_data::has_gate): Remove.
3327 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3328 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3329 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3330 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3331 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3332 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3333 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3334 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3335 gimple-low.c, gimple-ssa-isolate-paths.c,
3336 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3337 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3338 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3339 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3340 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3341 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3342 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3343 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3344 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3345 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3346 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3347 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3348 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3349 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3350 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3351 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3352 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3353 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3354 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3355 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3356 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3357 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3358 Adjust.
3359
3360 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3361
3362 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3363 declaration.
3364 * passes.c (pass_manager::register_dump_files_1): Merge into
3365 (pass_manager::register_dump_files): this, and remove its handling of
3366 properties since the pass always has the properties anyway.
3367 (pass_manager::pass_manager): Adjust.
3368
3369 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3370
3371 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3372 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3373 dealing with properties.
3374 (pass_manager::register_dump_files): Adjust.
3375
3376 2014-03-20 Mark Wielaard <mjw@redhat.com>
3377
3378 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3379 then represent the bound as normal constant value.
3380
3381 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3382
3383 PR target/60847
3384 Forward port from 4.8 branch
3385 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3386
3387 * config/i386/bmiintrin.h (_blsi_u32): New.
3388 (_blsi_u64): Ditto.
3389 (_blsr_u32): Ditto.
3390 (_blsr_u64): Ditto.
3391 (_blsmsk_u32): Ditto.
3392 (_blsmsk_u64): Ditto.
3393 (_tzcnt_u32): Ditto.
3394 (_tzcnt_u64): Ditto.
3395
3396 2014-04-17 Kito Cheng <kito@0xlab.org>
3397
3398 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3399
3400 2014-04-17 Richard Biener <rguenther@suse.de>
3401
3402 PR middle-end/60849
3403 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3404 boolean results for comparisons.
3405
3406 2014-04-17 Richard Biener <rguenther@suse.de>
3407
3408 PR tree-optimization/60836
3409 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3410 initial PHI args to be gimple values.
3411
3412 2014-04-17 Richard Biener <rguenther@suse.de>
3413
3414 PR tree-optimization/60841
3415 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3416 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3417 of stmts to SLP build.
3418 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3419 (vect_analyze_slp): Likewise.
3420 (vect_analyze_slp_instance): Likewise.
3421 (vect_build_slp_tree): Limit overall SLP tree growth.
3422 * tree-vectorizer.h (vect_analyze_data_refs,
3423 vect_analyze_slp): Adjust prototypes.
3424
3425 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3426
3427 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3428 Silvermont.
3429
3430 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3431
3432 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3433 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3434 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3435 for TARGET_SLOW_PSHUFB
3436
3437 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3438
3439 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3440 * config/i386/i386.c (intel_cost): Ditto.
3441
3442 2014-04-17 Joey Ye <joey.ye@arm.com>
3443
3444 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3445
3446 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3447
3448 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3449 with -fuse-profile.
3450
3451 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3452
3453 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3454 (type_all_derivations_known_p): New predicate.
3455 (type_all_ctors_visible_p): New predicate.
3456 (type_possibly_instantiated_p): New predicate.
3457 (get_odr_type): Compute all_derivations_known.
3458 (dump_odr_type): Dump the flag.
3459 (maybe_record_type): Cleanup.
3460 (record_target_from_binfo): Add bases_to_consider array;
3461 record bases for types w/o instances and skip CXX destructor.
3462 (possible_polymorphic_call_targets_1): Add bases_to_consider
3463 and consider_construction parameters; check if type may have instance.
3464 (get_polymorphic_call_info): Set maybe_in_construction to true
3465 when we know nothing.
3466 (record_targets_from_bases): Skip CXX destructors; they are
3467 never called for types in construction.
3468 (possible_polymorphic_call_targets): Do not record target when
3469 type may not have instance.
3470
3471 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3472
3473 PR ipa/60854
3474 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3475 external aliases alive, too.
3476
3477 2014-04-16 Andrew Pinski <apinski@cavium.com>
3478
3479 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3480 definition.
3481
3482 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3483
3484 * final.c (compute_alignments): Do not apply loop alignment to a block
3485 falling through to the exit.
3486
3487 2014-04-16 Catherine Moore <clm@codesourcery.com>
3488
3489 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3490 Adjust constraints for microMIPS store patterns.
3491
3492 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3493
3494 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3495
3496 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3497
3498 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3499 (append_use): Run at -O0.
3500 (append_vdef): Likewise.
3501 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3502 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3503
3504 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3505
3506 PR tree-optimization/60844
3507 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3508 (propagate_op_to_single_use, remove_visited_stmt_chain,
3509 linearize_expr, repropagate_negates, reassociate_bb): Use it
3510 instead of gsi_remove.
3511
3512 2014-04-16 Martin Jambor <mjambor@suse.cz>
3513
3514 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3515 ipa_transforms_to_apply.
3516 (cgraph_function_versioning): Assert that old_node has empty
3517 ipa_transforms_to_apply.
3518 * trans-mem.c (ipa_tm_create_version): Likewise.
3519 * tree-inline.c (tree_function_versioning): Do not duplicate
3520 ipa_transforms_to_apply.
3521
3522 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3523
3524 PR target/60817
3525 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3526 x86_64-*-* cases.
3527 Pass necessary as flags on 64-bit Solaris/x86.
3528 Use lowercase relocs for x86_64-*-*.
3529 * configure: Regenerate.
3530
3531 2014-04-15 Jan Hubicka <jh@suse.cz>
3532
3533 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3534 (maybe_record_node, likely_target_p): Use it.
3535
3536 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3537
3538 PR target/60839
3539 Revert following patch
3540
3541 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3542
3543 PR target/60735
3544 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3545 software floating point or no floating point registers, do not
3546 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3547 in GPRs that occurs after we tested for GPRs that would never be
3548 true.
3549
3550 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3551 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3552 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3553 specifically allow DDmode, since that does not use the SPE SIMD
3554 instructions.
3555
3556 2014-03-21 Mark Wielaard <mjw@redhat.com>
3557
3558 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3559 as unsigned or int depending on type and value used.
3560
3561 2014-04-15 Richard Biener <rguenther@suse.de>
3562
3563 PR rtl-optimization/56965
3564 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3565 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3566 ... here.
3567 * alias.c (true_dependence_1): Do not call
3568 nonoverlapping_component_refs_p.
3569 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3570 nonoverlapping_component_refs_p.
3571 (indirect_refs_may_alias_p): Likewise.
3572
3573 2014-04-15 Teresa Johnson <tejohnson@google.com>
3574
3575 * cfg.c (dump_bb_info): Fix flags check.
3576 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3577
3578 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3579
3580 PR rtl-optimization/60663
3581 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3582 avoid 0 cost.
3583
3584 2014-04-15 Richard Biener <rguenther@suse.de>
3585
3586 * lto-streamer.h (LTO_major_version): Bump to 4.
3587
3588 2014-04-15 Richard Biener <rguenther@suse.de>
3589
3590 * common.opt (lto_partition_model): New enum.
3591 (flto-partition=): Merge separate options with a single with argument,
3592 add -flto-partition=one support.
3593 * flag-types.h (enum lto_partition_model): Declare.
3594 * opts.c (finish_options): Remove duplicate -flto-partition=
3595 option check.
3596 * lto-wrapper.c (run_gcc): Adjust.
3597
3598 2014-04-15 Richard Biener <rguenther@suse.de>
3599
3600 * alias.c (ncr_compar): New function.
3601 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3602
3603 2014-04-15 Richard Biener <rguenther@suse.de>
3604
3605 * alias.c (record_component_aliases): Do not walk BINFOs.
3606
3607 2014-04-15 Richard Biener <rguenther@suse.de>
3608
3609 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3610 Add struct function argument and adjust.
3611 (find_func_aliases_for_call): Likewise.
3612 (find_func_aliases): Likewise.
3613 (find_func_clobbers): Likewise.
3614 (intra_create_variable_infos): Likewise.
3615 (compute_points_to_sets): Likewise.
3616 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3617
3618 2014-04-15 Richard Biener <rguenther@suse.de>
3619
3620 * tree.c (iterative_hash_expr): Use enum tree_code_class
3621 to store TREE_CODE_CLASS.
3622 (tree_block): Likewise.
3623 (tree_set_block): Likewise.
3624 * tree.h (fold_build_pointer_plus_loc): Use
3625 convert_to_ptrofftype_loc.
3626
3627 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3628
3629 PR plugins/59335
3630 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3631 added in 4.9.
3632
3633 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3634
3635 * cfgloop.h (struct loop): Move force_vectorize down.
3636 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3637 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3638 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3639 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3640 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3641 * tree-core.h (enum annot_expr_kind): Add new kind values.
3642 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3643 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3644 kinds.
3645 * tree.def (ANNOTATE_EXPR): Tweak comment.
3646
3647 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3648
3649 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3650 cxa_pure_virtual).
3651
3652 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3653
3654 * tree.h (TYPE_IDENTIFIER): Declare.
3655 * tree.c (subrange_type_for_debug_p): Use it.
3656 * godump.c (go_format_type): Likewise.
3657 * dwarf2out.c (is_cxx_auto, modified_type_die,
3658 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3659 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3660
3661 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3662
3663 PR lto/60820
3664 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3665
3666 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3667
3668 * config/i386/i386.c (examine_argument): Return bool. Return true if
3669 parameter should be passed in memory.
3670 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3671 (construct_container): Update calls to examine_argument.
3672 (function_arg_advance_64): Ditto.
3673 (return_in_memory_32): Merge with ix86_return_in_memory.
3674 (return_in_memory_64): Ditto.
3675 (return_in_memory_ms_64): Ditto.
3676
3677 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3678
3679 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3680 * coverage.c (coverage_compute_profile_id): Handle externally visible
3681 symbols.
3682
3683 2014-04-14 Martin Jambor <mjambor@suse.cz>
3684
3685 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3686 DECL_DISREGARD_INLINE_LIMITS functions.
3687
3688 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3689
3690 PR target/60827
3691 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3692
3693 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3694
3695 PR target/60827
3696 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3697 optimize_insn_for_speed_p instead of
3698 optimize_function_for_speed_p.
3699
3700 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3701
3702 * doc/invoke.texi (free): Document AArch64.
3703
3704 2014-04-14 Richard Biener <rguenther@suse.de>
3705
3706 PR tree-optimization/60042
3707 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3708 (insert_into_preds_of_block): Do not prevent PHI insertion
3709 for REFERENCE exprs here ...
3710 (eliminate_dom_walker::before_dom_children): ... but prevent
3711 their use here under similar conditions when applied to the
3712 IL after PRE optimizations.
3713
3714 2014-04-14 Richard Biener <rguenther@suse.de>
3715
3716 * passes.def: Move early points-to after early SRA.
3717
3718 2014-04-14 Richard Biener <rguenther@suse.de>
3719
3720 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3721 check for which sign-changes we allow when forwarding
3722 a converted value into a switch.
3723
3724 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3725
3726 * stor-layout.c (place_field): Finalize non-constant offset for the
3727 field, if any.
3728
3729 2014-04-14 Richard Biener <rguenther@suse.de>
3730
3731 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3732 as argument.
3733 (expand_switch_using_bit_tests_p): Likewise.
3734 (process_switch): Compute and pass on speed_p based on the
3735 switch stmt.
3736 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3737 optimize_bb_for_speed_p.
3738
3739 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3740
3741 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3742 * function.h (struct function): Rename has_force_vect_loops into
3743 has_force_vectorize_loops.
3744 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3745 (input_struct_function_base): Likewise.
3746 * lto-streamer-out.c (output_cfg): Likewise.
3747 (output_struct_function_base): Likewise.
3748 * omp-low.c (expand_omp_simd): Likewise.
3749 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3750 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3751 (version_loop_for_if_conversion): Likewise.
3752 (tree_if_conversion): Likewise.
3753 (main_tree_if_conversion): Likewise.
3754 (gate_tree_if_conversion): Likewise.
3755 * tree-inline.c (copy_loops): Likewise.
3756 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3757 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3758 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3759 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3760 * tree-vectorizer.c (vectorize_loops): Likewise.
3761 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3762
3763 2014-04-14 Richard Biener <rguenther@suse.de>
3764
3765 PR lto/60720
3766 * lto-streamer-out.c (wrap_refs): New function.
3767 (lto_output): Wrap symbol references in global initializes in
3768 type-preserving MEM_REFs.
3769
3770 2014-04-14 Christian Bruel <christian.bruel@st.com>
3771
3772 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3773
3774 2014-04-14 Christian Bruel <christian.bruel@st.com>
3775
3776 * config/sh/sh.md (setmemqi): New expand pattern.
3777 * config/sh/sh.h (CLEAR_RATIO): Define.
3778 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3779 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3780
3781 2014-04-14 Richard Biener <rguenther@suse.de>
3782
3783 PR middle-end/55022
3784 * fold-const.c (negate_expr_p): Don't negate directional rounding
3785 division.
3786 (fold_negate_expr): Likewise.
3787
3788 2014-04-14 Richard Biener <rguenther@suse.de>
3789
3790 PR tree-optimization/59817
3791 PR tree-optimization/60453
3792 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3793 recursion to catch all CHRECs in the scalar evolution and restrict
3794 the predicate for the remains appropriately.
3795
3796 2014-04-12 Catherine Moore <clm@codesourcery.com>
3797
3798 * config/mips/constraints.md: Add new register constraint "kb".
3799 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3800 (*movhi_internal): Likewise.
3801 (*movqi_internal): Likewise.
3802 * config/mips/mips.h (M16_STORE_REGS): New register class.
3803 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3804 (REG_CLASS_CONTENTS): Likewise.
3805 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3806
3807 2014-04-11 Tobias Burnus <burnus@net-b.de>
3808
3809 PR c/60194
3810 * doc/invoke.texi (-Wformat-signedness): Document it.
3811 (Wformat=2): Mention that this enables -Wformat-signedness.
3812
3813 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3814
3815 * common/config/epiphany/epiphany-common.c
3816 (epiphany_option_optimization_table): Enable section anchors by
3817 default at -O1 or higher.
3818 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3819 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3820 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3821 carries no extra cost.
3822 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3823 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3824 * config/epiphany/predicates.md (memclob_operand): New predicate.
3825 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3826 Use memclob_operand predicate and X constraint for operand 3.
3827
3828 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3829
3830 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3831 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3832 its operands.
3833
3834 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3835
3836 PR rtl-optimization/60651
3837 * mode-switching.c (optimize_mode_switching): Make sure to emit
3838 sets of a lower numbered entity before sets of a higher numbered
3839 entity to a mode of the same or lower priority.
3840 When creating a seginfo for a basic block that starts with a code
3841 label, move the insertion point past the code label.
3842 (new_seginfo): Document and enforce requirement that
3843 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3844 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3845 * doc/tm.texi: Regenerate.
3846
3847 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3848
3849 PR target/60811
3850 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3851
3852 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3853
3854 * BASE-VER: Set to 4.10.0.
3855
3856 2014-04-11 Tobias Burnus <burnus@net-b.de>
3857
3858 PR other/59055
3859 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3860 * doc/gcc.texi (Service): Update description in the @menu
3861 * doc/invoke.texi (Option Summary): Remove misplaced and
3862 duplicated @menu.
3863
3864 2014-04-11 Steve Ellcey <sellcey@mips.com>
3865 Jakub Jelinek <jakub@redhat.com>
3866
3867 PR middle-end/60556
3868 * expr.c (convert_move): Use emit_store_flag_force instead of
3869 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3870 argument to it.
3871
3872 2014-04-11 Richard Biener <rguenther@suse.de>
3873
3874 PR middle-end/60797
3875 * varasm.c (assemble_alias): Avoid endless error reporting
3876 recursion by setting TREE_ASM_WRITTEN.
3877
3878 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3879
3880 * config/s390/s390.md: Add a splitter for NOT rtx.
3881
3882 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3883
3884 PR rtl-optimization/60663
3885 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3886
3887 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3888 Jakub Jelinek <jakub@redhat.com>
3889
3890 PR lto/60567
3891 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3892 flag from decl_node to node.
3893
3894 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3895
3896 PR debug/60655
3897 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3898 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3899 ameliorating the cases where it can be.
3900
3901 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3902
3903 Revert
3904 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3905
3906 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3907 (loadsync_<mode>): Change mode.
3908 (load_quadpti, store_quadpti): New.
3909 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3910 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3911 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3912
3913 2014-04-09 Cong Hou <congh@google.com>
3914
3915 PR testsuite/60773
3916 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3917 documentation.
3918
3919 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3920
3921 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3922 instead of vnor to exploit possible fusion opportunity in the
3923 future.
3924 (altivec_expand_vec_perm_const_le): Likewise.
3925
3926 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3927
3928 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3929 (loadsync_<mode>): Change mode.
3930 (load_quadpti, store_quadpti): New.
3931 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3932 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3933
3934 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3935
3936 PR target/60763
3937 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3938 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3939 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3940
3941 2014-04-08 Richard Biener <rguenther@suse.de>
3942
3943 PR middle-end/60706
3944 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3945 a 64bit widest int print double-int similar to on HWI64 hosts.
3946
3947 2014-04-08 Richard Biener <rguenther@suse.de>
3948
3949 PR tree-optimization/60785
3950 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3951 default defs properly.
3952
3953 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3954
3955 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3956 (Weffc++): Likewise.
3957
3958 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3959
3960 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3961 set completep to false rather than true.
3962
3963 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3964
3965 PR target/60504
3966 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3967 ARM_TARGET2_DWARF_FORMAT.
3968
3969 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3970
3971 PR target/60609
3972 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3973 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3974 ADDR_DIFF_VEC.
3975
3976 2014-04-07 Richard Biener <rguenther@suse.de>
3977
3978 PR tree-optimization/60766
3979 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3980 (may_eliminate_iv): Convert cand_value_at result to desired type.
3981
3982 2014-04-07 Jason Merrill <jason@redhat.com>
3983
3984 PR c++/60731
3985 * common.opt (-fno-gnu-unique): Add.
3986 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3987
3988 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3989
3990 * haifa-sched.c: Fix outdated function reference and minor
3991 grammar errors in introductory comment.
3992
3993 2014-04-07 Richard Biener <rguenther@suse.de>
3994
3995 PR middle-end/60750
3996 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3997 for noreturn calls.
3998 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3999
4000 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4001
4002 PR debug/55794
4003 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4004 size accounting for thunks.
4005 (pa_asm_output_mi_thunk): Use final_start_function() and
4006 final_end_function() to output function start and end directives.
4007
4008 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4009
4010 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4011 device specific ISA/ feature information. Remove short_sp and
4012 errata_skip ds. Add avr_device_specific_features enum to have device
4013 specific info.
4014 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4015 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4016 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4017 updated device specific info.
4018 * config/avr/avr-mcus.def: Merge device specific details to
4019 dev_attribute field.
4020 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4021 errata_skip.
4022 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4023 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4024 assembler if RMW isa supported by current device.
4025 * config/avr/genmultilib.awk: Update as device info structure changed.
4026 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4027
4028 2014-04-04 Cong Hou <congh@google.com>
4029
4030 PR tree-optimization/60656
4031 * tree-vect-stmts.c (supportable_widening_operation):
4032 Fix a bug that elements in a vector with vect_used_by_reduction
4033 property are incorrectly reordered when the operation on it is not
4034 consistant with the one in reduction operation.
4035
4036 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4037
4038 PR rtl-optimization/60155
4039 * gcse.c (record_set_data): New function.
4040 (single_set_gcse): New function.
4041 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4042 (hoist_code): Likewise.
4043 (get_pressure_class_and_nregs): Likewise.
4044
4045 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4046
4047 * explow.c (probe_stack_range): Emit a final optimization blockage.
4048
4049 2014-04-04 Anthony Green <green@moxielogic.com>
4050
4051 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4052 typos.
4053
4054 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4055
4056 PR ipa/59626
4057 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4058 flags are set only during streaming.
4059 * ipa.c (process_references, walk_polymorphic_call_targets,
4060 symtab_remove_unreachable_nodes): Drop bodies of always inline
4061 after early inlining.
4062 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4063
4064 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4065 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4066
4067 PR debug/60655
4068 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4069 containing a NOT.
4070
4071 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4072
4073 PR bootstrap/60743
4074 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4075 duration.
4076 (cortex_a53_fdivd): Likewise.
4077
4078 2014-04-04 Martin Jambor <mjambor@suse.cz>
4079
4080 PR ipa/60640
4081 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4082 Adjust all callers.
4083 * cgraph.c (clone_of_p): Also return true if thunks match.
4084 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4085 cgraph_function_or_thunk_node and an obsolete comment.
4086 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4087 file.
4088 (build_function_decl_skip_args): Likewise.
4089 (set_new_clone_decl_and_node_flags): New function.
4090 (duplicate_thunk_for_node): Likewise.
4091 (redirect_edge_duplicating_thunks): Likewise.
4092 (cgraph_clone_node): New parameter args_to_skip, pass it to
4093 redirect_edge_duplicating_thunks which is called instead of
4094 cgraph_redirect_edge_callee.
4095 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4096 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4097
4098 2014-04-04 Jeff Law <law@redhat.com>
4099
4100 PR target/60657
4101 * config/arm/predicates.md (const_int_I_operand): New predicate.
4102 (const_int_M_operand): Similarly.
4103 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4104 const_int_operand.
4105 (insv_t2, extv_reg, extzv_t2): Likewise.
4106 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4107 (pop_multiple_with_writeback_and_return): Likewise.
4108 (vfp_pop_multiple_with_writeback): Likewise
4109
4110 2014-04-04 Richard Biener <rguenther@suse.de>
4111
4112 PR ipa/60746
4113 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4114 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4115 non-GIMPLE_LABELs.
4116 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4117 * gimplify.c (gimple_add_tmp_var_fn): New function.
4118 * gimple-expr.h (create_tmp_reg_fn): Declare.
4119 * gimple-expr.c (create_tmp_reg_fn): New function.
4120 * gimple-low.c (record_vars_into): Don't change cfun.
4121 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4122 code generation without cfun.
4123
4124 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4125
4126 PR bootstrap/60719
4127 * Makefile.in (install-driver): Fix shell scripting.
4128
4129 2014-04-03 Cong Hou <congh@google.com>
4130
4131 PR tree-optimization/60505
4132 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4133 threshold of number of iterations below which no vectorization
4134 will be done.
4135 * tree-vect-loop.c (new_loop_vec_info):
4136 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4137 * tree-vect-loop.c (vect_analyze_loop_operations):
4138 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4139 * tree-vect-loop.c (vect_transform_loop):
4140 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4141 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4142 of iterations of the loop and see if we should build the epilogue.
4143
4144 2014-04-03 Richard Biener <rguenther@suse.de>
4145
4146 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4147 (streamer_tree_cache_create): Adjust.
4148 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4149 to allow optional nodes array.
4150 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4151 (streamer_tree_cache_append): Likewise.
4152 (streamer_tree_cache_create): Create nodes array optionally
4153 as specified by parameter.
4154 * lto-streamer-out.c (create_output_block): Avoid maintaining
4155 the node array in the writer cache.
4156 (DFS_write_tree): Remove assertion.
4157 (produce_asm_for_decls): Free the out decl state hash table early.
4158 * lto-streamer-in.c (lto_data_in_create): Adjust for
4159 streamer_tree_cache_create prototype change.
4160
4161 2014-04-03 Richard Biener <rguenther@suse.de>
4162
4163 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4164 set TREE_CHAIN to NULL_TREE.
4165
4166 2014-04-03 Richard Biener <rguenther@suse.de>
4167
4168 PR tree-optimization/60740
4169 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4170 over all GIMPLE_COND operands.
4171
4172 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4173
4174 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4175 (Weffc++): Remove Scott's numbering, merge lists and reference
4176 Wnon-virtual-dtor.
4177
4178 2014-04-03 Nick Clifton <nickc@redhat.com>
4179
4180 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4181 properly.
4182
4183 2014-04-03 Martin Jambor <mjambor@suse.cz>
4184
4185 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4186 mention gcc_unreachable before failing.
4187 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4188 removed symbols.
4189
4190 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4191
4192 PR ipa/60659
4193 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4194 inconsistent code and instead mark the context inconsistent.
4195 (possible_polymorphic_call_targets): For inconsistent contexts
4196 return empty complete list.
4197
4198 2014-04-02 Anthony Green <green@moxielogic.com>
4199
4200 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4201 (extendqisi2, extendhisi2): Define.
4202 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4203 (WCHAR_TYPE): Change to unsigned int.
4204
4205 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4206
4207 PR tree-optimization/60733
4208 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4209 insertion point for PHI candidates to be the end of the feeding
4210 block for the PHI argument.
4211
4212 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4213
4214 PR rtl-optimization/60650
4215 * lra-constraints.c (process_alt_operands): Decrease reject for
4216 earlyclobber matching.
4217
4218 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4219
4220 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4221
4222 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4223
4224 * config/spu/spu.c (pad_bb): Do not crash when the last
4225 insn is CODE_FOR_blockage.
4226
4227 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4228
4229 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4230 lies outside the target mode.
4231
4232 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4233
4234 PR target/60735
4235 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4236 software floating point or no floating point registers, do not
4237 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4238 in GPRs that occurs after we tested for GPRs that would never be
4239 true.
4240
4241 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4242 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4243 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4244 specifically allow DDmode, since that does not use the SPE SIMD
4245 instructions.
4246
4247 2014-04-02 Richard Biener <rguenther@suse.de>
4248
4249 PR middle-end/60729
4250 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4251 MODE_INTs. Properly use negv_optab.
4252 (expand_abs): Likewise.
4253
4254 2014-04-02 Richard Biener <rguenther@suse.de>
4255
4256 PR bootstrap/60719
4257 * Makefile.in (install-driver): Guard extra installs with special
4258 names properly.
4259
4260 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4261
4262 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4263 Document vec_vgbbd.
4264
4265 2014-04-01 Richard Henderson <rth@redhat.com>
4266
4267 PR target/60704
4268 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4269 alternative enabled before register allocation.
4270
4271 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4272
4273 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4274 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4275 typo.
4276 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4277 (nios2_got_address): Update nios2_large_got_address call site.
4278 (nios2_delegitimize_address): New function.
4279 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4280 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4281 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4282
4283 2014-04-01 Martin Husemann <martin@duskware.de>
4284
4285 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4286 for -mabi=32.
4287
4288 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4289
4290 PR rtl-optimization/60604
4291 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4292 check from register_operand.
4293 (register_operand): Redefine in terms of general_operand.
4294 (nonmemory_operand): Use register_operand for the non-constant cases.
4295
4296 2014-04-01 Richard Biener <rguenther@suse.de>
4297
4298 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4299
4300 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4301
4302 * doc/invoke.texi (mapp-regs): Clarify.
4303
4304 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4305
4306 * config/i386/avx512fintrin.h (__v32hi): Define type.
4307 (__v64qi): Likewise.
4308 (_mm512_set1_epi8): Define.
4309 (_mm512_set1_epi16): Define.
4310 (_mm512_set4_epi32): Define.
4311 (_mm512_set4_epi64): Define.
4312 (_mm512_set4_pd): Define.
4313 (_mm512_set4_ps): Define.
4314 (_mm512_setr4_epi64): Define.
4315 (_mm512_setr4_epi32): Define.
4316 (_mm512_setr4_pd): Define.
4317 (_mm512_setr4_ps): Define.
4318 (_mm512_setzero_epi32): Define.
4319
4320 2014-03-31 Martin Jambor <mjambor@suse.cz>
4321
4322 PR middle-end/60647
4323 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4324 callsite_arguments_match_p. Updated all callers. Also check types of
4325 corresponding formal parameters and actual arguments.
4326 (not_all_callers_have_enough_arguments_p) Renamed to
4327 some_callers_have_mismatched_arguments_p.
4328
4329 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4330
4331 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4332
4333 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4334
4335 PR target/60034
4336 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4337 section anchor.
4338
4339 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4340
4341 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4342 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4343 Split out
4344 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4345 Use FMAMODE_NOVF512 mode iterator.
4346 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4347 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4348 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4349 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4350 Split out
4351 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4352 Use VF_128_256 mode iterator.
4353 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4354 Ditto.
4355
4356 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4357
4358 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4359 static chain if needed.
4360
4361 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4362
4363 PR target/60697
4364 * lra-constraints.c (index_part_to_reg): New.
4365 (process_address): Use it.
4366
4367 2014-03-27 Jeff Law <law@redhat.com>
4368 Jakub Jelinek <jakub@redhat.com>
4369
4370 PR target/60648
4371 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4372 gen_rtx_{PLUS,MULT} to build up the address expression.
4373
4374 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4375 creating non-canonical RTL.
4376
4377 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4378
4379 PR ipa/60243
4380 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4381 functions; reorganize to make cheap checks first.
4382 (inline_small_functions): Do not estimate growth when dumping;
4383 it is expensive.
4384 * ipa-inline.h (inline_summary): Add min_size.
4385 (growth_likely_positive): New function.
4386 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4387 (set_cond_stmt_execution_predicate): Cleanup.
4388 (estimate_edge_size_and_time): Compute min_size.
4389 (estimate_calls_size_and_time): Likewise.
4390 (estimate_node_size_and_time): Likewise.
4391 (inline_update_overall_summary): Update min_size.
4392 (do_estimate_edge_time): Likewise.
4393 (do_estimate_edge_size): Update.
4394 (do_estimate_edge_hints): Update.
4395 (growth_likely_positive): New function.
4396
4397 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4398
4399 PR target/60693
4400 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4401 also if addr has VOIDmode.
4402
4403 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4404
4405 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4406 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4407 Declare extern.
4408 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4409 instructions as well as AdvancedSIMD loads.
4410
4411 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4412
4413 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4414 Use crypto_aese type.
4415 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4416 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4417 crypto_aese, crypto_aesmc. Move to types.md.
4418 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4419 crypto_aesmc.
4420 * config/arm/iterators.md (crypto_type): Likewise.
4421
4422 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4423
4424 * cgraph.c: Include expr.h and tree-dfa.h.
4425 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4426 remove LHS.
4427
4428 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4429
4430 PR target/60675
4431 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4432 regs from checking multi-reg pseudos.
4433
4434 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4435
4436 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4437
4438 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4439
4440 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4441 if it would clobber the stack pointer, even temporarily.
4442
4443 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4444
4445 * mode-switching.c: Make small adjustments to the top comment.
4446
4447 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4448
4449 * config/rs6000/constraints.md (wD constraint): New constraint to
4450 match the constant integer to get the top DImode/DFmode out of a
4451 vector in a VSX register.
4452
4453 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4454 match the constant integer to get the top DImode/DFmode out of a
4455 vector in a VSX register.
4456
4457 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4458 for ISA 2.07.
4459
4460 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4461 vbpermq builtins.
4462
4463 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4464 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4465
4466 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4467 Optimize vec_extract of 64-bit values, where the value being
4468 extracted is in the top word, where we can use scalar
4469 instructions. Add direct move and store support. Combine the big
4470 endian/little endian vector select load support into a single insn.
4471 (vsx_extract_<mode>_internal1): Likewise.
4472 (vsx_extract_<mode>_internal2): Likewise.
4473 (vsx_extract_<mode>_load): Likewise.
4474 (vsx_extract_<mode>_store): Likewise.
4475 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4476 combined into vsx_extract_<mode>_load.
4477 (vsx_extract_<mode>_one_le): Likewise.
4478
4479 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4480 define the top 64-bit vector element.
4481
4482 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4483 constraint.
4484
4485 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4486 Document vec_vbpermq builtin.
4487
4488 PR target/60672
4489 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4490 enable use of xxsldwi and xxpermdi builtin functions.
4491 (vec_xxpermdi): Likewise.
4492
4493 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4494 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4495
4496 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4497
4498 PR rtl-optimization/60650
4499 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4500 first_p. Use it.
4501 (find_spills_for): New.
4502 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4503 Spill all pseudos on the second iteration.
4504
4505 2014-03-27 Marek Polacek <polacek@redhat.com>
4506
4507 PR c/50347
4508 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4509 types.
4510
4511 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4512
4513 * config/s390/s390.c (s390_can_use_return_insn): Check for
4514 call-saved FPRs on 31 bit.
4515
4516 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4517
4518 PR middle-end/60682
4519 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4520 if they need regimplification, just drop them instead of
4521 calling gimple_regimplify_operands on them.
4522
4523 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4524
4525 PR target/60580
4526 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4527 (aarch64_frame_pointer_required): Adjust logic.
4528 (aarch64_can_eliminate): Adjust logic.
4529 (aarch64_override_options_after_change): Adjust logic.
4530
4531 2014-03-27 Dehao Chen <dehao@google.com>
4532
4533 * ipa-inline.c (early_inliner): Update node's inline info.
4534
4535 2014-03-26 Dehao Chen <dehao@google.com>
4536
4537 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4538 compiler inserted conditional jumps for NAN float check.
4539
4540 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4541
4542 * ubsan.h (ubsan_create_data): Change second argument's type
4543 to const location_t *.
4544 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4545 _("<unknown>").
4546 (ubsan_create_data): Change second argument to const location_t *PLOC.
4547 Create Loc field whenever PLOC is non-NULL.
4548 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4549 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4550 callers.
4551
4552 PR other/59545
4553 * real.c (real_to_integer2): Change type of low to UHWI.
4554
4555 2014-03-26 Tobias Burnus <burnus@net-b.de>
4556
4557 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4558 (CILK_SELF_SPECS): New define.
4559 (driver_self_specs): Use it.
4560
4561 2014-03-26 Richard Biener <rguenther@suse.de>
4562
4563 * tree-pretty-print.c (percent_K_format): Implement special
4564 case for LTO and its stripped down BLOCK tree.
4565
4566 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4567
4568 PR sanitizer/60636
4569 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4570
4571 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4572 one range is range_int_cst_p, but not both, at least optimize
4573 addition/subtraction of 0 and multiplication by 0 or 1.
4574 * gimple-fold.c (gimple_fold_call): Fold
4575 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4576 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4577 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4578
4579 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4580
4581 PR rtl-optimization/60452
4582 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4583 <case REG>: Return 1 for invalid offsets from the frame pointer.
4584
4585 2014-03-26 Marek Polacek <polacek@redhat.com>
4586
4587 PR c/37428
4588 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4589 a structure/union.
4590
4591 2014-03-26 Marek Polacek <polacek@redhat.com>
4592
4593 PR c/39525
4594 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4595 field members.
4596
4597 2014-03-26 Marek Polacek <polacek@redhat.com>
4598
4599 PR other/59545
4600 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4601 multiplication in unsigned type.
4602
4603 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4604
4605 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4606
4607 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4608
4609 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4610
4611 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4612
4613 PR ipa/60315
4614 * cif-code.def (UNREACHABLE) New code.
4615 * ipa-inline.c (inline_small_functions): Skip edges to
4616 __builtlin_unreachable.
4617 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4618 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4619 predicate to __bulitin_unreachable.
4620 (set_cond_stmt_execution_predicate): Fix issue when
4621 invert_tree_comparison returns ERROR_MARK.
4622 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4623 propagate to inline clones.
4624 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4625 to unreachable.
4626 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4627 * cgraphclones.c (cgraph_clone_node): If call destination is already
4628 ureachable, do not redirect it back.
4629 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4630 unreachable.
4631
4632 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4633
4634 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4635 Do not modify inline clones.
4636
4637 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4638
4639 * config/i386/i386.md (general_sext_operand): New mode attr.
4640 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4641 don't generate (sign_extend (const_int)).
4642 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4643 operands[2]. Use We constraint instead of <i> and
4644 <general_sext_operand> predicate instead of <general_operand>.
4645 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4646 * config/i386/constraints.md (We): New constraint.
4647 * config/i386/predicates.md (x86_64_sext_operand,
4648 sext_operand): New predicates.
4649
4650 2014-03-25 Martin Jambor <mjambor@suse.cz>
4651
4652 PR ipa/60600
4653 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4654 inconsistent devirtualizations to __builtin_unreachable.
4655
4656 2014-03-25 Marek Polacek <polacek@redhat.com>
4657
4658 PR c/35449
4659 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4660
4661 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4662
4663 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4664 order of elements for big-endian.
4665
4666 2014-03-25 Richard Biener <rguenther@suse.de>
4667
4668 PR middle-end/60635
4669 * gimplify-me.c (gimple_regimplify_operands): Update the
4670 re-gimplifed stmt.
4671
4672 2014-03-25 Martin Jambor <mjambor@suse.cz>
4673
4674 PR ipa/59176
4675 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4676 (lto_output_varpool_node): Likewise.
4677 (input_overwrite_node): Likewise.
4678 (input_varpool_node): Likewise.
4679
4680 2014-03-25 Richard Biener <rguenther@suse.de>
4681
4682 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4683 (run_gcc): Likewise.
4684
4685 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4686
4687 * combine.c (simplify_compare_const): Add MODE argument.
4688 Handle mode_width 0 as very large mode_width.
4689 (try_combine, simplify_comparison): Adjust callers.
4690
4691 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4692 type to avoid signed integer overflow.
4693 * explow.c (plus_constant): Likewise.
4694
4695 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4696
4697 * doc/generic.texi: Correct typos.
4698
4699 2014-03-24 Tobias Burnus <burnus@net-b.de>
4700
4701 * doc/invoke.texi (-flto): Expand section about
4702 using static libraries with LTO.
4703
4704 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4705
4706 PR rtl-optimization/60501
4707 * optabs.def (addptr3_optab): New optab.
4708 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4709 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4710 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4711
4712 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4713
4714 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4715
4716 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4717
4718 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4719 _mm512_set1_pd.
4720
4721 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4722 (_mm256_undefined_ps): Define.
4723 (_mm256_undefined_pd): Define.
4724 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4725 (_mm_undefined_pd): Define.
4726 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4727 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4728 (_mm512_undefined_ps): Define.
4729 (_mm512_undefined_pd): Define.
4730 Use _mm*_undefined_*.
4731 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4732
4733 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4734
4735 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4736 (lshr_simd): DI mode added.
4737 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4738 (aarch64_ushr_simddi): Likewise.
4739 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4740 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4741 (vshrd_n_u64): Likewise.
4742
4743 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4744
4745 * Makefile.in (s-macro_list): Depend on cc1.
4746
4747 2014-03-23 Teresa Johnson <tejohnson@google.com>
4748
4749 * ipa-utils.c (ipa_print_order): Use specified dump file.
4750
4751 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4752
4753 PR rtl-optimization/60601
4754 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4755
4756 * gcc.c (eval_spec_function): Initialize save_growing_value.
4757
4758 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4759
4760 PR sanitizer/60613
4761 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4762 code == MINUS_EXPR, never swap op0 with op1.
4763
4764 * toplev.c (init_local_tick): Avoid signed integer multiplication
4765 overflow.
4766 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4767 shift by first operand's bitsize.
4768
4769 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4770
4771 PR target/60610
4772 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4773 redefine to 1 or 0.
4774 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4775 TARGET_ISA_64BIT_P(x).
4776
4777 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4778
4779 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4780 pattern for vector nor instead of subtract from splat(-1).
4781 (altivec_expand_vec_perm_const_le): Likewise.
4782
4783 2014-03-21 Richard Henderson <rth@twiddle.net>
4784
4785 PR target/60598
4786 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4787 related insns after epilogue_completed.
4788
4789 2014-03-21 Martin Jambor <mjambor@suse.cz>
4790
4791 PR ipa/59176
4792 * cgraph.h (symtab_node): New flag body_removed.
4793 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4794 when removing bodies.
4795 * symtab.c (dump_symtab_base): Dump body_removed flag.
4796 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4797 had their bodies removed.
4798
4799 2014-03-21 Martin Jambor <mjambor@suse.cz>
4800
4801 PR ipa/60419
4802 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4803 in the border.
4804
4805 2014-03-21 Richard Biener <rguenther@suse.de>
4806
4807 PR tree-optimization/60577
4808 * tree-core.h (struct tree_base): Document nothrow_flag use
4809 in VAR_DECL_NONALIASED.
4810 * tree.h (VAR_DECL_NONALIASED): New.
4811 (may_be_aliased): Adjust.
4812 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4813
4814 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4815
4816 * expr.c (expand_expr_real_1): Remove outdated comment.
4817
4818 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4819
4820 PR middle-end/60597
4821 * ira.c (adjust_cleared_regs): Call copy_rtx on
4822 *reg_equiv[REGNO (loc)].src_p before passing it to
4823 simplify_replace_fn_rtx.
4824
4825 PR target/60568
4826 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4827 into CONST, put pic register as first operand of PLUS. Use
4828 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4829
4830 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4831
4832 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4833
4834 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4835
4836 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4837 around for store forwarding issue in the FPU on the UT699.
4838 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4839 loads and operations if -mfix-ut699 is specified.
4840 (divtf3_hq): Tweak attribute.
4841 (sqrttf2_hq): Likewise.
4842
4843 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4844
4845 * calls.c (store_one_arg): Remove incorrect const qualification on the
4846 type of the temporary.
4847 * cfgexpand.c (expand_return): Likewise.
4848 * expr.c (expand_constructor): Likewise.
4849 (expand_expr_real_1): Likewise.
4850
4851 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4852
4853 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4854 of parts.
4855
4856 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4857
4858 PR target/60039
4859 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4860
4861 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4862
4863 * config/arm/aarch-common-protos.h
4864 (alu_cost_table): Fix spelling of "extend".
4865 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4866
4867 2014-03-19 Richard Biener <rguenther@suse.de>
4868
4869 PR middle-end/60553
4870 * tree-core.h (tree_type_common): Re-order pointer members
4871 to reduce recursion depth during GC walks.
4872
4873 2014-03-19 Marek Polacek <polacek@redhat.com>
4874
4875 PR sanitizer/60569
4876 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4877 before accessing it.
4878
4879 2014-03-19 Richard Biener <rguenther@suse.de>
4880
4881 PR lto/59543
4882 * lto-streamer-in.c (input_function): In WPA stage do not drop
4883 debug stmts.
4884
4885 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4886
4887 PR tree-optimization/60559
4888 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4889 with build_zero_cst assignment.
4890
4891 2014-03-18 Kai Tietz <ktietz@redhat.com>
4892
4893 PR rtl-optimization/56356
4894 * sdbout.c (sdbout_parms): Verify that parms'
4895 incoming argument is valid.
4896 (sdbout_reg_parms): Likewise.
4897
4898 2014-03-18 Richard Henderson <rth@redhat.com>
4899
4900 PR target/60562
4901 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4902 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4903 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4904
4905 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4906
4907 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4908 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4909 Italicize plugin event names in description. Explain that
4910 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4911 Remind that no GCC functions should be called after PLUGIN_FINISH.
4912 Explain what pragmas with expansion are.
4913
4914 2014-03-18 Martin Liska <mliska@suse.cz>
4915
4916 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4917 gimple call statement is update.
4918 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4919 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4920
4921 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4922
4923 PR sanitizer/60557
4924 * ubsan.c (ubsan_instrument_unreachable): Call
4925 initialize_sanitizer_builtins.
4926 (ubsan_pass): Likewise.
4927
4928 PR sanitizer/60535
4929 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4930 varpool_finalize_decl instead of rest_of_decl_compilation.
4931
4932 2014-03-18 Richard Biener <rguenther@suse.de>
4933
4934 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4935 by using bitmap_and_compl instead of bitmap_and_compl_into.
4936 (df_rd_transfer_function): Likewise.
4937
4938 2014-03-18 Richard Biener <rguenther@suse.de>
4939
4940 * doc/lto.texi (fresolution): Fix typo.
4941
4942 2014-03-18 Richard Biener <rguenther@suse.de>
4943
4944 * doc/invoke.texi (flto): Update for changes in 4.9.
4945
4946 2014-03-18 Richard Biener <rguenther@suse.de>
4947
4948 * doc/loop.texi: Remove section on the removed lambda framework.
4949 Update loop docs with recent changes in preserving loop structure.
4950
4951 2014-03-18 Richard Biener <rguenther@suse.de>
4952
4953 * doc/lto.texi (-fresolution): Document.
4954
4955 2014-03-18 Richard Biener <rguenther@suse.de>
4956
4957 * doc/contrib.texi: Adjust my name.
4958
4959 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4960
4961 PR ipa/58721
4962 * internal-fn.c: Include diagnostic-core.h.
4963 (expand_BUILTIN_EXPECT): New function.
4964 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4965 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4966 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4967 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4968 IFN_BUILTIN_EXPECT.
4969 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4970 Revert 3 argument __builtin_expect code.
4971 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4972 * gimple-fold.c (gimple_fold_call): Likewise.
4973 * tree.h (fold_builtin_expect): New prototype.
4974 * builtins.c (build_builtin_expect_predicate): Add predictor
4975 argument, if non-NULL, create 3 argument __builtin_expect.
4976 (fold_builtin_expect): No longer static. Add ARG2 argument,
4977 pass it through to build_builtin_expect_predicate.
4978 (fold_builtin_2): Adjust caller.
4979 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4980 * internal-fn.def (BUILTIN_EXPECT): New.
4981
4982 2014-03-18 Tobias Burnus <burnus@net-b.de>
4983
4984 PR ipa/58721
4985 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4986 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4987 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4988
4989 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4990
4991 PR ipa/58721
4992 * predict.c (combine_predictions_for_bb): Fix up formatting.
4993 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4994 fill what it points to if non-NULL.
4995 (tree_predict_by_opcode): Adjust caller, use the predictor.
4996 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4997
4998 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4999
5000 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5001 proper constant for the store mode.
5002
5003 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5004
5005 * symtab.c (change_decl_assembler_name): Fix transparent alias
5006 chain construction.
5007
5008 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5009
5010 * config/aarch64/aarch64.c: Correct the comments about the
5011 aarch64 stack layout.
5012
5013 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5014
5015 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5016 check for GF_OMP_FOR_KIND_FOR.
5017
5018 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5019
5020 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5021 ymm and zmm register names.
5022
5023 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5024
5025 PR target/60516
5026 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5027 note creation for the 2010-08-31 changes.
5028
5029 2014-03-17 Marek Polacek <polacek@redhat.com>
5030
5031 PR middle-end/60534
5032 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5033 as -fno-tree-loop-vectorize.
5034 (expand_omp_simd): Likewise.
5035
5036 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5037
5038 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5039 (eligible_for_call_delay): New prototype.
5040 * config/sparc/sparc.c (tls_call_delay): Rename into...
5041 (eligible_for_call_delay): ...this. Return false if the instruction
5042 cannot be put in the delay slot of a branch.
5043 (eligible_for_restore_insn): Simplify.
5044 (eligible_for_return_delay): Return false if the instruction cannot be
5045 put in the delay slot of a branch and simplify.
5046 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5047 put in the delay slot of a branch.
5048 * config/sparc/sparc.md (fix_ut699): New attribute.
5049 (tls_call_delay): Delete.
5050 (in_call_delay): Reimplement.
5051 (eligible_for_sibcall_delay): Rename into...
5052 (in_sibcall_delay): ...this.
5053 (eligible_for_return_delay): Rename into...
5054 (in_return_delay): ...this.
5055 (in_branch_delay): Reimplement.
5056 (in_uncond_branch_delay): Delete.
5057 (in_annul_branch_delay): Delete.
5058
5059 2014-03-14 Richard Henderson <rth@redhat.com>
5060
5061 PR target/60525
5062 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5063 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5064 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5065 (floathi splitters): Remove.
5066 (float<SWI48x>xf2): New pattern.
5067 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5068 code that tried to handle DImode for 32-bit, but which was excluded
5069 by the pattern's condition. Drop allocation of stack temporary.
5070 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5071 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5072 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5073 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5074 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5075 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5076 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5077 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5078 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5079 (*float<SWI48x><X87MODEF>2_i387): Remove.
5080 (all float _with_temp splitters): Remove.
5081 (*float<SWI48x><MODEF>2_i387): New pattern.
5082 (*float<SWI48><MODEF>2_sse): New pattern.
5083 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5084 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5085
5086 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5087 Marek Polacek <polacek@redhat.com>
5088
5089 PR middle-end/60484
5090 * common.opt (dump_base_name_prefixed): New Variable.
5091 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5092 if x_dump_base_name_prefixed is already set, set it at the end.
5093
5094 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5095
5096 PR rtl-optimization/60508
5097 * lra-constraints.c (get_reload_reg): Add new parameter
5098 in_subreg_p.
5099 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5100 Pass the new parameter values.
5101
5102 2014-03-14 Richard Biener <rguenther@suse.de>
5103
5104 * common.opt: Revert unintented changes from r205065.
5105 * opts.c: Likewise.
5106
5107 2014-03-14 Richard Biener <rguenther@suse.de>
5108
5109 PR middle-end/60518
5110 * cfghooks.c (split_block): Properly adjust all loops the
5111 block was a latch of.
5112
5113 2014-03-14 Martin Jambor <mjambor@suse.cz>
5114
5115 PR lto/60461
5116 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5117 and simplify it.
5118
5119 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5120
5121 PR target/59396
5122 * config/avr/avr.c (avr_set_current_function): Pass function name
5123 through default_strip_name_encoding before sanity checking instead
5124 of skipping the first char of the assembler name.
5125
5126 2014-03-13 Richard Henderson <rth@redhat.com>
5127
5128 PR debug/60438
5129 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5130 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5131 * config/i386/i386-protos.h: Likewise.
5132 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5133 in the expander instead of a splitter.
5134 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5135 any possibility of requiring a memory.
5136 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5137 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5138 (fp branch splitters): Update for ix86_split_fp_branch.
5139 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5140 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5141 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5142 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5143 (*fop_<MODEF>_3_i387): Likewise.
5144 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5145 (splitters for the fop_* register patterns): Remove.
5146 (fscalexf4_i387): Rename from *fscalexf4_i387.
5147 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5148
5149 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5150
5151 PR tree-optimization/59779
5152 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5153 type for bitsize and maxsize instead of HOST_WIDE_INT.
5154
5155 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5156
5157 PR rtl-optimization/57320
5158 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5159 the CFG after thread_prologue_and_epilogue_insns.
5160
5161 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5162
5163 PR rtl-optimization/57189
5164 * lra-constraints.c (process_alt_operands): Disfavor spilling
5165 vector pseudos.
5166
5167 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5168
5169 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5170
5171 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5172
5173 PR tree-optimization/59025
5174 PR middle-end/60418
5175 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5176 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5177
5178 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5179
5180 PR target/60486
5181 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5182 calls of avr_out_plus_1.
5183
5184 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5185
5186 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5187 BB's single pred and update the father loop's latch info later.
5188
5189 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5190
5191 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5192 (VEC_M): Likewise.
5193 (VEC_N): Likewise.
5194 (VEC_R): Likewise.
5195 (VEC_base): Likewise.
5196 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5197 registers, we need to swap double words in little endian mode.
5198
5199 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5200 to be a container mode for 128-bit integer operations added in ISA
5201 2.07. Unlike TImode and PTImode, the preferred register set is
5202 the Altivec/VMX registers for the 128-bit operations.
5203
5204 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5205 declarations.
5206 (rs6000_split_128bit_ok_p): Likewise.
5207
5208 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5209 macros for creating ISA 2.07 normal and overloaded builtin
5210 functions with 3 arguments.
5211 (BU_P8V_OVERLOAD_3): Likewise.
5212 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5213 for use as overloaded functions.
5214 (VPERM_1TI_UNS): Likewise.
5215 (VSEL_1TI): Likewise.
5216 (VSEL_1TI_UNS): Likewise.
5217 (ST_INTERNAL_1ti): Likewise.
5218 (LD_INTERNAL_1ti): Likewise.
5219 (XXSEL_1TI): Likewise.
5220 (XXSEL_1TI_UNS): Likewise.
5221 (VPERM_1TI): Likewise.
5222 (VPERM_1TI_UNS): Likewise.
5223 (XXPERMDI_1TI): Likewise.
5224 (SET_1TI): Likewise.
5225 (LXVD2X_V1TI): Likewise.
5226 (STXVD2X_V1TI): Likewise.
5227 (VEC_INIT_V1TI): Likewise.
5228 (VEC_SET_V1TI): Likewise.
5229 (VEC_EXT_V1TI): Likewise.
5230 (EQV_V1TI): Likewise.
5231 (NAND_V1TI): Likewise.
5232 (ORC_V1TI): Likewise.
5233 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5234 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5235 overloaded builtin.
5236 (VADDUQM): Likewise.
5237 (VSUBCUQ): Likewise.
5238 (VADDEUQM): Likewise.
5239 (VADDECUQ): Likewise.
5240 (VSUBEUQM): Likewise.
5241 (VSUBECUQ): Likewise.
5242
5243 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5244 __int128_t and __uint128_t types.
5245 (__uint128_type): Likewise.
5246 (altivec_categorize_keyword): Add support for vector __int128_t,
5247 vector __uint128_t, vector __int128, and vector unsigned __int128
5248 as a container type for TImode operations that need to be done in
5249 VSX/Altivec registers.
5250 (rs6000_macro_to_expand): Likewise.
5251 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5252 to support 128-bit integer instructions vaddcuq, vadduqm,
5253 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5254 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5255
5256 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5257 for V1TImode, and set up preferences to use VSX/Altivec registers.
5258 Setup VSX reload handlers.
5259 (rs6000_debug_reg_global): Likewise.
5260 (rs6000_init_hard_regno_mode_ok): Likewise.
5261 (rs6000_preferred_simd_mode): Likewise.
5262 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5263 (easy_altivec_constant): Likewise.
5264 (output_vec_const_move): Likewise.
5265 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5266 simple move.
5267 (rs6000_expand_vector_extract): Likewise.
5268 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5269 addressing.
5270 (rs6000_const_vec): Add support for V1TImode.
5271 (rs6000_emit_le_vsx_load): Swap double words when loading or
5272 storing TImode/V1TImode.
5273 (rs6000_emit_le_vsx_store): Likewise.
5274 (rs6000_emit_le_vsx_move): Likewise.
5275 (rs6000_emit_move): Add support for V1TImode.
5276 (altivec_expand_ld_builtin): Likewise.
5277 (altivec_expand_st_builtin): Likewise.
5278 (altivec_expand_vec_init_builtin): Likewise.
5279 (altivec_expand_builtin): Likewise.
5280 (rs6000_init_builtins): Add support for V1TImode type. Add
5281 support for ISA 2.07 128-bit integer builtins. Define type names
5282 for the VSX/Altivec vector types.
5283 (altivec_init_builtins): Add support for overloaded vector
5284 functions with V1TImode type.
5285 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5286 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5287 external function.
5288 (rs6000_split_128bit_ok_p): Likewise.
5289 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5290 __int128_t and vector __uint128_t.
5291
5292 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5293 and mode attributes.
5294 (VSX_M): Likewise.
5295 (VSX_M2): Likewise.
5296 (VSm): Likewise.
5297 (VSs): Likewise.
5298 (VSr): Likewise.
5299 (VSv): Likewise.
5300 (VS_scalar): Likewise.
5301 (VS_double): Likewise.
5302 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5303
5304 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5305 we support the ISA 2.07 128-bit integer arithmetic instructions.
5306 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5307 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5308 and TImode types for use with the builtin functions.
5309 (V1TI_type_node): Likewise.
5310 (unsigned_V1TI_type_node): Likewise.
5311 (intTI_type_internal_node): Likewise.
5312 (uintTI_type_internal_node): Likewise.
5313
5314 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5315 128-bit builtin functions.
5316 (UNSPEC_VADDEUQM): Likewise.
5317 (UNSPEC_VADDECUQ): Likewise.
5318 (UNSPEC_VSUBCUQ): Likewise.
5319 (UNSPEC_VSUBEUQM): Likewise.
5320 (UNSPEC_VSUBECUQ): Likewise.
5321 (VM): Add V1TImode to vector mode iterators.
5322 (VM2): Likewise.
5323 (VI_unit): Likewise.
5324 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5325 (altivec_vaddcuq): Likewise.
5326 (altivec_vsubuqm): Likewise.
5327 (altivec_vsubcuq): Likewise.
5328 (altivec_vaddeuqm): Likewise.
5329 (altivec_vaddecuq): Likewise.
5330 (altivec_vsubeuqm): Likewise.
5331 (altivec_vsubecuq): Likewise.
5332
5333 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5334 mode iterators.
5335 (BOOL_128): Likewise.
5336 (BOOL_REGS_OUTPUT): Likewise.
5337 (BOOL_REGS_OP1): Likewise.
5338 (BOOL_REGS_OP2): Likewise.
5339 (BOOL_REGS_UNARY): Likewise.
5340 (BOOL_REGS_AND_CR0): Likewise.
5341
5342 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5343 128-bit integer builtin support.
5344 (vec_vadduqm): Likewise.
5345 (vec_vaddecuq): Likewise.
5346 (vec_vaddeuqm): Likewise.
5347 (vec_vsubecuq): Likewise.
5348 (vec_vsubeuqm): Likewise.
5349 (vec_vsubcuq): Likewise.
5350 (vec_vsubuqm): Likewise.
5351
5352 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5353 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5354 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5355 128-bit integer add/subtract to ISA 2.07.
5356
5357 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5358
5359 * config/arc/arc.c (arc_predicate_delay_insns):
5360 Fix third argument passed to conditionalize_nonjump.
5361
5362 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5363
5364 * config/aarch64/aarch64-builtins.c
5365 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5366 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5367 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5368 instead of __builtin_lfloor.
5369 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5370
5371 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5372
5373 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5374 (tree_ssa_ifcombine_bb_1): New function.
5375 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5376 is an empty forwarder block to then_bb or vice versa and then_bb
5377 and else_bb are effectively swapped.
5378
5379 2014-03-12 Christian Bruel <christian.bruel@st.com>
5380
5381 PR target/60264
5382 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5383 REG_CFA_DEF_CFA note.
5384 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5385 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5386
5387 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5388
5389 PR tree-optimization/60454
5390 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5391
5392 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5393
5394 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5395 Do not define target_cpu_default2 to generic.
5396 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5397 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5398 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5399
5400 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5401 Marc Glisse <marc.glisse@inria.fr>
5402
5403 PR tree-optimization/60502
5404 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5405 instead of build_low_bits_mask.
5406
5407 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5408
5409 PR middle-end/60482
5410 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5411 if there are multiple uses, but op doesn't live on E edge.
5412 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5413 clobber stmts before __builtin_unreachable.
5414
5415 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5416
5417 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5418 hard_frame_pointer_rtx.
5419 * cse.c (cse_insn): Remove volatile check.
5420 * cselib.c (cselib_process_insn): Likewise.
5421 * dse.c (scan_insn): Likewise.
5422
5423 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5424
5425 * config/arc/arc.c (conditionalize_nonjump): New function,
5426 broken out of ...
5427 (arc_ifcvt): ... this.
5428 (arc_predicate_delay_insns): Use it.
5429
5430 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5431
5432 * config/arc/predicates.md (extend_operand): During/after reload,
5433 allow const_int_operand.
5434 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5435 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5436 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5437 to "i".
5438 (umulsi3_highpart_i): Likewise.
5439
5440 2014-03-11 Richard Biener <rguenther@suse.de>
5441
5442 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5443 Add asserts to guard possible wrong-code bugs.
5444
5445 2014-03-11 Richard Biener <rguenther@suse.de>
5446
5447 PR tree-optimization/60429
5448 PR tree-optimization/60485
5449 * tree-ssa-structalias.c (set_union_with_increment): Properly
5450 take into account all fields that overlap the shifted vars.
5451 (do_sd_constraint): Likewise.
5452 (do_ds_constraint): Likewise.
5453 (get_constraint_for_ptr_offset): Likewise.
5454
5455 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5456
5457 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5458 (nios2_compute_frame_layout):
5459 Add calculation of cfun->machine->fp_save_offset.
5460 (nios2_expand_prologue): Correct setting of frame pointer register
5461 in prologue.
5462 (nios2_expand_epilogue): Update recovery of stack pointer from
5463 frame pointer accordingly.
5464 (nios2_initial_elimination_offset): Update calculation of offset
5465 for eliminating to HARD_FRAME_POINTER_REGNUM.
5466
5467 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5468
5469 PR ipa/60457
5470 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5471 cgraph_get_create_node on VAR_DECLs.
5472
5473 2014-03-10 Richard Biener <rguenther@suse.de>
5474
5475 PR middle-end/60474
5476 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5477
5478 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5479
5480 * config/vms/vms.opt (vms_float_format): New variable.
5481
5482 2014-03-08 Tobias Burnus <burnus@net-b.de>
5483
5484 * doc/invoke.texi (-fcilkplus): Update implementation status.
5485
5486 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5487 Richard Biener <rguenther@suse.de>
5488
5489 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5490 consistently accross all TUs.
5491 (run_gcc): Enable -fshort-double automatically at link at link-time
5492 and disallow override.
5493
5494 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5495
5496 PR target/58271
5497 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5498 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5499 if they can't be used.
5500
5501 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5502
5503 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5504 for Solaris 11/x86 ld.
5505 * configure: Regenerate.
5506
5507 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5508
5509 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5510 (LIB_TLS_SPEC): Save as ld_tls_libs.
5511 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5512 (HAVE_AS_IX86_TLSLDM): New test.
5513 * configure, config.in: Regenerate.
5514 * config/i386/i386.c (legitimize_tls_address): Fall back to
5515 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5516 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5517 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5518 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5519
5520 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5521
5522 * common.opt (fira-loop-pressure): Mark as optimization.
5523
5524 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5525
5526 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5527 an OpenMP mappable type.
5528
5529 2014-03-06 Matthias Klose <doko@ubuntu.com>
5530
5531 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5532 MULTILIB_OSDIRNAMES is not defined.
5533
5534 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5535 Meador Inge <meadori@codesourcery.com>
5536
5537 PR target/58595
5538 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5539 (arm_legitimize_address): Call legitimize_tls_address for any
5540 arm_tls_referenced_p expression, handle constant addend. Call it
5541 before testing for !TARGET_ARM.
5542 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5543
5544 2014-03-06 Richard Biener <rguenther@suse.de>
5545
5546 PR middle-end/60445
5547 PR lto/60424
5548 PR lto/60427
5549 Revert
5550 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5551
5552 * tree-streamer.c (record_common_node): Assert we don't record
5553 nodes with type double.
5554 (preload_common_node): Skip type double, complex double and double
5555 pointer since it is now frontend dependent due to fshort-double option.
5556
5557 2014-03-06 Richard Biener <rguenther@suse.de>
5558
5559 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5560 or -fno-lto is specified and the linker has full plugin support.
5561 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5562 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5563 * lto-wrapper.c (merge_and_complain): Merge compile-time
5564 optimization levels.
5565 (run_gcc): And pass it through to the link options.
5566
5567 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5568
5569 PR debug/60381
5570 Revert:
5571 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5572 PR debug/59992
5573 * cselib.c (remove_useless_values): Skip to avoid quadratic
5574 behavior if the condition moved from...
5575 (cselib_process_insn): ... here holds.
5576
5577 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5578
5579 PR plugins/59335
5580 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5581 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5582
5583 PR plugins/59335
5584 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5585 (TM_H): Add x86-tune.def.
5586
5587 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5588
5589 * config/aarch64/aarch64.c (generic_tunings):
5590 Use cortexa57_extra_costs.
5591
5592 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5593
5594 PR lto/60404
5595 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5596 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5597 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5598 cost for in_lto_p.
5599
5600 2014-03-04 Heiher <r@hev.cc>
5601
5602 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5603 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5604
5605 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5606
5607 * config/i386/predicates.md (const2356_operand): Change to ...
5608 (const2367_operand): ... this.
5609 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5610 const2367_operand.
5611 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5612 (*avx512pf_scatterpf<mode>sf): Ditto.
5613 (avx512pf_scatterpf<mode>df): Ditto.
5614 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5615 (*avx512pf_scatterpf<mode>df): Ditto.
5616 * config/i386/i386.c (ix86_expand_builtin): Update
5617 incorrect hint operand error message.
5618
5619 2014-03-04 Richard Biener <rguenther@suse.de>
5620
5621 * lto-section-in.c (lto_get_section_data): Fix const cast.
5622
5623 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5624
5625 * tree-streamer.c (record_common_node): Assert we don't record
5626 nodes with type double.
5627 (preload_common_node): Skip type double, complex double and double
5628 pointer since it is now frontend dependent due to fshort-double option.
5629
5630 2014-03-04 Richard Biener <rguenther@suse.de>
5631
5632 PR lto/60405
5633 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5634 (lto_input_toplevel_asms): Likewise.
5635 * lto-section-in.c (lto_get_section_data): Instead do it here
5636 for every section.
5637
5638 2014-03-04 Richard Biener <rguenther@suse.de>
5639
5640 PR tree-optimization/60382
5641 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5642 dead PHIs a reduction.
5643
5644 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5645
5646 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5647 hint value.
5648 (_mm_prefetch): Move out of GCC target("sse") pragma.
5649 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5650 GCC target("prfchw") pragma.
5651 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5652 for locality <= 2.
5653 * config/i386/i386.c (ix86_option_override_internal): Enable
5654 -mprfchw with -mprefetchwt1.
5655
5656 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5657
5658 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5659 Mark as varying.
5660
5661 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5662
5663 * opts.h (CL_PCH_IGNORE): Define.
5664 * targhooks.c (option_affects_pch_p):
5665 Return false for options that have CL_PCH_IGNORE set.
5666 * opt-functions.awk: Process PchIgnore.
5667 * doc/options.texi: Document PchIgnore.
5668
5669 * config/arc/arc.opt (misize): Add PchIgnore property.
5670
5671 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5672
5673 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5674 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5675 constraint on constants to permit them being loaded into
5676 GENERAL_REGS or BASE_REGS.
5677
5678 2014-03-03 Nick Clifton <nickc@redhat.com>
5679
5680 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5681 anti-cacnonical alternatives.
5682 (negandhi3_real): New pattern.
5683 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5684
5685 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5686
5687 * config/avr/avr-mcus.def: Remove atxmega16x1.
5688 * config/avr/avr-tables.opt: Regenerate.
5689 * config/avr/t-multilib: Regenerate.
5690 * doc/avr-mmcu.texi: Regenerate.
5691
5692 2014-03-03 Tobias Grosser <tobias@grosser.es>
5693 Mircea Namolaru <mircea.namolaru@inria.fr>
5694
5695 PR tree-optimization/58028
5696 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5697 scalar dimensions.
5698
5699 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5700
5701 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5702 not handled by recognizers.
5703
5704 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5705
5706 PR middle-end/60175
5707 * function.c (expand_function_end): Don't emit
5708 clobber_return_register sequence if clobber_after is a BARRIER.
5709 * cfgexpand.c (construct_exit_block): Append instructions before
5710 return_label to prev_bb.
5711
5712 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5713
5714 * config/rs6000/constraints.md: Document reserved use of "wc".
5715
5716 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5717
5718 PR ipa/60150
5719 * ipa.c (function_and_variable_visibility): When dissolving comdat
5720 group, also set all symbols to local.
5721
5722 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5723
5724 PR ipa/60306
5725
5726 Revert:
5727 2013-12-14 Jan Hubicka <jh@suse.cz>
5728 PR middle-end/58477
5729 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5730
5731 2014-03-02 Jon Beniston <jon@beniston.com>
5732
5733 PR bootstrap/48230
5734 PR bootstrap/50927
5735 PR bootstrap/52466
5736 PR target/46898
5737 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5738 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5739 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5740 (simple_return, *simple_return): New patterns
5741 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5742 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5743
5744 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5745
5746 * dwarf2out.c (gen_subprogram_die): Tidy.
5747
5748 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5749
5750 PR target/60071
5751 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5752 (*mov_t_msb_neg_negc): ... this new insn.
5753
5754 2014-02-28 Jason Merrill <jason@redhat.com>
5755
5756 PR c++/58678
5757 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5758 function.
5759
5760 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5761
5762 PR c++/60314
5763 * dwarf2out.c (decltype_auto_die): New static.
5764 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5765 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5766 (is_cxx_auto): Likewise.
5767
5768 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5769
5770 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5771 we are not using general regs only.
5772
5773 2014-02-28 Richard Biener <rguenther@suse.de>
5774
5775 PR target/60280
5776 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5777 previous fix and only allow to remove trivial pre-headers
5778 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5779 (remove_forwarder_block): Properly update the latch of a loop.
5780
5781 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5782
5783 PR debug/59992
5784 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5785 (cselib_preserved_hash_table): New.
5786 (preserve_constants_and_equivs): Move preserved vals to it.
5787 (cselib_find_slot): Look it up first.
5788 (cselib_init): Initialize it.
5789 (cselib_finish): Release it.
5790 (dump_cselib_table): Dump it.
5791
5792 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5793
5794 PR debug/59992
5795 * cselib.c (remove_useless_values): Skip to avoid quadratic
5796 behavior if the condition moved from...
5797 (cselib_process_insn): ... here holds.
5798
5799 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5800
5801 PR debug/57232
5802 * var-tracking.c (vt_initialize): Apply the same condition to
5803 preserve the CFA base value.
5804
5805 2014-02-28 Joey Ye <joey.ye@arm.com>
5806
5807 PR target/PR60169
5808 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5809 if reload in progress or completed.
5810
5811 2014-02-28 Tobias Burnus <burnus@net-b.de>
5812
5813 PR middle-end/60147
5814 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5815 NAMELIST_DECL.
5816
5817 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5818
5819 * doc/tm.texi.in (Condition Code Status): Update documention for
5820 relative locations of cc0-setter and cc0-user.
5821
5822 2014-02-27 Jeff Law <law@redhat.com>
5823
5824 PR rtl-optimization/52714
5825 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5826 into two independent simple sets, if I3 is a jump, ensure the
5827 pattern we place into I3 is a (set (pc) ...).
5828
5829 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5830 Jeff Law <law@redhat.com>
5831
5832 PR rtl-optimization/49847
5833 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5834 are in different blocks.
5835 * doc/tm.texi (Condition Code Status): Update documention for
5836 relative locations of cc0-setter and cc0-user.
5837
5838 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5839
5840 PR target/59222
5841 * lra.c (lra_emit_add): Check SUBREG too.
5842
5843 2014-02-27 Andreas Schwab <schwab@suse.de>
5844
5845 * config/m68k/m68k.c (m68k_option_override): Disable
5846 -flive-range-shrinkage for classic m68k.
5847 (m68k_override_options_after_change): Likewise.
5848
5849 2014-02-27 Marek Polacek <polacek@redhat.com>
5850
5851 PR middle-end/59223
5852 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5853 -Wmaybe-uninitialized.
5854
5855 2014-02-27 Alan Modra <amodra@gmail.com>
5856
5857 PR target/57936
5858 * reload1.c (emit_input_reload_insns): When reload_override_in,
5859 set old to rl->in_reg when rl->in_reg is a subreg.
5860
5861 2014-02-26 Richard Biener <rguenther@suse.de>
5862
5863 PR bootstrap/60343
5864 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5865
5866 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5867
5868 * common/config/i386/predicates.md (const1256_operand): Remove.
5869 (const2356_operand): New.
5870 (const_1_to_2_operand): Remove.
5871 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5872 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5873 (*avx512pf_gatherpf<mode>sf): Ditto.
5874 (avx512pf_gatherpf<mode>df): Ditto.
5875 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5876 (*avx512pf_gatherpf<mode>df): Ditto.
5877 (avx512pf_scatterpf<mode>sf): Ditto.
5878 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5879 (*avx512pf_scatterpf<mode>sf): Ditto.
5880 (avx512pf_scatterpf<mode>df): Ditto.
5881 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5882 (*avx512pf_scatterpf<mode>df): Ditto.
5883 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5884
5885 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5886
5887 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5888 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5889 (_mm512_mask_testn_epi64_mask): Move to ...
5890 * config/i386/avx512cdintrin.h: Here.
5891 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5892 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5893 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5894 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5895 TARGET_AVX512F from TARGET_AVX512CD.
5896
5897 2014-02-26 Richard Biener <rguenther@suse.de>
5898
5899 PR ipa/60327
5900 * ipa.c (walk_polymorphic_call_targets): Properly guard
5901 call to inline_update_overall_summary.
5902
5903 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5904
5905 PR target/60280
5906 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5907 and latches only if requested. Fix latch if it is removed.
5908 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5909 LOOPS_HAVE_PREHEADERS.
5910
5911 2014-02-25 Andrew Pinski <apinski@cavium.com>
5912
5913 * builtins.c (expand_builtin_thread_pointer): Create a new target
5914 when the target is NULL.
5915
5916 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5917
5918 PR rtl-optimization/60317
5919 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5920 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5921 * lra-assigns.c: Include params.h.
5922 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5923 other reload pseudos considerations.
5924
5925 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5926
5927 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5928 to use canonical form for nor<mode>3.
5929
5930 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5931
5932 PR target/55426
5933 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5934 conversions.
5935
5936 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5937
5938 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5939 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5940 (ix86_handle_option): Handle OPT_mprefetchwt1.
5941 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5942 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5943 PREFETCHWT1 CPUID.
5944 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5945 OPTION_MASK_ISA_PREFETCHWT1.
5946 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5947 (PTA_PREFETCHWT1): New.
5948 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5949 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5950 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5951 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5952 (*prefetch_avx512pf_<mode>_: Change into ...
5953 (*prefetch_prefetchwt1_<mode>: This.
5954 * config/i386/i386.opt (mprefetchwt1): New.
5955 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5956 (_mm_prefetch): Handle intent to write.
5957 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5958
5959 2014-02-25 Richard Biener <rguenther@suse.de>
5960
5961 PR middle-end/60291
5962 * emit-rtl.c (mem_attrs_htab): Remove.
5963 (mem_attrs_htab_hash): Likewise.
5964 (mem_attrs_htab_eq): Likewise.
5965 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5966 (init_emit_once): Do not allocate mem_attrs_htab.
5967
5968 2014-02-25 Richard Biener <rguenther@suse.de>
5969
5970 PR lto/60319
5971 * lto-opts.c (lto_write_options): Output non-explicit conservative
5972 -fwrapv, -fno-trapv and -fno-strict-overflow.
5973 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5974 (run_gcc): And pass them through.
5975
5976 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5977
5978 * sel-sched.c (calculate_new_fences): New parameter ptime.
5979 Calculate it as a maximum over all fence cycles.
5980 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5981 Print the final schedule timing when sched_verbose.
5982
5983 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5984
5985 PR rtl-optimization/60292
5986 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5987 bit fot the fence instruction.
5988
5989 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5990
5991 * calls.h: Fix typo in comment.
5992
5993 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5994
5995 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5996 adjusting offsetable addresses.
5997
5998 2014-02-24 Guozhi Wei <carrot@google.com>
5999
6000 * sparseset.h (sparseset_pop): Fix the wrong index.
6001
6002 2014-02-24 Walter Lee <walt@tilera.com>
6003
6004 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6005 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6006 triplet.
6007 * common/config/tilegx/tilegx-common.c
6008 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6009 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6010 (LINK_SPEC): Ditto.
6011 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6012 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6013 (tilegx_gimplify_va_arg_expr): Handle big endian.
6014 (tilegx_expand_unaligned_load): Ditto.
6015 (tilegx_expand_unaligned_store): Ditto.
6016 (TARGET_RETURN_IN_MSB): New.
6017 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6018 (TARGET_ENDIAN_DEFAULT): New.
6019 (TARGET_BIG_ENDIAN): Handle big endian.
6020 (BYTES_BIG_ENDIAN): Ditto.
6021 (WORDS_BIG_ENDIAN): Ditto.
6022 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6023 (ENDIAN_SPEC): New.
6024 (EXTRA_SPECS): New.
6025 * config/tilegx/tilegx.md (extv): Handle big endian.
6026 (extzv): Ditto.
6027 (insn_st<n>): Ditto.
6028 (insn_st<n>_add<bitsuffix>): Ditto.
6029 (insn_stnt<n>): Ditto.
6030 (insn_stnt<n>_add<bitsuffix>):Ditto.
6031 (vec_interleave_highv8qi): Handle big endian.
6032 (vec_interleave_highv8qi_be): New.
6033 (vec_interleave_highv8qi_le): New.
6034 (insn_v1int_h): Handle big endian.
6035 (vec_interleave_lowv8qi): Handle big endian.
6036 (vec_interleave_lowv8qi_be): New.
6037 (vec_interleave_lowv8qi_le): New.
6038 (insn_v1int_l): Handle big endian.
6039 (vec_interleave_highv4hi): Handle big endian.
6040 (vec_interleave_highv4hi_be): New.
6041 (vec_interleave_highv4hi_le): New.
6042 (insn_v2int_h): Handle big endian.
6043 (vec_interleave_lowv4hi): Handle big endian.
6044 (vec_interleave_lowv4hi_be): New.
6045 (vec_interleave_lowv4hi_le): New.
6046 (insn_v2int_l): Handle big endian.
6047 (vec_interleave_highv2si): Handle big endian.
6048 (vec_interleave_highv2si_be): New.
6049 (vec_interleave_highv2si_le): New.
6050 (insn_v4int_h): Handle big endian.
6051 (vec_interleave_lowv2si): Handle big endian.
6052 (vec_interleave_lowv2si_be): New.
6053 (vec_interleave_lowv2si_le): New.
6054 (insn_v4int_l): Handle big endian.
6055 * config/tilegx/tilegx.opt (mbig-endian): New option.
6056 (mlittle-endian): New option.
6057 * doc/install.texi: Document tilegxbe-linux.
6058 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6059
6060 2014-02-24 Martin Jambor <mjambor@suse.cz>
6061
6062 PR ipa/60266
6063 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6064 there are no parameter descriptors.
6065
6066 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6067
6068 PR rtl-optimization/60268
6069 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6070 initialization to ...
6071 (sched_rgn_init): ... here.
6072 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6073
6074 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6075
6076 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6077 names.
6078
6079 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6080
6081 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6082 definition.
6083
6084 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6085
6086 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6087 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6088
6089 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6090
6091 * config/microblaze/predicates.md: Add cmp_op predicate.
6092 * config/microblaze/microblaze.md: Add branch_compare instruction
6093 which uses cmp_op predicate and emits cmp insn before branch.
6094 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6095 to microblaze_expand_conditional_branch and consolidate logic.
6096 (microblaze_expand_conditional_branch): emit branch_compare
6097 insn instead of handling cmp op separate from branch insn.
6098
6099 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6100
6101 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6102 to permit subregs.
6103
6104 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6105
6106 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6107 define_insn with define_expand and new define_insn
6108 *altivec_lve<VI_char>x_internal.
6109 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6110 and new define_insn *altivec_stve<VI_char>x_internal.
6111 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6112 prototype.
6113 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6114 lve*x built-ins.
6115 (altivec_expand_stvex_be): New function.
6116
6117 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6118
6119 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6120 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6121 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6122 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6123
6124 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6125
6126 PR target/60298
6127 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6128 instead of emit_move_insn.
6129
6130 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6131
6132 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6133 vspltw with vsldoi.
6134 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6135 gen_altivec_vsumsws.
6136
6137 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6138
6139 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6140 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6141 (altivec_lvxl_<mode>): New define_expand incorporating
6142 -maltivec=be semantics where needed.
6143 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6144 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6145 semantics where needed.
6146 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6147 (altivec_stvx_<mode>): New define_expand incorporating
6148 -maltivec=be semantics where needed.
6149 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6150 VM2 iterator instead of V4SI.
6151 (altivec_stvxl_<mode>): New define_expand incorporating
6152 -maltivec=be semantics where needed.
6153 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6154 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6155 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6156 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6157 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6158 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6159 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6160 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6161 ALTIVEC_BUILTIN_STVXL.
6162 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6163 (altivec_expand_stvx_be): Likewise.
6164 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6165 (altivec_expand_lvx_be): Likewise.
6166 (altivec_expand_stvx_be): Likewise.
6167 (altivec_expand_builtin): Add cases for
6168 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6169 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6170 (altivec_init_builtins): Add definitions for
6171 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6172 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6173
6174 2014-02-21 Catherine Moore <clm@codesourcery.com>
6175
6176 * doc/invoke.texi (mvirt, mno-virt): Document.
6177 * config/mips/mips.opt (mvirt): New option.
6178 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6179
6180 2014-02-21 Richard Biener <rguenther@suse.de>
6181
6182 PR tree-optimization/60276
6183 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6184 (STMT_VINFO_MIN_NEG_DIST): New macro.
6185 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6186 STMT_VINFO_MIN_NEG_DIST.
6187 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6188 made for negative dependence distances still hold.
6189
6190 2014-02-21 Richard Biener <rguenther@suse.de>
6191
6192 PR middle-end/60291
6193 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6194 DECL_INITIAL for globals not in the current function context.
6195
6196 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6197
6198 PR tree-optimization/56490
6199 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6200 * tree-ssa-uninit.c: Include params.h.
6201 (compute_control_dep_chain): Add num_calls argument, return false
6202 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6203 num_calls to recursive call.
6204 (find_predicates): Change dep_chain into normal array,
6205 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6206 variable and adjust compute_control_dep_chain caller.
6207 (find_def_preds): Likewise.
6208
6209 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6210
6211 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6212 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6213
6214 2014-02-21 Nick Clifton <nickc@redhat.com>
6215
6216 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6217 (pushhi1): Likewise.
6218 (popqi1): Add mode to pre_dec.
6219 (pophi1): Likewise.
6220
6221 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6222
6223 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6224 mode for mask of V8SFmode permutation.
6225
6226 2014-02-20 Richard Henderson <rth@redhat.com>
6227
6228 PR c++/60272
6229 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6230 a new pseudo for OLDVAL.
6231
6232 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6233
6234 PR target/57896
6235 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6236 gen_reg_rtx if d->testing_p.
6237 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6238 if d->testing_p and we will certainly return true.
6239 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6240 if d->testing_p.
6241
6242 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6243
6244 * emit-rtl.c (gen_reg_rtx): Assert that
6245 crtl->emit.regno_pointer_align_length is non-zero.
6246
6247 2014-02-20 Richard Henderson <rth@redhat.com>
6248
6249 PR c++/60272
6250 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6251 on failure the store back into EXPECT.
6252
6253 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6254 Sandra Loosemore <sandra@codesourcery.com>
6255
6256 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6257 * config/nios2/nios2.c (nios2_function_profiler): Add
6258 -fPIC (flag_pic == 2) support.
6259 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6260 (nios2_large_offset_p): New function.
6261 (nios2_unspec_reloc_p): Move up position, update to use
6262 nios2_large_offset_p.
6263 (nios2_unspec_address): Remove function.
6264 (nios2_unspec_offset): New function.
6265 (nios2_large_got_address): New function.
6266 (nios2_got_address): Add large offset support.
6267 (nios2_legitimize_tls_address): Update usage of removed and new
6268 functions.
6269 (nios2_symbol_binds_local_p): New function.
6270 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6271 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6272 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6273 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6274 processing for (const (unspec ...)).
6275 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6276
6277 2014-02-20 Richard Biener <rguenther@suse.de>
6278
6279 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6280 doing the substitution.
6281 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6282
6283 2014-02-20 Martin Jambor <mjambor@suse.cz>
6284
6285 PR ipa/55260
6286 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6287 info when checking whether lattices are bottom.
6288
6289 2014-02-20 Richard Biener <rguenther@suse.de>
6290
6291 PR middle-end/60221
6292 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6293 regions at -O0.
6294
6295 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6296
6297 PR ipa/58555
6298 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6299 parameter specifying the scaling.
6300 (inline_call): Update.
6301 (want_inline_recursively): Guard division by zero.
6302 (recursive_inlining): Update.
6303 * ipa-inline.h (clone_inlined_nodes): Update.
6304
6305 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6306
6307 PR target/60204
6308 * config/i386/i386.c (classify_argument): Pass structures of size
6309 64 bytes or less in register.
6310
6311 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6312 Kirill Yukhin <kirill.yukhin@intel.com>
6313
6314 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6315 (_mm_rcp28_round_ss): Ditto.
6316 (_mm_rsqrt28_round_sd): Ditto.
6317 (_mm_rsqrt28_round_ss): Ditto.
6318 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6319 (_mm_rcp14_round_ss): Ditto.
6320 (_mm_rsqrt14_round_sd): Ditto.
6321 (_mm_rsqrt14_round_ss): Ditto.
6322 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6323 the first input operand, get rid of match_dup.
6324 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6325 attribute to sse.
6326 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6327 Ditto.
6328 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6329 operand as the first input operand, set type attribute.
6330 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6331 Set type attribute.
6332 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6333 operand as the first input operand, set type attribute.
6334
6335 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6336
6337 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6338 bit of zero.
6339
6340 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6341
6342 PR target/60207
6343 * config/i386/i386.c (construct_container): Remove TFmode check
6344 for X86_64_INTEGER_CLASS.
6345
6346 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6347
6348 PR target/59794
6349 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6350 only when -Wpsabi is enabled.
6351
6352 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6353
6354 PR target/59799
6355 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6356 passing arrays in registers are the same as for structs, so remove the
6357 special case for them.
6358
6359 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6360
6361 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6362 destination type, extract only the valid bits if the source type is not
6363 integral and has a different mode.
6364
6365 2014-02-19 Richard Biener <rguenther@suse.de>
6366
6367 PR ipa/60243
6368 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6369 for all calls.
6370
6371 2014-02-19 Richard Biener <rguenther@suse.de>
6372
6373 PR ipa/60243
6374 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6375 (ipa_modify_call_arguments): Emit an argument load explicitely and
6376 preserve virtual SSA form there and for the replacement call.
6377 Do not update SSA form nor free dominance info.
6378
6379 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6380
6381 * ipa.c (function_and_variable_visibility): Also clear WEAK
6382 flag when disolving COMDAT_GROUP.
6383
6384 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6385
6386 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6387 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6388 not devirtualizing.
6389 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6390 do more sanity checks.
6391 (detect_type_change): Return true when giving up early.
6392 (compute_complex_assign_jump_func): Fix type parameter of
6393 ipa_set_ancestor_jf.
6394 (compute_complex_ancestor_jump_func): Likewise.
6395 (update_jump_functions_after_inlining): Fix updating of
6396 ancestor function.
6397 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6398
6399 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6400
6401 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6402 inline clones when edge disappears.
6403
6404 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6405
6406 PR target/60203
6407 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6408 Split 64-bit moves into 2 patterns. Do not allow the use of
6409 direct move for TDmode in little endian, since the decimal value
6410 has little endian bytes within a word, but the 64-bit pieces are
6411 ordered in a big endian fashion, and normal subreg's of TDmode are
6412 not allowed.
6413 (mov<mode>_64bit_dm): Likewise.
6414 (movtd_64bit_nodm): Likewise.
6415
6416 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6417
6418 PR tree-optimization/60174
6419 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6420 statement of an SSA_NAME that occurs in an abnormal PHI node.
6421
6422 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6423
6424 PR sanitizer/60142
6425 * final.c (SEEN_BB): Remove.
6426 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6427 (final_scan_insn): Don't force_source_line on second
6428 NOTE_INSN_BASIC_BLOCK.
6429
6430 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6431
6432 PR target/60205
6433 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6434 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6435 (type_natural_mode): Warn ABI change when %zmm register is not
6436 available for AVX512F vector value passing.
6437
6438 2014-02-18 Kai Tietz <ktietz@redhat.com>
6439
6440 PR target/60193
6441 * config/i386/i386.c (ix86_expand_prologue): Use value in
6442 rax register as displacement when restoring %r10 or %rax.
6443 Fix wrong offset when restoring both registers.
6444
6445 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6446
6447 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6448 assertion with conditional return.
6449
6450 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6451 Uros Bizjak <ubizjak@gmail.com>
6452
6453 PR driver/60233
6454 * config/i386/driver-i386.c (host_detect_local_cpu): If
6455 YMM state is not saved by the OS, also clear has_f16c. Move
6456 CPUID 0x80000001 handling before YMM state saving checking.
6457
6458 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6459
6460 PR rtl-optimization/58960
6461 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6462 factored out from ...
6463 (sched_init): ... here.
6464 (free_global_sched_pressure_data): New, factored out from ...
6465 (sched_finish): ... here.
6466 * sched-int.h (free_global_sched_pressure_data): Declare.
6467 * sched-rgn.c (nr_regions_initial): New static global.
6468 (haifa_find_rgns): Initialize it.
6469 (schedule_region): Disable sched-pressure for the newly
6470 generated regions.
6471
6472 2014-02-17 Richard Biener <rguenther@suse.de>
6473
6474 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6475 release SSA defs of pattern stmts.
6476
6477 2014-02-17 Richard Biener <rguenther@suse.de>
6478
6479 * tree-inline.c (expand_call_inline): Release the virtual
6480 operand defined by the call we are about to inline.
6481
6482 2014-02-17 Richard Biener <rguenther@suse.de>
6483
6484 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6485
6486 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6487 Ilya Tocar <ilya.tocar@intel.com>
6488
6489 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6490 arguments order in builtin.
6491 (_mm512_permutexvar_epi64): Ditto.
6492 (_mm512_mask_permutexvar_epi64): Ditto
6493 (_mm512_maskz_permutexvar_epi32): Ditto
6494 (_mm512_permutexvar_epi32): Ditto
6495 (_mm512_mask_permutexvar_epi32): Ditto
6496
6497 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6498
6499 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6500 (p8_vmrgow): Likewise.
6501
6502 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6503
6504 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6505 endian targets.
6506
6507 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6508
6509 PR target/60203
6510 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6511 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6512 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6513 using direct move instructions on ISA 2.07. Also adjust
6514 instruction length for 64-bit.
6515 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6516 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6517
6518 2014-02-15 Alan Modra <amodra@gmail.com>
6519
6520 PR target/58675
6521 PR target/57935
6522 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6523 find_replacement on parts of insn rtl that might be reloaded.
6524
6525 2014-02-15 Richard Biener <rguenther@suse.de>
6526
6527 PR tree-optimization/60183
6528 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6529 (tree_ssa_phiprop): Calculate and free post-dominators.
6530
6531 2014-02-14 Jeff Law <law@redhat.com>
6532
6533 PR rtl-optimization/60131
6534 * ree.c (get_extended_src_reg): New function.
6535 (combine_reaching_defs): Use it rather than assuming location of REG.
6536 (find_and_remove_re): Verify first operand of extension is
6537 a REG before adding the insns to the copy list.
6538
6539 2014-02-14 Roland McGrath <mcgrathr@google.com>
6540
6541 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6542 * configure: Regenerated.
6543 * config.in: Regenerated.
6544 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6545 instead of ASM_SHORT.
6546
6547 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6548 Richard Earnshaw <rearnsha@arm.com>
6549
6550 PR rtl-optimization/59535
6551 * lra-constraints.c (process_alt_operands): Encourage alternative
6552 when unassigned pseudo class is superset of the alternative class.
6553 (inherit_reload_reg): Don't inherit when optimizing for code size.
6554 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6555 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6556 modes not less than 4 for Thumb1.
6557
6558 2014-02-14 Kyle McMartin <kyle@redhat.com>
6559
6560 PR pch/60010
6561 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6562
6563 2014-02-14 Richard Biener <rguenther@suse.de>
6564
6565 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6566 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6567 Do not strip INDIRECT_REFs.
6568
6569 2014-02-14 Richard Biener <rguenther@suse.de>
6570
6571 PR lto/60179
6572 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6573 DECL_FUNCTION_SPECIFIC_TARGET.
6574 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6575 * tree-streamer-out.c (pack_ts_target_option): Remove.
6576 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6577 (write_ts_function_decl_tree_pointers): Do not stream
6578 DECL_FUNCTION_SPECIFIC_TARGET.
6579 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6580 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6581 (lto_input_ts_function_decl_tree_pointers): Do not stream
6582 DECL_FUNCTION_SPECIFIC_TARGET.
6583
6584 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6585
6586 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6587 (get_initial_def_for_induction, vectorizable_induction): Ignore
6588 debug stmts when looking for exit_phi.
6589 (vectorizable_live_operation): Fix up condition.
6590
6591 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6592
6593 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6594 nreverse() because it changes the content of original tree list.
6595
6596 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6597
6598 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6599 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6600
6601 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6602
6603 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6604 GNU coding standards.
6605
6606 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6607
6608 PR debug/60152
6609 * dwarf2out.c (gen_subprogram_die): Don't call
6610 add_calling_convention_attribute if subr_die is old_die.
6611
6612 2014-02-13 Sharad Singhai <singhai@google.com>
6613
6614 * doc/optinfo.texi: Fix order of nodes.
6615
6616 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6617
6618 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6619 operands[2], not operands[3].
6620
6621 2014-02-13 Richard Biener <rguenther@suse.de>
6622
6623 PR bootstrap/59878
6624 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6625 mention the possibility of an in-tree build.
6626 (CLooG): Update recommended version to 0.18.1, mention the
6627 possibility of an in-tree build and clarify that the ISL
6628 bundled with CLooG does not work.
6629
6630 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6631
6632 PR target/43546
6633 * expr.c (compress_float_constant): If x is a hard register,
6634 extend into a pseudo and then move to x.
6635
6636 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6637
6638 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6639 caused by bad second argument to warning_at() with -mhotpatch and
6640 nested functions (e.g. with gfortran).
6641
6642 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6643
6644 * opts.c (option_name): Remove "enabled by default" rider.
6645
6646 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6647
6648 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6649
6650 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6651 Uros Bizjak <ubizjak@gmail.com>
6652
6653 PR target/60151
6654 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6655 * configure: Regenerated.
6656
6657 2014-02-12 Richard Biener <rguenther@suse.de>
6658
6659 * vec.c (vec_prefix::calculate_allocation): Move as
6660 inline variant to vec.h.
6661 (vec_prefix::calculate_allocation_1): New out-of-line version.
6662 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6663 (vec_prefix::m_has_auto_buf): Rename to ...
6664 (vec_prefix::m_using_auto_storage): ... this.
6665 (vec_prefix::calculate_allocation): Inline the easy cases
6666 and dispatch to calculate_allocation_1 which doesn't need the
6667 prefix address.
6668 (va_heap::reserve): Use gcc_checking_assert.
6669 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6670 m_using_auto_storage.
6671 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6672 member and adjust.
6673 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6674 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6675 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6676
6677 2014-02-12 Richard Biener <rguenther@suse.de>
6678
6679 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6680 when we found a dependence.
6681
6682 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6683
6684 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6685 common code...
6686 (maybe_fold_stmt): ... into this new function.
6687 * omp-low.c (lower_omp): Update comment.
6688
6689 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6690 last use.
6691
6692 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6693 dereference.
6694
6695 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6696
6697 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6698 identifiers in comments.
6699 (cortexa53_extra_costs): Likewise.
6700 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6701 (cortexa7_extra_costs): Likewise.
6702 (cortexa12_extra_costs): Likewise.
6703 (cortexa15_extra_costs): Likewise.
6704 (v7m_extra_costs): Likewise.
6705
6706 2014-02-12 Richard Biener <rguenther@suse.de>
6707
6708 PR middle-end/60092
6709 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6710 of posix_memalign being successful.
6711 (lower_stmt): Restrict lowering of posix_memalign to when
6712 -ftree-bit-ccp is enabled.
6713
6714 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6715
6716 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6717 arg_loc.
6718 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6719
6720 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6721
6722 PR rtl-optimization/60116
6723 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6724 other_insn once the combination has been validated.
6725
6726 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6727
6728 PR lto/59468
6729 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6730 and wrapper.
6731 * ipa-devirt.c: Include demangle.h
6732 (odr_violation_reported): New static variable.
6733 (add_type_duplicate): Update odr_violations.
6734 (maybe_record_node): Add completep parameter; update it.
6735 (record_target_from_binfo): Add COMPLETEP parameter;
6736 update it as needed.
6737 (possible_polymorphic_call_targets_1): Likewise.
6738 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6739 rename FINAL to COMPLETE.
6740 (record_targets_from_bases): Sanity check we found the binfo;
6741 fix COMPLETEP updating.
6742 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6743 parameter, fix computing of COMPLETEP.
6744 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6745 at LTO time do demangling.
6746 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6747 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6748 parameter.
6749 (gimple_get_virt_method_for_binfo): Likewise.
6750 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6751 gimple_get_virt_method_for_vtable): Update prototypes.
6752
6753 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6754
6755 PR target/49008
6756 * genautomata.c (add_presence_absence): Fix typo with
6757 {final_}presence_list.
6758
6759 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6760
6761 PR target/60137
6762 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6763 for VSX/Altivec vectors that land in GPR registers.
6764
6765 2014-02-11 Richard Henderson <rth@redhat.com>
6766 Jakub Jelinek <jakub@redhat.com>
6767
6768 PR debug/59776
6769 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6770 around drhs if type conversion to lacc->type is not useless.
6771
6772 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6773
6774 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6775 tuning struct.
6776 (cortex-a57.cortex-a53): Likewise.
6777 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6778
6779 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6780
6781 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6782 arm_restrict_it.
6783
6784 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6785
6786 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6787 add_options_for_arm_vfp3.
6788
6789 2014-02-11 Jeff Law <law@redhat.com>
6790
6791 PR middle-end/54041
6792 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6793 object with an undesirable mode.
6794
6795 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6796
6797 PR libgomp/60107
6798 * config/i386/sol2-9.h: New file.
6799 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6800 *-*-solaris2.9*): Use it.
6801
6802 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6803
6804 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6805 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6806
6807 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6808
6809 * config/microblaze/microblaze.c: Extend mcpu version format
6810
6811 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6812
6813 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6814
6815 2014-02-10 Richard Henderson <rth@redhat.com>
6816
6817 PR target/59927
6818 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6819 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6820 ms-abi vs -mno-accumulate-outgoing-args.
6821 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6822 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6823 respect to ms-abi.
6824
6825 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6826
6827 PR middle-end/60080
6828 * cfgexpand.c (expand_asm_operands): Attach source location to
6829 ASM_INPUT rtx objects.
6830 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6831
6832 2014-02-10 Nick Clifton <nickc@redhat.com>
6833
6834 * config/mn10300/mn10300.c (popcount): New function.
6835 (mn10300_expand_prologue): Include saved registers in stack usage
6836 count.
6837
6838 2014-02-10 Jeff Law <law@redhat.com>
6839
6840 PR middle-end/52306
6841 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6842 when changing the SET_DEST of a prior insn to avoid an input reload.
6843
6844 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6845
6846 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6847 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6848 -mcall-openbsd, or -mcall-linux.
6849 (CC1_ENDIAN_BIG_SPEC): Remove.
6850 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6851 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6852 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6853 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6854 and %cc1_endian_default.
6855 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6856
6857 2014-02-10 Richard Biener <rguenther@suse.de>
6858
6859 PR tree-optimization/60115
6860 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6861 MEM_REF handling. Properly verify that the accesses are not
6862 out of the objects bound.
6863
6864 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6865
6866 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6867 coretex to cortex.
6868
6869 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6870
6871 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6872 proper constants and fix formatting.
6873 (possible_polymorphic_call_targets): Fix formatting.
6874
6875 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6876 Ilya Tocar <ilya.tocar@intel.com>
6877
6878 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6879 (_mm512_loadu_epi32): Renamed into...
6880 (_mm512_loadu_si512): This.
6881 (_mm512_storeu_epi32): Renamed into...
6882 (_mm512_storeu_si512): This.
6883 (_mm512_maskz_ceil_ps): Removed.
6884 (_mm512_maskz_ceil_pd): Ditto.
6885 (_mm512_maskz_floor_ps): Ditto.
6886 (_mm512_maskz_floor_pd): Ditto.
6887 (_mm512_floor_round_ps): Ditto.
6888 (_mm512_floor_round_pd): Ditto.
6889 (_mm512_ceil_round_ps): Ditto.
6890 (_mm512_ceil_round_pd): Ditto.
6891 (_mm512_mask_floor_round_ps): Ditto.
6892 (_mm512_mask_floor_round_pd): Ditto.
6893 (_mm512_mask_ceil_round_ps): Ditto.
6894 (_mm512_mask_ceil_round_pd): Ditto.
6895 (_mm512_maskz_floor_round_ps): Ditto.
6896 (_mm512_maskz_floor_round_pd): Ditto.
6897 (_mm512_maskz_ceil_round_ps): Ditto.
6898 (_mm512_maskz_ceil_round_pd): Ditto.
6899 (_mm512_expand_pd): Ditto.
6900 (_mm512_expand_ps): Ditto.
6901 * config/i386/i386.c (ix86_builtins): Remove
6902 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6903 (bdesc_args): Ditto.
6904 * config/i386/predicates.md (const1256_operand): New.
6905 (const_1_to_2_operand): Ditto.
6906 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6907 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6908 (*avx512pf_gatherpf<mode>sf): Ditto.
6909 (avx512pf_gatherpf<mode>df): Ditto.
6910 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6911 (*avx512pf_gatherpf<mode>df): Ditto.
6912 (avx512pf_scatterpf<mode>sf): Ditto.
6913 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6914 (*avx512pf_scatterpf<mode>sf): Ditto.
6915 (avx512pf_scatterpf<mode>df): Ditto.
6916 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6917 (*avx512pf_scatterpf<mode>df): Ditto.
6918 (avx512f_expand<mode>): Removed.
6919 (<shift_insn><mode>3<mask_name>): Change predicate type.
6920
6921 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6922
6923 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6924 not at the end of datarefs vector use ordered_remove to avoid
6925 reordering datarefs vector.
6926
6927 PR c/59984
6928 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6929 mark local addressable non-static vars as GOVD_PRIVATE
6930 instead of GOVD_LOCAL.
6931 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6932 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6933 than copying them.
6934
6935 PR middle-end/60092
6936 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6937 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6938 assume_aligned or alloc_align attributes.
6939 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6940 arguments. Handle also assume_aligned and alloc_align attributes.
6941 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6942 calls to functions with assume_aligned or alloc_align attributes.
6943 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6944
6945 2014-02-08 Terry Guo <terry.guo@arm.com>
6946
6947 * doc/invoke.texi: Document ARM -march=armv7e-m.
6948
6949 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6950
6951 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6952 flag on __cilkrts_rethrow builtin.
6953
6954 PR ipa/60026
6955 * ipa-cp.c (determine_versionability): Fail at -O0
6956 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6957 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6958
6959 Revert:
6960 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6961
6962 PR ipa/60026
6963 * tree-inline.c (copy_forbidden): Fail for
6964 __attribute__((optimize (0))) functions.
6965
6966 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6967
6968 * varpool.c: Include pointer-set.h.
6969 (varpool_remove_unreferenced_decls): Variables in other partitions
6970 will not be output; be however careful to not lose information
6971 about partitioning.
6972
6973 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6974
6975 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6976 lookup in the vtable constructor.
6977
6978 2014-02-07 Jeff Law <law@redhat.com>
6979
6980 PR target/40977
6981 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6982 define_insn_and_split.
6983
6984 * ipa-inline.c (inline_small_functions): Fix typos.
6985
6986 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6987
6988 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6989 (s390_can_use_return_insn): Declare.
6990 * config/s390/s390.h (EPILOGUE_USES): Define.
6991 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6992 instructions.
6993 (s390_chunkify_start): Handle return JUMP_LABELs.
6994 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6995 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6996 (s390_can_use_return_insn): New functions.
6997 (s390_fix_long_loop_prediction): Handle conditional returns.
6998 (TARGET_SET_UP_BY_PROLOGUE): Define.
6999 * config/s390/s390.md (ANY_RETURN): New code iterator.
7000 (*creturn, *csimple_return, return, simple_return): New patterns.
7001
7002 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7003
7004 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7005 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7006 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7007 REG_CFA_RESTORE list when deciding not to restore a register.
7008
7009 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7010
7011 * config/s390/s390.c: Include tree-pass.h and context.h.
7012 (s390_early_mach): New function, split out from...
7013 (s390_emit_prologue): ...here.
7014 (pass_data_s390_early_mach): New pass structure.
7015 (pass_s390_early_mach): New class.
7016 (s390_option_override): Create and register early_mach pass.
7017 Move to end of file.
7018
7019 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7020
7021 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7022 to match for the exit block.
7023
7024 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7025
7026 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7027 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7028 Reject misaligned operands.
7029
7030 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7031
7032 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7033
7034 2014-02-07 Richard Biener <rguenther@suse.de>
7035
7036 PR middle-end/60092
7037 * gimple-low.c (lower_builtin_posix_memalign): New function.
7038 (lower_stmt): Call it to lower posix_memalign in a way
7039 to make alignment info accessible.
7040
7041 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7042
7043 PR c++/60082
7044 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7045 __builtin_setjmp_receiver.
7046
7047 2014-02-07 Richard Biener <rguenther@suse.de>
7048
7049 PR middle-end/60092
7050 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7051 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7052 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7053 Handle BUILT_IN_POSIX_MEMALIGN.
7054 (find_func_clobbers): Likewise.
7055 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7056 (call_may_clobber_ref_p_1): Likewise.
7057
7058 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7059
7060 PR ipa/59918
7061 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7062 sanity check.
7063
7064 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7065
7066 PR ipa/59469
7067 * lto-cgraph.c (lto_output_node): Use
7068 symtab_get_symbol_partitioning_class.
7069 (lto_output_varpool_node): likewise.
7070 (symtab_get_symbol_partitioning_class): Move here from
7071 lto/lto-partition.c
7072 * cgraph.h (symbol_partitioning_class): Likewise.
7073 (symtab_get_symbol_partitioning_class): Declare.
7074
7075 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7076
7077 * ggc.h (ggc_internal_cleared_alloc): New macro.
7078 * vec.h (vec_safe_copy): Handle memory stats.
7079 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7080 * target-globals.c (save_target_globals): Likewise.
7081
7082 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7083
7084 PR target/60077
7085 * expr.c (emit_move_resolve_push): Export; be bit more selective
7086 on when to clear alias set.
7087 * expr.h (emit_move_resolve_push): Declare.
7088 * function.h (struct function): Add tail_call_marked.
7089 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7090 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7091 * config/i386/i386.md (TImode move expander): De not call
7092 ix86_expand_push.
7093 (FP push expanders): Preserve memory attributes.
7094 * config/i386/sse.md (push<mode>1): Remove.
7095 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7096 (ix86_expand_push): Remove.
7097 * config/i386/mmx.md (push<mode>1): Remove.
7098
7099 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7100
7101 PR rtl-optimization/60030
7102 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7103 lopart with paradoxical subreg before shifting it up by hprec.
7104
7105 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7106
7107 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7108 Remove extra newline at end of file.
7109 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7110 (arm_issue_rate): Handle cortexa57.
7111 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7112 (cortex-a57.cortex-a53): Likewise.
7113
7114 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7115
7116 PR target/59575
7117 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7118 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7119 bitmask.
7120 (arm_expand_prologue): Adjust all callers.
7121 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7122 info, registers also at the lowest numbered registers side. Use
7123 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7124 XEXP.
7125
7126 PR debug/59992
7127 * var-tracking.c (adjust_mems): Before adding a SET to
7128 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7129
7130 2014-02-06 Alan Modra <amodra@gmail.com>
7131
7132 PR target/60032
7133 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7134 change SDmode to DDmode when lra_in_progress.
7135
7136 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7137
7138 PR middle-end/59150
7139 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7140 free_data_ref on the dr first, and before goto again also set dr
7141 to the next dr. For simd_lane_access, free old datarefs[i] before
7142 overwriting it. For get_vectype_for_scalar_type failure, don't
7143 free_data_ref if simd_lane_access.
7144
7145 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7146
7147 PR target/60062
7148 * tree.h (opts_for_fn): New inline function.
7149 (opt_for_fn): Define.
7150 * config/i386/i386.c (ix86_function_regparm): Use
7151 opt_for_fn (decl, optimize) instead of optimize.
7152
7153 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7154
7155 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7156 for SYMBOL_REF in large memory model.
7157
7158 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7159
7160 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7161 and crypto support.
7162 (cortex-a57): Likewise.
7163 (cortex-a57.cortex-a53): Likewise.
7164
7165 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7166 Kugan Vivekanandarajah <kuganv@linaro.org>
7167
7168 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7169 unaligned_access.
7170 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7171
7172 2014-02-06 Richard Biener <rguenther@suse.de>
7173
7174 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7175 set_loop_copy and initialize_original_copy_tables.
7176
7177 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7178
7179 * config/aarch64/aarch64-simd.md
7180 (aarch64_ashr_simddi): Change QI to SI.
7181
7182 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7183 Jakub Jelinek <jakub@redhat.com>
7184
7185 PR middle-end/60013
7186 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7187 of the dataflow.
7188
7189 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7190
7191 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7192 CODE_FOR_altivec_vpku[hw]um to
7193 CODE_FOR_altivec_vpku[hw]um_direct.
7194 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7195 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7196 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7197 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7198
7199 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7200
7201 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7202 generation for -maltivec=be.
7203 (altivec_vsumsws): Simplify redundant test.
7204
7205 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7206
7207 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7208 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7209 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7210 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7211 gen_altivec_vpkuwum.
7212 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7213 BYTES_BIG_ENDIAN.
7214 (altivec_vpks<VI_char>ss): Likewise.
7215 (altivec_vpks<VI_char>us): Likewise.
7216 (altivec_vpku<VI_char>us): Likewise.
7217 (altivec_vpku<VI_char>um): Likewise.
7218 (altivec_vpku<VI_char>um_direct): New (copy of
7219 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7220 internal use).
7221 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7222 target is little endian and -maltivec=be is not specified.
7223 (*altivec_vupkhs<VU_char>_direct): New (copy of
7224 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7225 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7226 target is little endian and -maltivec=be is not specified.
7227 (*altivec_vupkls<VU_char>_direct): New (copy of
7228 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7229 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7230 little endian and -maltivec=be is not specified.
7231 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7232 little endian and -maltivec=be is not specified.
7233
7234 2014-02-05 Richard Henderson <rth@redhat.com>
7235
7236 PR debug/52727
7237 * combine-stack-adj.c: Revert r206943.
7238 * sched-int.h (struct deps_desc): Add last_args_size.
7239 * sched-deps.c (init_deps): Initialize it.
7240 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7241 contain REG_ARGS_SIZE notes.
7242
7243 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7244
7245 * lto-cgraph.c (asm_nodes_output): Make global.
7246 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7247 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7248 (driver_handle_option): Handle OPT_fwpa.
7249
7250 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7251
7252 PR ipa/59947
7253 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7254 a comment typo and formatting issue. If odr_hash hasn't been
7255 created, return vNULL and set *completep to false.
7256
7257 PR middle-end/57499
7258 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7259 bb with no successors.
7260
7261 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7262
7263 PR target/59718
7264 * doc/invoke.texi (-march): Clarify documentation for ARM.
7265 (-mtune): Likewise.
7266 (-mcpu): Likewise.
7267
7268 2014-02-05 Richard Biener <rguenther@suse.de>
7269
7270 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7271 when not vectorizing because of too many alias checks.
7272 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7273 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7274
7275 2014-02-05 Nick Clifton <nickc@redhat.com>
7276
7277 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7278 accept extended registers in any mode when compiling for the MN10300.
7279
7280 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7281
7282 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7283 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7284 sanitization attributes.
7285 (can_inline_edge_p): Likewise.
7286 (sanitize_attrs_match_for_inline_p): New function.
7287
7288 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7289
7290 * ipa-prop.c (detect_type_change): Shor circuit testing of
7291 type changes on THIS pointer.
7292
7293 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7294
7295 PR target/59777
7296 * config/pa/pa.c (legitimize_tls_address): Return original address
7297 if not passed a SYMBOL_REF rtx.
7298 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7299 addresses.
7300 (pa_emit_move_sequence): Simplify TLS source operands.
7301 (pa_legitimate_constant_p): Reject all TLS constants.
7302 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7303 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7304
7305 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7306
7307 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7308 groups when we know they are controlled by LTO.
7309 * varasm.c (default_binds_local_p_1): If object is in other partition,
7310 it will be resolved locally.
7311
7312 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7313
7314 * config/host-linux.c (linux_gt_pch_use_address): Don't
7315 use SSIZE_MAX because it is not always defined.
7316
7317 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7318
7319 PR bootstrap/59913
7320 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7321 threshold for pseudo splitting.
7322 (update_ebb_live_info): Process call argument hard registers and
7323 hard registers from insn definition too.
7324 (max_small_class_regs_num): New constant.
7325 (inherit_in_ebb): Update live hard regs through EBBs. Update
7326 reloads_num only for small register classes. Don't split for
7327 outputs of jumps.
7328
7329 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7330
7331 PR ipa/60058
7332 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7333 is non-null.
7334
7335 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7336
7337 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7338 visibility is safe.
7339
7340 2014-02-04 Marek Polacek <polacek@redhat.com>
7341
7342 * gdbinit.in (pel): Define.
7343
7344 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7345
7346 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7347 behavior.
7348
7349 2014-02-04 Richard Biener <rguenther@suse.de>
7350
7351 PR lto/59723
7352 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7353 in function context local.
7354 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7355 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7356 similar to LTO_imported_decl_ref.
7357
7358 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7359
7360 PR tree-optimization/60002
7361 * cgraphclones.c (build_function_decl_skip_args): Clear
7362 DECL_LANG_SPECIFIC.
7363
7364 PR tree-optimization/60023
7365 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7366 false to gsi_replace.
7367 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7368 has been in some EH region and vec_stmt could throw, add
7369 vec_stmt into the same EH region.
7370 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7371 has no lhs, ignore it.
7372 * internal-fn.c (expand_MASK_LOAD): Likewise.
7373
7374 PR ipa/60026
7375 * tree-inline.c (copy_forbidden): Fail for
7376 __attribute__((optimize (0))) functions.
7377
7378 PR other/58712
7379 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7380 is set, copy one less argument.
7381 (expand_simd_clones): Don't subtract clone_info->inbranch
7382 from simd_clone_struct_alloc argument.
7383
7384 PR rtl-optimization/57915
7385 * recog.c (simplify_while_replacing): If all unary/binary/relational
7386 operation arguments are constant, attempt to simplify those.
7387
7388 PR middle-end/59261
7389 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7390 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7391
7392 2014-02-04 Richard Biener <rguenther@suse.de>
7393
7394 PR tree-optimization/60012
7395 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7396 TBAA disambiguation to all DDRs.
7397
7398 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7399
7400 PR target/59788
7401 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7402 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7403
7404 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7405
7406 PR ipa/59882
7407 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7408
7409 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7410
7411 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7412 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7413
7414 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7415
7416 PR ipa/59831
7417 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7418 to figure out targets of polymorphic calls with known decl.
7419 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7420 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7421 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7422 (get_polymorphic_call_info): ... here.
7423 (get_polymorphic_call_info_from_invariant): New function.
7424
7425 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7426
7427 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7428 lookup via vtable pointer; check for type consistency
7429 and turn inconsitent facts into UNREACHABLE.
7430 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7431 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7432 type inconsistent querries; return UNREACHABLE instead.
7433
7434 2014-02-03 Richard Henderson <rth@twiddle.net>
7435
7436 PR tree-opt/59924
7437 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7438 already processed this node.
7439 (normalize_one_pred_1): Pass along mark_set.
7440 (normalize_one_pred): Create and destroy a pointer_set_t.
7441 (normalize_one_pred_chain): Likewise.
7442
7443 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7444
7445 PR gcov-profile/58602
7446 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7447
7448 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7449
7450 PR ipa/59831
7451 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7452 -fno-devirtualize; try to devirtualize by the knowledge of
7453 virtual table pointer given by aggregate propagation.
7454 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7455 (ipa_print_node_jump_functions): Dump also offset that
7456 is relevant for polymorphic calls.
7457 (determine_known_aggregate_parts): Add arg_type parameter; use it
7458 instead of determining the type from pointer type.
7459 (ipa_compute_jump_functions_for_edge): Update call of
7460 determine_known_aggregate_parts.
7461 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7462 (gimple_get_virt_method_for_binfo): ... here; simplify using
7463 vtable_pointer_value_to_vtable.
7464 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7465 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7466 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7467 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7468 POINTER_PLUS_EXPR.
7469 (vtable_pointer_value_to_binfo): ... here.
7470 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7471
7472 2014-02-03 Teresa Johnson <tejohnson@google.com>
7473
7474 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7475 redef of outer loop index variable.
7476
7477 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7478
7479 PR c++/53017
7480 PR c++/59211
7481 * doc/extend.texi (Function Attributes): Typo.
7482
7483 2014-02-03 Cong Hou <congh@google.com>
7484
7485 PR tree-optimization/60000
7486 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7487 if the vectorized statement is a store. A store statement can only
7488 appear at the end of pattern statements.
7489
7490 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7491
7492 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7493 (ix86_option_override_internal): Default long double to 64-bit for
7494 32-bit Bionic and to 128-bit for 64-bit Bionic.
7495
7496 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7497 TARGET_LONG_DOUBLE_128 is true.
7498 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7499
7500 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7501 (mlong-double-64): Negate -mlong-double-128.
7502 (mlong-double-128): New option.
7503
7504 * config/i386/i386-c.c (ix86_target_macros): Define
7505 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7506
7507 * doc/invoke.texi: Document -mlong-double-128.
7508
7509 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7510
7511 PR rtl-optimization/60024
7512 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7513
7514 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7515
7516 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7517
7518 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7519
7520 PR rtl-optimization/57662
7521 * sel-sched.c (code_motion_path_driver): Do not mark already not
7522 existing blocks in the visiting bitmap.
7523
7524 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7525
7526 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7527 on the insn being emitted.
7528
7529 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7530 Will Deacon <will.deacon@arm.com>
7531
7532 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7533
7534 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7535
7536 * config/arm/arm-tables.opt: Regenerate.
7537
7538 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7539
7540 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7541 for vector types other than V16QImode.
7542 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7543 define_expand, and call altivec_expand_vec_perm_le when producing
7544 code with little endian element order.
7545 (*altivec_vperm_<mode>_internal): New insn having previous
7546 behavior of altivec_vperm_<mode>.
7547 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7548 altivec_expand_vec_perm_le when producing code with little endian
7549 element order.
7550 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7551 behavior of altivec_vperm_<mode>_uns.
7552
7553 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7554
7555 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7556 (altivec_vsumsws): Add handling for -maltivec=be with a little
7557 endian target.
7558 (altivec_vsumsws_direct): New.
7559 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7560 gen_altivec_vsumsws.
7561
7562 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7563
7564 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7565 vtable_pointer_value_to_binfo): New functions.
7566 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7567 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7568
7569 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7570
7571 * config/nios2/nios2.md (load_got_register): Initialize GOT
7572 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7573 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7574
7575 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7576
7577 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7578 preserverd by passthrough, do not propagate the type.
7579
7580 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7581
7582 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7583 (mips_atomic_assign_expand_fenv): New function.
7584 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7585
7586 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7587
7588 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7589 (__builtin_mips_set_fcsr): Likewise.
7590 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7591 MIPS_USI_FTYPE_VOID.
7592 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7593 (mips16_expand_set_fcsr): Likewise.
7594 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7595 (mips16_set_fcsr_stub): Likewise.
7596 (mips16_get_fcsr_one_only_stub): New class.
7597 (mips16_set_fcsr_one_only_stub): Likewise.
7598 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7599 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7600 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7601 (hard_float): New availability predicate.
7602 (mips_builtins): Add get_fcsr and set_fcsr.
7603 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7604 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7605 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7606 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7607 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7608 patterns.
7609
7610 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7611
7612 * config/mips/mips.c (mips_one_only_stub): New class.
7613 (mips_need_mips16_rdhwr_p): Replace with...
7614 (mips16_rdhwr_stub): ...this new variable.
7615 (mips16_stub_call_address): New function.
7616 (mips16_rdhwr_one_only_stub): New class.
7617 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7618 (mips_output_mips16_rdhwr): Delete.
7619 (mips_finish_stub): New function.
7620 (mips_code_end): Use it to handle rdhwr stubs.
7621
7622 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7623
7624 PR target/60017
7625 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7626 when calculating size of integer atomic types.
7627
7628 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7629
7630 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7631
7632 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7633
7634 PR tree-optimization/60003
7635 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7636 * profile.c (branch_prob): Use gimple_call_builtin_p
7637 to check for BUILT_IN_SETJMP_RECEIVER.
7638 * tree-inline.c (copy_bb): Call notice_special_calls.
7639
7640 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7641
7642 PR bootstrap/59985
7643 * lra-constraints.c (process_alt_operands): Update reload_sum only
7644 on the first pass.
7645
7646 2014-01-31 Richard Henderson <rth@redhat.com>
7647
7648 PR middle-end/60004
7649 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7650 until after else_eh is processed.
7651
7652 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7653
7654 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7655 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7656 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7657 in smmintrin.h, remove them.
7658 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7659 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7660 * config/i386/i386.md (ROUND_SAE): Fix value.
7661 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7662 (const48_operand): New.
7663 * config/i386/subst.md (round), (round_expand): Use
7664 const_4_or_8_to_11_operand.
7665 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7666
7667 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7668
7669 * config/i386/constraints.md (Yk): Swap meaning with k.
7670 * config/i386/i386.md (movhi_internal): Change Yk to k.
7671 (movqi_internal): Ditto.
7672 (*k<logic><mode>): Ditto.
7673 (*andhi_1): Ditto.
7674 (*andqi_1): Ditto.
7675 (kandn<mode>): Ditto.
7676 (*<code>hi_1): Ditto.
7677 (*<code>qi_1): Ditto.
7678 (kxnor<mode>): Ditto.
7679 (kortestzhi): Ditto.
7680 (kortestchi): Ditto.
7681 (kunpckhi): Ditto.
7682 (*one_cmplhi2_1): Ditto.
7683 (*one_cmplqi2_1): Ditto.
7684 * config/i386/sse.md (): Change k to Yk.
7685 (avx512f_load<mode>_mask): Ditto.
7686 (avx512f_blendm<mode>): Ditto.
7687 (avx512f_store<mode>_mask): Ditto.
7688 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7689 (avx512f_storedqu<mode>_mask): Ditto.
7690 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7691 Ditto.
7692 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7693 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7694 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7695 (avx512f_maskcmp<mode>3): Ditto.
7696 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7697 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7698 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7699 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7700 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7701 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7702 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7703 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7704 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7705 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7706 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7707 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7708 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7709 (vec_extract_lo_<mode>_maskm): Ditto.
7710 (vec_extract_hi_<mode>_maskm): Ditto.
7711 (avx512f_vternlog<mode>_mask): Ditto.
7712 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7713 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7714 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7715 (avx512f_<code>v8div16qi2_mask): Ditto.
7716 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7717 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7718 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7719 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7720 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7721 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7722 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7723 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7724 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7725 (avx512cd_maskb_vec_dupv8di): Ditto.
7726 (avx512cd_maskw_vec_dupv16si): Ditto.
7727 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7728 (avx512f_vpermi2var<mode>3_mask): Ditto.
7729 (avx512f_vpermi2var<mode>3_mask): Ditto.
7730 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7731 (*avx512f_gathersi<mode>): Ditto.
7732 (*avx512f_gathersi<mode>_2): Ditto.
7733 (*avx512f_gatherdi<mode>): Ditto.
7734 (*avx512f_gatherdi<mode>_2): Ditto.
7735 (*avx512f_scattersi<mode>): Ditto.
7736 (*avx512f_scatterdi<mode>): Ditto.
7737 (avx512f_compress<mode>_mask): Ditto.
7738 (avx512f_compressstore<mode>_mask): Ditto.
7739 (avx512f_expand<mode>_mask): Ditto.
7740 * config/i386/subst.md (mask): Change k to Yk.
7741 (mask_scalar_merge): Ditto.
7742 (sd): Ditto.
7743
7744 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7745
7746 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7747
7748 2014-01-31 Richard Biener <rguenther@suse.de>
7749
7750 PR middle-end/59990
7751 * builtins.c (fold_builtin_memory_op): Make sure to not
7752 use a floating-point mode or a boolean or enumeral type for
7753 the copy operation.
7754
7755 2014-01-30 DJ Delorie <dj@redhat.com>
7756
7757 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7758 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7759 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7760 whenever main() has an epilogue.
7761
7762 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7763
7764 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7765 unused variable "field".
7766 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7767 (vsx_mergeh_<mode>): Likewise.
7768 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7769 (altivec_vmrghh): Likewise.
7770 (altivec_vmrghw): Likewise.
7771 (altivec_vmrglb): Likewise.
7772 (altivec_vmrglh): Likewise.
7773 (altivec_vmrglw): Likewise.
7774 (altivec_vspltb): Add missing uses.
7775 (altivec_vsplth): Likewise.
7776 (altivec_vspltw): Likewise.
7777 (altivec_vspltsf): Likewise.
7778
7779 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7780
7781 PR target/59923
7782 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7783 frame related instructions.
7784
7785 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7786
7787 PR rtl-optimization/59959
7788 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7789 any reload of register whose subreg is invalid.
7790
7791 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7792
7793 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7794 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7795 Add missing return type - void.
7796
7797 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7798
7799 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7800 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7801 remove element index adjustment for endian (now handled in vsx.md
7802 and altivec.md).
7803 (altivec_expand_vec_perm_const): Use
7804 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7805 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7806 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7807 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7808 define_expand and a new define_insn *altivec_vspltb_internal;
7809 adjust for -maltivec=be on a little endian target.
7810 (altivec_vspltb_direct): New.
7811 (altivec_vsplth): Divide into a define_expand and a new
7812 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7813 little endian target.
7814 (altivec_vsplth_direct): New.
7815 (altivec_vspltw): Divide into a define_expand and a new
7816 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7817 little endian target.
7818 (altivec_vspltw_direct): New.
7819 (altivec_vspltsf): Divide into a define_expand and a new
7820 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7821 a little endian target.
7822
7823 2014-01-30 Richard Biener <rguenther@suse.de>
7824
7825 PR tree-optimization/59993
7826 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7827 can propagate form the earlier stmt and avoid the transform
7828 when the intermediate result is needed.
7829
7830 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7831
7832 * README.Portability: Fix typo.
7833
7834 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7835
7836 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7837 comparison_operator with ordered_comparison_operator.
7838
7839 2014-01-30 Nick Clifton <nickc@redhat.com>
7840
7841 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7842 Rename to mn10300_store_multiple_regs.
7843 * config/mn10300/mn10300.c: Likewise.
7844 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7845 store_multiple_regs.
7846 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7847 Call mn10300_store_multiple_regs.
7848
7849 2014-01-30 Nick Clifton <nickc@redhat.com>
7850 DJ Delorie <dj@redhat.com>
7851
7852 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7853 %fp 2 to keep registers after it properly word-aligned.
7854 (rl78_alloc_physical_registers_umul): Handle the case where both
7855 input operands are the same.
7856
7857 2014-01-30 Richard Biener <rguenther@suse.de>
7858
7859 PR tree-optimization/59903
7860 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7861 check properly.
7862
7863 2014-01-30 Jason Merrill <jason@redhat.com>
7864
7865 PR c++/59633
7866 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7867
7868 PR c++/59645
7869 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7870
7871 2014-01-30 Richard Biener <rguenther@suse.de>
7872
7873 PR tree-optimization/59951
7874 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7875
7876 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7877
7878 PR target/59784
7879 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7880 SFmode to DFmode case.
7881
7882 2014-01-29 DJ Delorie <dj@redhat.com>
7883
7884 * config/msp430/msp430.opt (-minrt): New.
7885 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7886 if -minrt given.
7887 (ENDFILE_SPEC): Likewise.
7888
7889 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7890
7891 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7892 (estimate_function_body_sizes): Use it.
7893
7894 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7895
7896 PR c++/58561
7897 * dwarf2out.c (is_cxx_auto): New.
7898 (is_base_type): Use it.
7899 (gen_type_die_with_usage): Likewise.
7900
7901 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7902
7903 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7904 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7905 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7906 -maltivec=be with LE targets.
7907 (vsx_mergeh_<mode>): Likewise.
7908 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7909 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7910 (altivec_vmrghb): Replace with define_expand and new
7911 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7912 (altivec_vmrghb_direct): New define_insn.
7913 (altivec_vmrghh): Replace with define_expand and new
7914 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7915 (altivec_vmrghh_direct): New define_insn.
7916 (altivec_vmrghw): Replace with define_expand and new
7917 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7918 (altivec_vmrghw_direct): New define_insn.
7919 (*altivec_vmrghsf): Adjust for endianness.
7920 (altivec_vmrglb): Replace with define_expand and new
7921 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7922 (altivec_vmrglb_direct): New define_insn.
7923 (altivec_vmrglh): Replace with define_expand and new
7924 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7925 (altivec_vmrglh_direct): New define_insn.
7926 (altivec_vmrglw): Replace with define_expand and new
7927 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7928 (altivec_vmrglw_direct): New define_insn.
7929 (*altivec_vmrglsf): Adjust for endianness.
7930 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7931 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7932 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7933 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7934 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7935 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7936 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7937 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7938
7939 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7940
7941 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7942 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7943 whitespace.
7944
7945 2014-01-29 Richard Biener <rguenther@suse.de>
7946
7947 PR tree-optimization/58742
7948 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7949 associate_pointerplus_align.
7950 (associate_pointerplus_diff): New function.
7951 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7952 and associate_pointerplus_diff.
7953
7954 2014-01-29 Richard Biener <rguenther@suse.de>
7955
7956 * lto-streamer.h (LTO_major_version): Bump to 3.
7957 (LTO_minor_version): Reset to 0.
7958
7959 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7960
7961 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7962 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7963 (arm_file_start): Generate correct asm header for armv7ve.
7964 * config/arm/bpabi.h: Add multilib support for armv7ve.
7965 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7966 and cortex-a15 to armv7ve.
7967 * config/arm/t-aprofile: Add multilib support for armv7ve.
7968 * doc/invoke.texi: Document -march=armv7ve.
7969
7970 2014-01-29 Richard Biener <rguenther@suse.de>
7971
7972 PR tree-optimization/58742
7973 * tree-ssa-forwprop.c (associate_plusminus): Return true
7974 if we changed sth, defer EH cleanup to ...
7975 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7976 (simplify_mult): New function.
7977
7978 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7979
7980 PR middle-end/59917
7981 PR tree-optimization/59920
7982 * tree.c (build_common_builtin_nodes): Remove
7983 __builtin_setjmp_dispatcher initialization.
7984 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7985 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7986 instead of gsi_after_labels + manually skipping debug stmts.
7987 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7988 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7989 * tree-inline.c (copy_edges_for_bb): Remove
7990 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7991 argument. Ignore computed_goto_p stmts. Don't call
7992 make_abnormal_goto_edges. If a call might need abnormal edges
7993 for non-local gotos, see if it already has an edge to
7994 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7995 with true argument, don't do anything then, otherwise add
7996 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7997 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7998 caller.
7999 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8000 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8001 (lower_stmt): Don't set data->calls_builtin_setjmp.
8002 (lower_builtin_setjmp): Adjust comment.
8003 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8004 * tree-cfg.c (found_computed_goto): Remove.
8005 (factor_computed_gotos): Remove.
8006 (make_goto_expr_edges): Return bool, true for computed gotos.
8007 Don't call make_abnormal_goto_edges.
8008 (build_gimple_cfg): Don't set found_computed_goto, don't call
8009 factor_computed_gotos.
8010 (computed_goto_p): No longer static.
8011 (make_blocks): Don't set found_computed_goto.
8012 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8013 (make_edges): If make_goto_expr_edges returns true, push bb
8014 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8015 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8016 vector. Record mapping between bbs and OpenMP regions if there
8017 are any, adjust make_gimple_omp_edges caller. Call
8018 handle_abnormal_edges.
8019 (make_abnormal_goto_edges): Remove.
8020 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8021 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8022 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8023 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8024 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8025 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8026 filling *region also set *region_idx to (*region)->entry->index.
8027
8028 PR other/58712
8029 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8030 For REGs set ORIGINAL_REGNO.
8031
8032 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8033
8034 * doc/md.texi: Mention that a target shouldn't implement
8035 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8036 than hi/lo pair.
8037
8038 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8039
8040 PR tree-optimization/59594
8041 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8042 a copy of the datarefs vector rather than the vector itself.
8043
8044 2014-01-28 Jason Merrill <jason@redhat.com>
8045
8046 PR c++/53756
8047 * dwarf2out.c (auto_die): New static.
8048 (gen_type_die_with_usage): Handle C++1y 'auto'.
8049 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8050 on definition.
8051
8052 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8053
8054 PR target/59672
8055 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8056 (SPEC_X32): Likewise.
8057 (SPEC_64): Likewise.
8058 * config/i386/i386.c (ix86_option_override_internal): Turn off
8059 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8060 for TARGET_16BIT.
8061 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8062 * config/i386/i386.h (TARGET_16BIT): New macro.
8063 (TARGET_16BIT_P): Likewise.
8064 * config/i386/i386.opt: Add m16.
8065 * doc/invoke.texi: Document -m16.
8066
8067 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8068
8069 PR preprocessor/59935
8070 * input.c (location_get_source_line): Bail out on when line number
8071 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8072
8073 2014-01-28 Richard Biener <rguenther@suse.de>
8074
8075 PR tree-optimization/58742
8076 * tree-ssa-forwprop.c (associate_plusminus): Handle
8077 pointer subtraction of the form (T)(P + A) - (T)P.
8078
8079 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8080
8081 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8082 at const_int_cost.
8083
8084 2014-01-28 Richard Biener <rguenther@suse.de>
8085
8086 Revert
8087 2014-01-28 Richard Biener <rguenther@suse.de>
8088
8089 PR rtl-optimization/45364
8090 PR rtl-optimization/59890
8091 * var-tracking.c (local_get_addr_clear_given_value): Handle
8092 already cleared slot.
8093 (val_reset): Handle not allocated local_get_addr_cache.
8094 (vt_find_locations): Use post-order on the inverted CFG.
8095
8096 2014-01-28 Richard Biener <rguenther@suse.de>
8097
8098 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8099
8100 2014-01-28 Richard Biener <rguenther@suse.de>
8101
8102 PR rtl-optimization/45364
8103 PR rtl-optimization/59890
8104 * var-tracking.c (local_get_addr_clear_given_value): Handle
8105 already cleared slot.
8106 (val_reset): Handle not allocated local_get_addr_cache.
8107 (vt_find_locations): Use post-order on the inverted CFG.
8108
8109 2014-01-28 Alan Modra <amodra@gmail.com>
8110
8111 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8112 * configure.ac <recursive call for build != host>: Define
8113 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8114 and LD_FOR_BUILD too.
8115 * configure: Regenerate.
8116
8117 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8118
8119 * config/i386/i386.c (get_builtin_code_for_version): Separate
8120 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8121 Broadwell from Haswell.
8122
8123 2014-01-27 Steve Ellcey <sellcey@mips.com>
8124
8125 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8126 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8127 * config/mips/mips.c (mips_option_override): Change setting
8128 of TARGET_DSP.
8129 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8130 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8131 Change from Mask to Var.
8132
8133 2014-01-27 Jeff Law <law@redhat.com>
8134
8135 * ipa-inline.c (inline_small_functions): Fix typo.
8136
8137 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8138
8139 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8140 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8141 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8142 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8143 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8144 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8145 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8146 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8147 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8148 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8149 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8150 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8151 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8152 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8153 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8154 (_mm512_storeu_epi64): Ditto.
8155 (_mm512_cmpge_epi32_mask): Ditto.
8156 (_mm512_cmpge_epu32_mask): Ditto.
8157 (_mm512_cmpge_epi64_mask): Ditto.
8158 (_mm512_cmpge_epu64_mask): Ditto.
8159 (_mm512_cmple_epi32_mask): Ditto.
8160 (_mm512_cmple_epu32_mask): Ditto.
8161 (_mm512_cmple_epi64_mask): Ditto.
8162 (_mm512_cmple_epu64_mask): Ditto.
8163 (_mm512_cmplt_epi32_mask): Ditto.
8164 (_mm512_cmplt_epu32_mask): Ditto.
8165 (_mm512_cmplt_epi64_mask): Ditto.
8166 (_mm512_cmplt_epu64_mask): Ditto.
8167 (_mm512_cmpneq_epi32_mask): Ditto.
8168 (_mm512_cmpneq_epu32_mask): Ditto.
8169 (_mm512_cmpneq_epi64_mask): Ditto.
8170 (_mm512_cmpneq_epu64_mask): Ditto.
8171 (_mm512_expand_pd): Ditto.
8172 (_mm512_expand_ps): Ditto.
8173 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8174 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8175 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8176 * config/i386/i386.c (ix86_builtins): Add
8177 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8178 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8179 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8180 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8181 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8182 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8183 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8184 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8185 IX86_BUILTIN_PMOVUSQW512_MEM.
8186 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8187 __builtin_ia32_pmovsqd512mem_mask,
8188 __builtin_ia32_pmovqd512mem_mask,
8189 __builtin_ia32_pmovusqw512mem_mask,
8190 __builtin_ia32_pmovsqw512mem_mask,
8191 __builtin_ia32_pmovqw512mem_mask,
8192 __builtin_ia32_pmovusdw512mem_mask,
8193 __builtin_ia32_pmovsdw512mem_mask,
8194 __builtin_ia32_pmovdw512mem_mask,
8195 __builtin_ia32_pmovqb512mem_mask,
8196 __builtin_ia32_pmovusqb512mem_mask,
8197 __builtin_ia32_pmovsqb512mem_mask,
8198 __builtin_ia32_pmovusdb512mem_mask,
8199 __builtin_ia32_pmovsdb512mem_mask,
8200 __builtin_ia32_pmovdb512mem_mask.
8201 (bdesc_args): Add __builtin_ia32_expanddf512,
8202 __builtin_ia32_expandsf512.
8203 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8204 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8205 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8206 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8207 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8208 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8209 (avx512f_<code>v8div16qi2_mask_store): This.
8210 (avx512f_expand<mode>): New.
8211
8212 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8213
8214 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8215 New.
8216 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8217 (_mm512_prefetch_i32scatter_pd): Ditto.
8218 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8219 (_mm512_prefetch_i64scatter_pd): Ditto.
8220 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8221 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8222 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8223 (_mm512_prefetch_i32scatter_ps): Ditto.
8224 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8225 (_mm512_prefetch_i64scatter_ps): Ditto.
8226 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8227 * config/i386/i386-builtin-types.def: Define
8228 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8229 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8230 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8231 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8232 IX86_BUILTIN_SCATTERPFQPD.
8233 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8234 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8235 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8236 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8237 __builtin_ia32_scatterpfqps.
8238 (ix86_expand_builtin): Expand new built-ins.
8239 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8240 fix memory access data type.
8241 (*avx512pf_gatherpf<mode>_mask): Ditto.
8242 (*avx512pf_gatherpf<mode>): Ditto.
8243 (avx512pf_scatterpf<mode>): Ditto.
8244 (*avx512pf_scatterpf<mode>_mask): Ditto.
8245 (*avx512pf_scatterpf<mode>): Ditto.
8246 (GATHER_SCATTER_SF_MEM_MODE): New.
8247 (avx512pf_gatherpf<mode>df): Ditto.
8248 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8249 (*avx512pf_scatterpf<mode>df): Ditto.
8250
8251 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8252
8253 PR bootstrap/59934
8254 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8255 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8256 reached.
8257
8258 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8259
8260 * common/config/arm/arm-common.c
8261 (arm_rewrite_mcpu): Handle multiple names.
8262 * config/arm/arm.h
8263 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8264
8265 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8266
8267 * gimple-builder.h (create_gimple_tmp): Delete.
8268
8269 2014-01-27 Christian Bruel <christian.bruel@st.com>
8270
8271 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8272 words comparisons.
8273
8274 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8275
8276 * config/pa/pa.md (call): Generate indirect long calls to non-local
8277 functions when outputing 32-bit code.
8278 (call_value): Likewise except for special call to buggy powf function.
8279
8280 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8281 portable runtime and PIC indirect calls.
8282 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8283 and PIC call sequences. Use ldo instead of blr to set return register
8284 in PIC call sequence.
8285
8286 2014-01-25 Walter Lee <walt@tilera.com>
8287
8288 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8289 avoid clobbering a live register.
8290
8291 2014-01-25 Walter Lee <walt@tilera.com>
8292
8293 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8294 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8295 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8296 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8297
8298 2014-01-25 Walter Lee <walt@tilera.com>
8299
8300 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8301 arguments on even registers.
8302 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8303 STACK_BOUNDARY.
8304 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8305 (BIGGEST_ALIGNMENT): Ditto.
8306 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8307
8308 2014-01-25 Walter Lee <walt@tilera.com>
8309
8310 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8311 insns before bundling.
8312 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8313
8314 2014-01-25 Walter Lee <walt@tilera.com>
8315
8316 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8317 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8318 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8319
8320 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8321
8322 * config/mips/constraints.md (kl): Delete.
8323 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8324 define expands, using...
8325 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8326 instructions for MIPS16.
8327 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8328 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8329
8330 2014-01-25 Walter Lee <walt@tilera.com>
8331
8332 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8333 (clzdi2): Ditto.
8334 (ffsdi2): Ditto.
8335
8336 2014-01-25 Walter Lee <walt@tilera.com>
8337
8338 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8339 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8340
8341 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8342
8343 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8344 Handle XOR.
8345
8346 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8347
8348 * print-rtl.c (in_call_function_usage): New var.
8349 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8350 EXPR_LIST mode as mode and not as reg note name.
8351
8352 PR middle-end/59561
8353 * cfgloopmanip.c (copy_loop_info): If
8354 loop->warned_aggressive_loop_optimizations, make sure
8355 the flag is set in target loop too.
8356
8357 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8358
8359 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8360 flag_cilkplus.
8361 * builtins.def: Likewise.
8362 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8363 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8364 * ira.c (ira_setup_eliminable_regset): Likewise.
8365 * omp-low.c (gate_expand_omp): Likewise.
8366 (execute_lower_omp): Likewise.
8367 (diagnose_sb_0): Likewise.
8368 (gate_diagnose_omp_blocks): Likewise.
8369 (simd_clone_clauses_extract): Likewise.
8370 (gate): Likewise.
8371
8372 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8373
8374 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8375 correction for little endian...
8376 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8377 here.
8378
8379 2014-01-24 Jeff Law <law@redhat.com>
8380
8381 PR tree-optimization/59919
8382 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8383 for non-returning calls.
8384
8385 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8386
8387 * common/config/aarch64/aarch64-common.c
8388 (aarch64_rewrite_mcpu): Handle multiple names.
8389 * config/aarch64/aarch64.h
8390 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8391
8392 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8393
8394 * input.c (add_file_to_cache_tab): Handle the case where fopen
8395 returns NULL.
8396
8397 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8398
8399 PR target/59929
8400 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8401 from push operand if code of push isn't PRE_DEC.
8402
8403 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8404
8405 PR target/59909
8406 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8407 -mquad-memory-atomic. Update -mquad-memory documentation to say
8408 it is only used for non-atomic loads/stores.
8409
8410 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8411 -mquad-memory or -mquad-memory-atomic switches.
8412
8413 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8414 -mquad-memory-atomic to ISA 2.07 support.
8415
8416 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8417 to separate support of normal quad word memory operations (ldq, stq)
8418 from the atomic quad word memory operations.
8419
8420 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8421 support to separate non-atomic quad word operations from atomic
8422 quad word operations. Disable non-atomic quad word operations in
8423 little endian mode so that we don't have to swap words after the
8424 load and before the store.
8425 (quad_load_store_p): Add comment about atomic quad word support.
8426 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8427 options printed with -mdebug=reg.
8428
8429 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8430 -mquad-memory-atomic as the test for whether we have quad word
8431 atomic instructions.
8432 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8433 or -mp8-vector are used, allow byte/half-word atomic operations.
8434
8435 * config/rs6000/sync.md (load_lockedti): Insure that the address
8436 is a proper indexed or indirect address for the lqarx instruction.
8437 On little endian systems, swap the hi/lo registers after the lqarx
8438 instruction.
8439 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8440 insure the address is valid for the lqarx instruction.
8441 (store_conditionalti): Insure that the address is a proper indexed
8442 or indirect address for the stqcrx. instruction. On little endian
8443 systems, swap the hi/lo registers before doing the stqcrx.
8444 instruction.
8445 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8446 insure the address is valid for the stqcrx. instruction.
8447
8448 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8449 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8450 type of quad memory support is available.
8451
8452 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8453
8454 PR regression/59915
8455 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8456 there is a danger of looping.
8457
8458 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8459
8460 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8461 force flag_ira_loop_pressure if set via command line.
8462
8463 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8464
8465 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8466 (ashr_simd): New builtin handling DI mode.
8467 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8468 (aarch64_sshr_simddi): New match pattern.
8469 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8470 (vshrd_n_s64): Likewise.
8471 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8472
8473 2014-01-23 Nick Clifton <nickc@redhat.com>
8474
8475 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8476 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8477 favour of mcu specific scripts.
8478 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8479 430x multilibs.
8480
8481 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8482 Alex Velenko <Alex.Velenko@arm.com>
8483
8484 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8485 (vaddv_s16): Likewise.
8486 (vaddv_s32): Likewise.
8487 (vaddv_u8): Likewise.
8488 (vaddv_u16): Likewise.
8489 (vaddv_u32): Likewise.
8490 (vaddvq_s8): Likewise.
8491 (vaddvq_s16): Likewise.
8492 (vaddvq_s32): Likewise.
8493 (vaddvq_s64): Likewise.
8494 (vaddvq_u8): Likewise.
8495 (vaddvq_u16): Likewise.
8496 (vaddvq_u32): Likewise.
8497 (vaddvq_u64): Likewise.
8498 (vaddv_f32): Likewise.
8499 (vaddvq_f32): Likewise.
8500 (vaddvq_f64): Likewise.
8501 (vmaxv_f32): Likewise.
8502 (vmaxv_s8): Likewise.
8503 (vmaxv_s16): Likewise.
8504 (vmaxv_s32): Likewise.
8505 (vmaxv_u8): Likewise.
8506 (vmaxv_u16): Likewise.
8507 (vmaxv_u32): Likewise.
8508 (vmaxvq_f32): Likewise.
8509 (vmaxvq_f64): Likewise.
8510 (vmaxvq_s8): Likewise.
8511 (vmaxvq_s16): Likewise.
8512 (vmaxvq_s32): Likewise.
8513 (vmaxvq_u8): Likewise.
8514 (vmaxvq_u16): Likewise.
8515 (vmaxvq_u32): Likewise.
8516 (vmaxnmv_f32): Likewise.
8517 (vmaxnmvq_f32): Likewise.
8518 (vmaxnmvq_f64): Likewise.
8519 (vminv_f32): Likewise.
8520 (vminv_s8): Likewise.
8521 (vminv_s16): Likewise.
8522 (vminv_s32): Likewise.
8523 (vminv_u8): Likewise.
8524 (vminv_u16): Likewise.
8525 (vminv_u32): Likewise.
8526 (vminvq_f32): Likewise.
8527 (vminvq_f64): Likewise.
8528 (vminvq_s8): Likewise.
8529 (vminvq_s16): Likewise.
8530 (vminvq_s32): Likewise.
8531 (vminvq_u8): Likewise.
8532 (vminvq_u16): Likewise.
8533 (vminvq_u32): Likewise.
8534 (vminnmv_f32): Likewise.
8535 (vminnmvq_f32): Likewise.
8536 (vminnmvq_f64): Likewise.
8537
8538 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8539
8540 * config/aarch64/aarch64-simd.md
8541 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8542 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8543 (*aarch64_mul3_elt<mode>): Likewise.
8544 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8545 (*aarch64_mul3_elt_to_64v2df): Likewise.
8546 (*aarch64_mla_elt<mode>): Likewise.
8547 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8548 (*aarch64_mls_elt<mode>): Likewise.
8549 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8550 (*aarch64_fma4_elt<mode>): Likewise.
8551 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8552 (*aarch64_fma4_elt_to_64v2df): Likewise.
8553 (*aarch64_fnma4_elt<mode>): Likewise.
8554 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8555 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8556 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8557 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8558 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8559 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8560 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8561 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8562 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8563
8564 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8565
8566 * config/aarch64/aarch64-simd.md
8567 (aarch64_be_checked_get_lane<mode>): New define_expand.
8568 * config/aarch64/aarch64-simd-builtins.def
8569 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8570 New builtin definition.
8571 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8572 Use new safe be builtin.
8573
8574 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8575
8576 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8577 New define_insn.
8578 (aarch64_be_st1<mode>): Likewise.
8579 (aarch_ld1<VALL:mode>): Define_expand modified.
8580 (aarch_st1<VALL:mode>): Likewise.
8581 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8582 (UNSPEC_ST1): Likewise.
8583
8584 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8585
8586 * config/microblaze/microblaze.md: Add trap insn and attribute
8587
8588 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8589
8590 PR preprocessor/58580
8591 * input.h (location_get_source_line): Take an additional line_size
8592 parameter.
8593 (void diagnostics_file_cache_fini): Declare new function.
8594 * input.c (struct fcache): New type.
8595 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8596 New static constants.
8597 (diagnostic_file_cache_init, total_lines_num)
8598 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8599 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8600 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8601 (get_next_line, read_next_line, goto_next_line, read_line_num):
8602 New static function definitions.
8603 (diagnostic_file_cache_fini): New function.
8604 (location_get_source_line): Take an additional output line_len
8605 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8606 read_line_num.
8607 * diagnostic.c (diagnostic_finish): Call
8608 diagnostic_file_cache_fini.
8609 (adjust_line): Take an additional input parameter for the length
8610 of the line, rather than calculating it with strlen.
8611 (diagnostic_show_locus): Adjust the use of
8612 location_get_source_line and adjust_line with respect to their new
8613 signature. While displaying a line now, do not stop at the first
8614 null byte. Rather, display the zero byte as a space and keep
8615 going until we reach the size of the line.
8616 * Makefile.in: Add vec.o to OBJS-libcommon
8617
8618 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8619 Ilya Tocar <ilya.tocar@intel.com>
8620
8621 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8622 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8623 (__builtin_ia32_kmov16): Ditto.
8624 * config/i386/i386.md (UNSPEC_KMOV): New.
8625 (kmovw): Ditto.
8626
8627 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8628
8629 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8630 (_mm512_storeu_si512): Ditto.
8631
8632 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8633
8634 PR target/52125
8635 * rtl.h (get_referenced_operands): Declare.
8636 * recog.c (get_referenced_operands): New function.
8637 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8638 operands have been referenced when recording LO_SUM references.
8639
8640 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8641
8642 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8643
8644 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8645
8646 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8647 Enable for generic and recent AMD targets.
8648
8649 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8650
8651 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8652 ARG_SIZE note when adjustment was eliminated.
8653
8654 2014-01-22 Jeff Law <law@redhat.com>
8655
8656 PR tree-optimization/59597
8657 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8658 in file. Accept new argument REGISTERING and use it to modify
8659 dump output appropriately.
8660 (register_jump_thread): Corresponding changes.
8661 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8662 thread paths involving joiner blocks. Add code to dump cancelled
8663 jump threading paths.
8664
8665 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8666
8667 PR rtl-optimization/59477
8668 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8669 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8670
8671 2014-01-22 Tom Tromey <tromey@redhat.com>
8672
8673 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8674 PARAMS.
8675 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8676
8677 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8678
8679 PR rtl-optimization/59896
8680 * lra-constraints.c (process_alt_operands): Check unused note for
8681 matched operands of insn with no output reloads.
8682
8683 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8684
8685 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8686 (mips_move_from_gpr_cost): Likewise.
8687
8688 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8689
8690 PR rtl-optimization/59858
8691 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8692 ira_class_hard_regs_num.
8693 (process_alt_operands): Increase reject for dying matched operand.
8694
8695 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8696
8697 PR target/59003
8698 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8699 smaller than size, perform several stores or loads and stores
8700 at dst + count - size to store or copy all of size bytes, rather
8701 than just last modesize bytes.
8702
8703 2014-01-20 DJ Delorie <dj@redhat.com>
8704
8705 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8706 that CLOBBERs are REGs before propogating their values.
8707
8708 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8709
8710 PR middle-end/59789
8711 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8712 (cgraph_inline_failed_type): New function.
8713 * cgraph.h (DEFCIFCODE): Add type.
8714 (cgraph_inline_failed_type_t): New enum.
8715 (cgraph_inline_failed_type): New prototype.
8716 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8717 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8718 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8719 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8720 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8721 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8722 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8723 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8724 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8725 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8726 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8727 OPTIMIZATION_MISMATCH.
8728 * tree-inline.c (expand_call_inline): Emit errors during
8729 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8730
8731 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8732
8733 PR target/59685
8734 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8735 mode attribute in insn output.
8736
8737 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8738
8739 * output.h (output_constant): Delete.
8740 * varasm.c (output_constant): Make private.
8741
8742 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8743
8744 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8745
8746 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8747
8748 PR middle-end/59860
8749 * tree.h (fold_builtin_strcat): New prototype.
8750 * builtins.c (fold_builtin_strcat): No longer static. Add len
8751 argument, if non-NULL, don't call c_strlen. Optimize
8752 directly into __builtin_memcpy instead of __builtin_strcpy.
8753 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8754 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8755
8756 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8757
8758 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8759 for SImode_address_operand operands, having only a REG argument.
8760
8761 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8762
8763 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8764 loader name using mbig-endian.
8765 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8766
8767 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8768
8769 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8770 (-mtune): Likewise.
8771 (-mcpu): Likewise.
8772
8773 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8774
8775 * config/aarch64/aarch64-protos.h
8776 (aarch64_cannot_change_mode_class_ptr): Declare.
8777 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8778 aarch64_cannot_change_mode_class_ptr): New.
8779 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8780 backend hook aarch64_cannot_change_mode_class.
8781
8782 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8783
8784 * common/config/aarch64/aarch64-common.c
8785 (aarch64_handle_option): Don't handle any option order logic here.
8786 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8787 selected_cpu, warn on architecture version mismatch.
8788 (aarch64_override_options): Fix parsing order for option strings.
8789
8790 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8791 Iain Sandoe <iain@codesourcery.com>
8792
8793 PR bootstrap/59496
8794 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8795 warning. Amend comment to reflect current functionality.
8796
8797 2014-01-20 Richard Biener <rguenther@suse.de>
8798
8799 PR middle-end/59860
8800 * builtins.c (fold_builtin_strcat): Remove case better handled
8801 by tree-ssa-strlen.c.
8802
8803 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8804
8805 * config/aarch64/aarch64.opt
8806 (mcpu, march, mtune): Make case-insensitive.
8807
8808 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8809
8810 PR target/59880
8811 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8812 if operands[1] is a REG or ZERO_EXTEND of a REG.
8813
8814 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8815
8816 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8817
8818 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8819
8820 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8821 long non-pic millicode calls.
8822
8823 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8824
8825 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8826
8827 2014-01-19 Kito Cheng <kito@0xlab.org>
8828
8829 * builtins.c (expand_movstr): Check movstr expand done or fail.
8830
8831 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8832 H.J. Lu <hongjiu.lu@intel.com>
8833
8834 PR target/59379
8835 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8836 to DImode for zero-extended addresses.
8837
8838 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8839
8840 PR rtl-optimization/57763
8841 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8842 on the new indirect jump_insn and increment LABEL_NUSES (label).
8843
8844 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8845
8846 PR bootstrap/59580
8847 PR bootstrap/59583
8848 * config.gcc (x86_archs): New variable.
8849 (x86_64_archs): Likewise.
8850 (x86_cpus): Likewise.
8851 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8852 --with-arch/--with-cpu= options.
8853 Support --with-arch=/--with-cpu={nehalem,westmere,
8854 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8855
8856 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8857
8858 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8859 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8860
8861 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8862
8863 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8864
8865 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8866
8867 PR target/58944
8868 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8869 clear cpp_get_options (parse_in)->warn_unused_macros for
8870 ix86_target_macros_internal with cpp_define.
8871
8872 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8873
8874 * jump.c (delete_related_insns): Keep (use (insn))s.
8875 * reorg.c (redundant_insn): Check for barriers too.
8876
8877 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8878
8879 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8880
8881 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8882
8883 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8884 call to $$dyncall when TARGET_LONG_CALLS is true.
8885
8886 2014-01-17 Jeff Law <law@redhat.com>
8887
8888 * ree.c (combine_set_extension): Temporarily disable test for
8889 changing number of hard registers.
8890
8891 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8892
8893 PR middle-end/58125
8894 * ipa-inline-analysis.c (inline_free_summary):
8895 Do not free summary of aliases.
8896
8897 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8898
8899 PR middle-end/59706
8900 * gimplify.c (gimplify_expr): Use create_tmp_var
8901 instead of create_tmp_var_raw. If cond doesn't have
8902 integral type, don't add the IFN_ANNOTATE builtin at all.
8903
8904 2014-01-17 Martin Jambor <mjambor@suse.cz>
8905
8906 PR ipa/59736
8907 * ipa-cp.c (prev_edge_clone): New variable.
8908 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8909 Also resize prev_edge_clone vector.
8910 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8911 (ipcp_edge_removal_hook): New function.
8912 (ipcp_driver): Register ipcp_edge_removal_hook.
8913
8914 2014-01-17 Andrew Pinski <apinski@cavium.com>
8915 Steve Ellcey <sellcey@mips.com>
8916
8917 PR target/59462
8918 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8919 of operator mode.
8920
8921 2014-01-17 Jeff Law <law@redhat.com>
8922
8923 PR middle-end/57904
8924 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8925 so that pass_ccp runs first.
8926
8927 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8928
8929 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8930 (ix86_adjust_cost): Use !TARGET_XXX.
8931 (do_reorder_for_imul): Likewise.
8932 (swap_top_of_ready_list): Likewise.
8933 (ix86_sched_reorder): Likewise.
8934
8935 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8936
8937 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8938 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8939 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8940 (intel_memset): New. Duplicate slm_memset.
8941 (intel_cost): New. Duplicate slm_cost.
8942 (m_INTEL): New macro.
8943 (processor_target_table): Add "intel".
8944 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8945 with PROCESSOR_INTEL for "intel".
8946 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8947 PROCESSOR_SILVERMONT.
8948 (ix86_issue_rate): Likewise.
8949 (ix86_adjust_cost): Likewise.
8950 (ia32_multipass_dfa_lookahead): Likewise.
8951 (swap_top_of_ready_list): Likewise.
8952 (ix86_sched_reorder): Likewise.
8953 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8954 instead of TARGET_OPT_AGU.
8955 * config/i386/i386.h (TARGET_INTEL): New.
8956 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8957 (processor_type): Add PROCESSOR_INTEL.
8958 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8959 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8960
8961 2014-01-17 Marek Polacek <polacek@redhat.com>
8962
8963 PR c/58346
8964 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8965 size is zero.
8966
8967 2014-01-17 Richard Biener <rguenther@suse.de>
8968
8969 PR tree-optimization/46590
8970 * opts.c (default_options_table): Add entries for
8971 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8972 all enabled at -O1 but not for -Og.
8973 * common.opt (fbranch-count-reg): Remove Init(1).
8974 (fmove-loop-invariants): Likewise.
8975 (ftree-pta): Likewise.
8976
8977 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8978
8979 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8980 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8981 decls to at least the GCC 4.8 used alignments.
8982
8983 PR fortran/59440
8984 * tree-nested.c (convert_nonlocal_reference_stmt,
8985 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8986 of GIMPLE_BIND stmts, adjust associated decls.
8987
8988 2014-01-17 Richard Biener <rguenther@suse.de>
8989
8990 PR tree-optimization/46590
8991 * vec.h (vec<>::bseach): New member function implementing
8992 binary search according to C89 bsearch.
8993 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8994 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8995 bitmap pointer again. Make accesses_in_loop a flat array.
8996 (mem_ref_obstack): New global.
8997 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8998 (mark_ref_stored): Likewise.
8999 (ref_indep_loop_p_2): Likewise.
9000 (set_ref_stored_in_loop): New helper function.
9001 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9002 (memref_free): Adjust.
9003 (record_mem_ref_loc): Simplify.
9004 (gather_mem_refs_stmt): Adjust.
9005 (sort_locs_in_loop_postorder_cmp): New function.
9006 (analyze_memory_references): Sort accesses_in_loop after
9007 loop postorder number.
9008 (find_ref_loc_in_loop_cmp): New function.
9009 (for_all_locs_in_loop): Find relevant cluster of locs in
9010 accesses_in_loop and iterate without recursion.
9011 (execute_sm): Avoid uninit warning.
9012 (struct ref_always_accessed): Simplify.
9013 (ref_always_accessed::operator ()): Likewise.
9014 (ref_always_accessed_p): Likewise.
9015 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9016 loop postorder numbers here.
9017 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9018 numbers.
9019
9020 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9021
9022 PR c++/57945
9023 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9024 on decls for which assemble_alias has been called.
9025
9026 2014-01-17 Nick Clifton <nickc@redhat.com>
9027
9028 * config/msp430/msp430.opt: (mcpu): New option.
9029 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9030 (msp430_option_override): Parse target_cpu. If the MCU name
9031 matches a generic string, clear target_mcu.
9032 (msp430_attr): Allow numeric interrupt values up to 63.
9033 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9034 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9035 option.
9036 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9037 Add mcpu matches.
9038 * config/msp430/msp430.md (popm): Use %J rather than %I.
9039 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9040 (addhi_cy_i): Use immediate_operand for operand 2.
9041 * doc/invoke.texi: Document -mcpu option.
9042
9043 2014-01-17 Richard Biener <rguenther@suse.de>
9044
9045 PR rtl-optimization/38518
9046 * df.h (df_analyze_loop): Declare.
9047 * df-core.c: Include cfgloop.h.
9048 (df_analyze_1): Split out main part of df_analyze.
9049 (df_analyze): Adjust.
9050 (loop_inverted_post_order_compute): New function.
9051 (loop_post_order_compute): Likewise.
9052 (df_analyze_loop): New function avoiding whole-function
9053 postorder computes.
9054 * loop-invariant.c (find_defs): Use df_analyze_loop.
9055 (find_invariants): Adjust.
9056 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9057
9058 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9059
9060 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9061 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9062
9063 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9064
9065 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9066 traversal when removing references.
9067
9068 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9069
9070 PR ipa/59775
9071 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9072
9073 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9074
9075 PR middle-end/56791
9076 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9077 pushing a reload for an autoinc when we had previously reloaded an
9078 inner part of the address.
9079
9080 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9081
9082 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9083 field.
9084 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9085 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9086 when not giving up or versioning for alias only because of
9087 loop->safelen.
9088 (vect_analyze_data_ref_dependences): Set to true.
9089 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9090 is a GIMPLE_PHI.
9091 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9092 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9093 to the condition.
9094
9095 PR middle-end/58344
9096 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9097
9098 PR target/59839
9099 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9100 operand 0 predicate for gathers, use a new pseudo as subtarget.
9101
9102 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9103
9104 PR middle-end/59609
9105 * lra-constraints.c (process_alt_operands): Add printing debug info.
9106 Check absence of input/output reloads for matched operands too.
9107
9108 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9109
9110 PR rtl-optimization/59835
9111 * ira.c (ira_init_register_move_cost): Increase cost for
9112 impossible modes.
9113
9114 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9115
9116 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9117
9118 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9119
9120 PR target/59780
9121 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9122 non-register objects. Use gen_(high/low)part more consistently.
9123 Fix assertions.
9124
9125 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9126
9127 PR target/59844
9128 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9129 endian support, remove tests for WORDS_BIG_ENDIAN.
9130 (p8_mfvsrd_3_<mode>): Likewise.
9131 (reload_gpr_from_vsx<mode>): Likewise.
9132 (reload_gpr_from_vsxsf): Likewise.
9133 (p8_mfvsrd_4_disf): Likewise.
9134
9135 2014-01-16 Richard Biener <rguenther@suse.de>
9136
9137 PR rtl-optimization/46590
9138 * lcm.c (compute_antinout_edge): Use postorder iteration.
9139 (compute_laterin): Use inverted postorder iteration.
9140
9141 2014-01-16 Nick Clifton <nickc@redhat.com>
9142
9143 PR middle-end/28865
9144 * varasm.c (output_constant): Return the number of bytes actually
9145 emitted.
9146 (output_constructor_array_range): Update the field size with the
9147 number of bytes emitted by output_constant.
9148 (output_constructor_regular_field): Likewise. Also do not
9149 complain if the total number of bytes emitted is now greater
9150 than the expected fieldpos.
9151 * output.h (output_constant): Update prototype and descriptive comment.
9152
9153 2014-01-16 Marek Polacek <polacek@redhat.com>
9154
9155 PR middle-end/59827
9156 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9157 it is error_mark_node.
9158
9159 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9160
9161 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9162 VALID_AVX256_REG_OR_OI_MODE.
9163
9164 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9165
9166 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9167 current procedure should be profiled.
9168
9169 2014-01-15 Andrew Pinski <apinski@cavium.com>
9170
9171 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9172 of moving from/to the STACK_REG register class.
9173
9174 2014-01-15 Richard Henderson <rth@redhat.com>
9175
9176 PR debug/54694
9177 * reginfo.c (global_regs_decl): Globalize.
9178 * rtl.h (global_regs_decl): Declare.
9179 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9180 reserved via global_regs.
9181
9182 2014-01-15 Teresa Johnson <tejohnson@google.com>
9183
9184 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9185
9186 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9187
9188 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9189 and vmulosh rather than call gen_vec_widen_smult_*.
9190 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9191 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9192 (vec_widen_smult_even_v16qi): Likewise.
9193 (vec_widen_umult_even_v8hi): Likewise.
9194 (vec_widen_smult_even_v8hi): Likewise.
9195 (vec_widen_umult_odd_v16qi): Likewise.
9196 (vec_widen_smult_odd_v16qi): Likewise.
9197 (vec_widen_umult_odd_v8hi): Likewise.
9198 (vec_widen_smult_odd_v8hi): Likewise.
9199 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9200 vmuloub rather than call gen_vec_widen_umult_*.
9201 (vec_widen_umult_lo_v16qi): Likewise.
9202 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9203 vmulosb rather than call gen_vec_widen_smult_*.
9204 (vec_widen_smult_lo_v16qi): Likewise.
9205 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9206 rather than call gen_vec_widen_umult_*.
9207 (vec_widen_umult_lo_v8hi): Likewise.
9208 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9209 rather than call gen_vec_widen_smult_*.
9210 (vec_widen_smult_lo_v8hi): Likewise.
9211
9212 2014-01-15 Jeff Law <law@redhat.com>
9213
9214 PR tree-optimization/59747
9215 * ree.c (find_and_remove_re): Properly handle case where a second
9216 eliminated extension requires widening a copy created for elimination
9217 of a prior extension.
9218 (combine_set_extension): Ensure that the number of hard regs needed
9219 for a destination register does not change when we widen it.
9220
9221 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9222
9223 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9224 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9225 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9226 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9227 (avr-*-rtems*): Likewise.
9228 (bfin*-rtems*): Likewise.
9229 (moxie-*-rtems*): Likewise.
9230 (h8300-*-rtems*): Likewise.
9231 (i[34567]86-*-rtems*): Likewise.
9232 (lm32-*-rtems*): Likewise.
9233 (m32r-*-rtems*): Likewise.
9234 (m68k-*-rtems*): Likewise.
9235 (microblaze*-*-rtems*): Likewise.
9236 (mips*-*-rtems*): Likewise.
9237 (powerpc-*-rtems*): Likewise.
9238 (sh-*-rtems*): Likewise.
9239 (sparc-*-rtems*): Likewise.
9240 (sparc64-*-rtems*): Likewise.
9241 (v850-*-rtems*): Likewise.
9242 (m32c-*-rtems*): Likewise.
9243
9244 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9245
9246 PR rtl-optimization/59511
9247 * ira.c (ira_init_register_move_cost): Use memory costs for some
9248 cases of register move cost calculations.
9249 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9250 instead of BB frequency.
9251 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9252 * lra-assigns.c (find_hard_regno_for): Ditto.
9253
9254 2014-01-15 Richard Biener <rguenther@suse.de>
9255
9256 PR tree-optimization/59822
9257 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9258 (vectorizable_load): Use it to hoist defs of uses of invariant
9259 loads out of the loop.
9260
9261 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9262 Kugan Vivekanandarajah <kuganv@linaro.org>
9263
9264 PR target/59695
9265 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9266 truncation.
9267
9268 2014-01-15 Richard Biener <rguenther@suse.de>
9269
9270 PR rtl-optimization/59802
9271 * lcm.c (compute_available): Use inverted postorder to seed
9272 the initial worklist.
9273
9274 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9275
9276 PR target/59803
9277 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9278 ADDR_REGS for invalid symrefs in non-PIC code.
9279
9280 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9281
9282 PR other/58712
9283 * builtins.c (determine_block_size): Initialize *probable_max_size
9284 even if len_rtx is CONST_INT.
9285
9286 2014-01-14 Andrew Pinski <apinski@cavium.com>
9287
9288 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9289 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9290 (cortexa53_tunings): Likewise.
9291 (aarch64_sched_issue_rate): New function.
9292 (TARGET_SCHED_ISSUE_RATE): Define.
9293
9294 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9295
9296 * ira-costs.c (find_costs_and_classes): Add missed
9297 ira_init_register_move_cost_if_necessary.
9298
9299 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9300
9301 PR target/59787
9302 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9303
9304 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9305
9306 PR target/59794
9307 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9308 to indicate if type is used for function return value. Warn ABI
9309 change if the vector mode isn't available for function return value.
9310 (ix86_function_arg_advance): Pass false to type_natural_mode.
9311 (ix86_function_arg): Likewise.
9312 (ix86_gimplify_va_arg): Likewise.
9313 (function_arg_32): Don't warn ABI change.
9314 (ix86_function_value): Pass true to type_natural_mode.
9315 (ix86_return_in_memory): Likewise.
9316 (ix86_struct_value_rtx): Removed.
9317 (TARGET_STRUCT_VALUE_RTX): Likewise.
9318
9319 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9320
9321 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9322 converting a conditional jump into a conditional return.
9323
9324 2014-01-14 Richard Biener <rguenther@suse.de>
9325
9326 PR tree-optimization/58921
9327 PR tree-optimization/59006
9328 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9329 hoisting invariant stmts.
9330 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9331 invariant loads on the preheader edge if possible.
9332
9333 2014-01-14 Joey Ye <joey.ye@arm.com>
9334
9335 * doc/plugin.texi (Building GCC plugins): Update to C++.
9336
9337 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9338
9339 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9340 (_mm_rcp28_round_ss): Ditto.
9341 (_mm_rsqrt28_round_sd): Ditto.
9342 (_mm_rsqrt28_round_ss): Ditto.
9343 (_mm_rcp28_sd): Ditto.
9344 (_mm_rcp28_ss): Ditto.
9345 (_mm_rsqrt28_sd): Ditto.
9346 (_mm_rsqrt28_ss): Ditto.
9347 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9348 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9349 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9350 (IX86_BUILTIN_RCP28SD): Ditto.
9351 (IX86_BUILTIN_RCP28SS): Ditto.
9352 (IX86_BUILTIN_RSQRT28SD): Ditto.
9353 (IX86_BUILTIN_RSQRT28SS): Ditto.
9354 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9355 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9356 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9357 (ix86_expand_special_args_builtin): Expand new FTYPE.
9358 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9359 (srcp14<mode>): Make insn unary.
9360 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9361 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9362 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9363 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9364 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9365 Fix rounding: make it SAE only.
9366 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9367 Ditto.
9368 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9369 Ditto.
9370 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9371 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9372 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9373 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9374 (round_saeonly_mask_scalar_operand4): Ditto.
9375 (round_saeonly_mask_scalar_op3): Ditto.
9376 (round_saeonly_mask_scalar_op4): Ditto.
9377
9378 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9379
9380 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9381 Implement -maltivec=be for vec_insert and vec_extract.
9382
9383 2014-01-10 DJ Delorie <dj@redhat.com>
9384
9385 * config/msp430/msp430.md (call_internal): Don't allow memory
9386 references with SP as the base register.
9387 (call_value_internal): Likewise.
9388 * config/msp430/constraints.md (Yc): New. For memory references
9389 that don't use SP as a base register.
9390
9391 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9392 "an integer without a # prefix"
9393 * config/msp430/msp430.md (epilogue_helper): Use it.
9394
9395 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9396
9397 PR target/59617
9398 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9399 AVX512F gather builtins.
9400 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9401 on gather decls with INTEGER_TYPE masktype.
9402 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9403 directly into the builtin rather than hoisting it before loop.
9404
9405 PR tree-optimization/59387
9406 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9407 (scev_const_prop): If folded_casts and type has undefined overflow,
9408 use force_gimple_operand instead of force_gimple_operand_gsi and
9409 for each added stmt if it is assign with
9410 arith_code_with_undefined_signed_overflow, call
9411 rewrite_to_defined_overflow.
9412 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9413 gimple-fold.h instead.
9414 (arith_code_with_undefined_signed_overflow,
9415 rewrite_to_defined_overflow): Moved to ...
9416 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9417 rewrite_to_defined_overflow): ... here. No longer static.
9418 Include gimplify-me.h.
9419 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9420 rewrite_to_defined_overflow): New prototypes.
9421
9422 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9423
9424 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9425
9426 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9427
9428 * builtins.c (get_object_alignment_2): Minor tweak.
9429 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9430
9431 2014-01-13 Christian Bruel <christian.bruel@st.com>
9432
9433 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9434 optimized non constant lengths.
9435
9436 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9437
9438 PR libgomp/59194
9439 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9440 load as __atomic_load_N if possible.
9441
9442 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9443
9444 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9445 target parameter.
9446 (rs6000_expand_builtin): Adjust call.
9447
9448 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9449
9450 PR target/58115
9451 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9452 * config/rs6000/rs6000.c: Include target-globals.h.
9453 (rs6000_set_current_function): Instead of doing target_reinit
9454 unconditionally, use save_target_globals_default_opts and
9455 restore_target_globals.
9456
9457 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9458 FPSCR.
9459 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9460 (rs6000_expand_builtin): Handle mffs and mtfsf.
9461 (rs6000_init_builtins): Define mffs and mtfsf.
9462 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9463 (rs6000_mffs): New pattern.
9464 (rs6000_mtfsf): New pattern.
9465
9466 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9467
9468 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9469 Start narrowing with START. Apply candidate-use pair
9470 and check overall cost in narrowing.
9471 (iv_ca_prune): Pass new argument.
9472
9473 2014-01-10 Jeff Law <law@redhat.com>
9474
9475 PR middle-end/59743
9476 * ree.c (combine_reaching_defs): Ensure the defining statement
9477 occurs before the extension when optimizing extensions with
9478 different source and destination hard registers.
9479
9480 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9481
9482 PR ipa/58585
9483 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9484 vtables into the type inheritance graph.
9485
9486 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9487
9488 PR rtl-optimization/59754
9489 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9490 modes in the REGNO != REGNO case.
9491
9492 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9493
9494 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9495
9496 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9497
9498 PR tree-optimization/59745
9499 * tree-predcom.c (tree_predictive_commoning_loop): Call
9500 free_affine_expand_cache if giving up because components is NULL.
9501
9502 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9503 GC in payload of target_globals struct instead of allocating them on
9504 the heap and the larger structs separately using GC.
9505 * target-globals.h (struct target_globals): Make regs, hard_regs,
9506 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9507 of GTY((skip)) and change type to void *.
9508 (reset_target_globals): Cast loads from those fields to corresponding
9509 types.
9510
9511 2014-01-10 Steve Ellcey <sellcey@mips.com>
9512
9513 PR plugins/59335
9514 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9515 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9516 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9517
9518 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9519
9520 PR target/59744
9521 * aarch64-modes.def (CC_Zmode): New flags mode.
9522 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9523 represents an equality.
9524 (aarch64_get_condition_code): Handle CC_Zmode.
9525 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9526
9527 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9528
9529 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9530 extraction in good case.
9531
9532 2014-01-10 Richard Biener <rguenther@suse.de>
9533
9534 PR tree-optimization/59374
9535 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9536 checking after SLP discovery. Mark stmts not participating
9537 in any SLP instance properly.
9538
9539 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9540
9541 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9542 when handling a SET rtx.
9543
9544 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9545
9546 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9547 (cortex-a57): Likewise.
9548 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9549
9550 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9551
9552 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9553 non-iwmmxt builtins.
9554
9555 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9556
9557 PR ipa/58252
9558 PR ipa/59226
9559 * ipa-devirt.c record_target_from_binfo): Take as argument
9560 stack of binfos and lookup matching one for virtual inheritance.
9561 (possible_polymorphic_call_targets_1): Update.
9562
9563 2014-01-10 Huacai Chen <chenhc@lemote.com>
9564
9565 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9566 kernel strings for Loongson-2E/2F/3A.
9567
9568 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9569
9570 PR middle-end/59670
9571 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9572 is_gimple_call before calling gimple_call_internal_p.
9573
9574 2014-01-09 Steve Ellcey <sellcey@mips.com>
9575
9576 * Makefile.in (TREE_FLOW_H): Remove.
9577 (TREE_SSA_H): Add file names from tree-flow.h.
9578 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9579 * tree.h: Remove tree-flow.h reference.
9580 * hash-table.h: Remove tree-flow.h reference.
9581 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9582 reference with tree-ssa-loop.h.
9583
9584 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9585
9586 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9587 default element-order behavior for -maltivec.
9588 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9589 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9590 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9591 when targeting big endian, at least for now.
9592 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9593
9594 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9595
9596 PR middle-end/47735
9597 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9598 var satisfies use_register_for_decl, just take into account type
9599 alignment, rather than decl alignment.
9600
9601 PR tree-optimization/59622
9602 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9603 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9604 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9605 Don't devirtualize for inplace at all. For targets.length () == 1,
9606 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9607
9608 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9609
9610 * config/i386/i386.md (cpu): Remove the unused btver1.
9611
9612 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9613
9614 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9615
9616 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9617
9618 PR target/58115
9619 * tree-core.h (struct target_globals): New forward declaration.
9620 (struct tree_target_option): Add globals field.
9621 * tree.h (TREE_TARGET_GLOBALS): Define.
9622 (prepare_target_option_nodes_for_pch): New prototype.
9623 * target-globals.h (struct target_globals): Define even if
9624 !SWITCHABLE_TARGET.
9625 * tree.c (prepare_target_option_node_for_pch,
9626 prepare_target_option_nodes_for_pch): New functions.
9627 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9628 * config/i386/i386.c: Include target-globals.h.
9629 (ix86_set_current_function): Instead of doing target_reinit
9630 unconditionally, use save_target_globals_default_opts and
9631 restore_target_globals.
9632
9633 2014-01-09 Richard Biener <rguenther@suse.de>
9634
9635 PR tree-optimization/59715
9636 * tree-cfg.h (split_critical_edges): Declare.
9637 * tree-cfg.c (split_critical_edges): Export.
9638 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9639
9640 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9641
9642 * cfgexpand.c (expand_stack_vars): Optionally disable
9643 asan stack protection.
9644 (expand_used_vars): Likewise.
9645 (partition_stack_vars): Likewise.
9646 * asan.c (asan_emit_stack_protection): Optionally disable
9647 after return stack usage.
9648 (instrument_derefs): Optionally disable memory access instrumentation.
9649 (instrument_builtin_call): Likewise.
9650 (instrument_strlen_call): Likewise.
9651 (asan_protect_global): Optionally disable global variables protection.
9652 * doc/invoke.texi: Added doc for new options.
9653 * params.def: Added new options.
9654 * params.h: Likewise.
9655
9656 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9657
9658 PR rtl-optimization/59724
9659 * ifcvt.c (cond_exec_process_if_block): Don't call
9660 flow_find_head_matching_sequence with 0 longest_match.
9661 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9662 non-active insns if !stop_after.
9663 (try_head_merge_bb): Revert 2014-01-07 changes.
9664
9665 2014-01-08 Jeff Law <law@redhat.com>
9666
9667 * ree.c (get_sub_rtx): New function, extracted from...
9668 (merge_def_and_ext): Here.
9669 (combine_reaching_defs): Use get_sub_rtx.
9670
9671 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9672
9673 * cgraph.h (varpool_variable_node): Do not choke on null node.
9674
9675 2014-01-08 Catherine Moore <clm@codesourcery.com>
9676
9677 * config/mips/mips.md (simple_return): Attempt to use JRC
9678 for microMIPS.
9679 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9680
9681 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9682
9683 PR rtl-optimization/59137
9684 * reorg.c (steal_delay_list_from_target): Call update_block for
9685 elided insns.
9686 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9687
9688 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9689
9690 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9691 two duplicate entries.
9692
9693 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9694
9695 Revert:
9696 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9697
9698 * config/mips/mips.c (mips_truncated_op_cost): New function.
9699 (mips_rtx_costs): Adjust test for BADDU.
9700 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9701
9702 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9703
9704 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9705 (*baddu_si): ...this new pattern.
9706
9707 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9708
9709 PR ipa/59722
9710 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9711
9712 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9713
9714 PR middle-end/57748
9715 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9716 inner_reference_p.
9717 (expand_expr, expand_normal): Adjust.
9718 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9719 inner_reference_p. Use inner_reference_p to expand inner references.
9720 (store_expr): Adjust.
9721 * cfgexpand.c (expand_call_stmt): Adjust.
9722
9723 2014-01-08 Rong Xu <xur@google.com>
9724
9725 * gcov-io.c (gcov_var): Move from gcov-io.h.
9726 (gcov_position): Ditto.
9727 (gcov_is_error): Ditto.
9728 (gcov_rewrite): Ditto.
9729 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9730 only part to libgcc/libgcov.h.
9731
9732 2014-01-08 Marek Polacek <polacek@redhat.com>
9733
9734 PR middle-end/59669
9735 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9736
9737 2014-01-08 Marek Polacek <polacek@redhat.com>
9738
9739 PR sanitizer/59667
9740 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9741
9742 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9743
9744 PR rtl-optimization/59649
9745 * stor-layout.c (get_mode_bounds): For BImode return
9746 0 and STORE_FLAG_VALUE.
9747
9748 2014-01-08 Richard Biener <rguenther@suse.de>
9749
9750 PR middle-end/59630
9751 * gimple.h (is_gimple_builtin_call): Remove.
9752 (gimple_builtin_call_types_compatible_p): New.
9753 (gimple_call_builtin_p): New overload.
9754 * gimple.c (is_gimple_builtin_call): Remove.
9755 (validate_call): Rename to ...
9756 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9757 check return types.
9758 (validate_type): New static function.
9759 (gimple_call_builtin_p): New overload and adjust.
9760 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9761 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9762 (gimple_fold_stmt_to_constant_1): Likewise.
9763 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9764
9765 2014-01-08 Richard Biener <rguenther@suse.de>
9766
9767 PR middle-end/59471
9768 * gimplify.c (gimplify_expr): Gimplify register-register type
9769 VIEW_CONVERT_EXPRs to separate stmts.
9770
9771 2014-01-07 Jeff Law <law@redhat.com>
9772
9773 PR middle-end/53623
9774 * ree.c (combine_set_extension): Handle case where source
9775 and destination registers in an extension insn are different.
9776 (combine_reaching_defs): Allow source and destination registers
9777 in extension to be different under limited circumstances.
9778 (add_removable_extension): Remove restriction that the
9779 source and destination registers in the extension are the same.
9780 (find_and_remove_re): Emit a copy from the extension's
9781 destination to its source after the defining insn if
9782 the source and destination registers are different.
9783
9784 PR middle-end/59285
9785 * ifcvt.c (merge_if_block): If we are merging a block with more than
9786 one successor with a block with no successors, remove any BARRIER
9787 after the second block.
9788
9789 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9790
9791 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9792
9793 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9794
9795 PR target/59652
9796 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9797 for 14-bit register offsets when INT14_OK_STRICT is false.
9798
9799 2014-01-07 Roland Stigge <stigge@antcom.de>
9800 Michael Meissner <meissner@linux.vnet.ibm.com>
9801
9802 PR 57386/target
9803 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9804 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9805
9806 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9807
9808 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9809 -mcpu.
9810
9811 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9812
9813 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9814 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9815 rtx is const0_rtx or not.
9816
9817 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9818
9819 PR target/58115
9820 * target-globals.c (save_target_globals): Remove this_fn_optab
9821 handling.
9822 * toplev.c: Include optabs.h.
9823 (target_reinit): Temporarily restore the global options if another
9824 set of options are in force.
9825
9826 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9827
9828 PR rtl-optimization/58668
9829 * cfgcleanup.c (flow_find_cross_jump): Don't count
9830 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9831 to determine what is counted.
9832 (flow_find_head_matching_sequence): Use active_insn_p to determine
9833 what is counted.
9834 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9835 counting change.
9836 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9837 determine what is counted.
9838
9839 PR tree-optimization/59643
9840 * tree-predcom.c (split_data_refs_to_components): If one dr is
9841 read and one write, determine_offset fails and the write isn't
9842 in the bad component, just put the read into the bad component.
9843
9844 2014-01-07 Mike Stump <mikestump@comcast.net>
9845 Jakub Jelinek <jakub@redhat.com>
9846
9847 PR pch/59436
9848 * tree-core.h (struct tree_optimization_option): Change optabs
9849 type from unsigned char * to void *.
9850 * optabs.c (init_tree_optimization_optabs): Adjust
9851 TREE_OPTIMIZATION_OPTABS initialization.
9852
9853 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9854
9855 PR target/59644
9856 * config/i386/i386.h (struct machine_function): Add
9857 no_drap_save_restore field.
9858 * config/i386/i386.c (ix86_save_reg): Use
9859 !cfun->machine->no_drap_save_restore instead of
9860 crtl->stack_realign_needed.
9861 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9862 this function clears frame_pointer_needed. Set
9863 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9864 and DRAP reg is needed.
9865
9866 2014-01-06 Marek Polacek <polacek@redhat.com>
9867
9868 PR c/57773
9869 * doc/implement-c.texi: Mention that other integer types are
9870 permitted as bit-field types in strictly conforming mode.
9871
9872 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9873
9874 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9875 is newly allocated.
9876
9877 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9878
9879 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9880
9881 2014-01-06 Martin Jambor <mjambor@suse.cz>
9882
9883 PR ipa/59008
9884 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9885 to int.
9886 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9887
9888 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9889
9890 PR debug/59350
9891 PR debug/59510
9892 * var-tracking.c (add_stores): Preserve the value of the source even if
9893 we don't record the store.
9894
9895 2014-01-06 Terry Guo <terry.guo@arm.com>
9896
9897 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9898
9899 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9900
9901 PR bootstrap/59541
9902 * config/darwin.c (darwin_function_section): Adjust return values to
9903 correspond to optimisation changes made in r206070.
9904
9905 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9906
9907 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9908 from prefetch_block tune setting.
9909 (nocona_cost): Correct size of prefetch block to 64.
9910
9911 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9912
9913 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9914 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9915 used to save the static chain register in the computation of the offset
9916 from which the FP registers need to be restored.
9917
9918 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9919
9920 PR tree-optimization/59519
9921 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9922 ICE if get_current_def (current_new_name) is already non-NULL, as long
9923 as it is a phi result of some other phi in *new_exit_bb that has
9924 the same argument.
9925
9926 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9927 or vmovdqu* for misaligned_operand.
9928 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9929 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9930 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9931 aligned_mem for AVX512F masked aligned load and store builtins and for
9932 non-temporal moves.
9933
9934 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9935
9936 PR tree-optimization/59651
9937 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9938 Address range for negative step should be added by TYPE_SIZE_UNIT.
9939
9940 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9941
9942 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9943 overlapping registers also for registers other than the stack pointer.
9944
9945 2014-01-03 Marek Polacek <polacek@redhat.com>
9946
9947 PR other/59661
9948 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9949 __builtin_FILE.
9950
9951 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9952
9953 PR target/59625
9954 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9955 asm goto as jump.
9956
9957 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9958 (push splitter): Use <P:MODE_SIZE> instead of
9959 GET_MODE_SIZE (<P:MODE>mode).
9960 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9961 (mov -1, reg peephole2): Likewise.
9962 * config/i386/sse.md (*mov<mode>_internal,
9963 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9964 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9965 *<code><mode>3, *andnot<mode>3<mask_name>,
9966 <mask_codefor><code><mode>3<mask_name>): Likewise.
9967 * config/i386/subst.md (mask_mode512bit_condition,
9968 sd_mask_mode512bit_condition): Likewise.
9969
9970 2014-01-02 Xinliang David Li <davidxl@google.com>
9971
9972 PR tree-optimization/59303
9973 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9974 (dump_predicates): Better output format.
9975 (pred_equal_p): New function.
9976 (is_neq_relop_p): Ditto.
9977 (is_neq_zero_form_p): Ditto.
9978 (pred_expr_equal_p): Ditto.
9979 (pred_neg_p): Ditto.
9980 (simplify_pred): Ditto.
9981 (simplify_preds_2): Ditto.
9982 (simplify_preds_3): Ditto.
9983 (simplify_preds_4): Ditto.
9984 (simplify_preds): Ditto.
9985 (push_pred): Ditto.
9986 (push_to_worklist): Ditto.
9987 (get_pred_info_from_cmp): Ditto.
9988 (is_degenerated_phi): Ditto.
9989 (normalize_one_pred_1): Ditto.
9990 (normalize_one_pred): Ditto.
9991 (normalize_one_pred_chain): Ditto.
9992 (normalize_preds): Ditto.
9993 (normalize_cond_1): Remove function.
9994 (normalize_cond): Ditto.
9995 (is_gcond_subset_of): Ditto.
9996 (is_subset_of_any): Ditto.
9997 (is_or_set_subset_of): Ditto.
9998 (is_and_set_subset_of): Ditto.
9999 (is_norm_cond_subset_of): Ditto.
10000 (pred_chain_length_cmp): Ditto.
10001 (convert_control_dep_chain_into_preds): Type change.
10002 (find_predicates): Ditto.
10003 (find_def_preds): Ditto.
10004 (destroy_predicates_vecs): Ditto.
10005 (find_matching_predicates_in_rest_chains): Ditto.
10006 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10007 (is_pred_expr_subset): Ditto.
10008 (is_pred_chain_subset_of): Ditto.
10009 (is_included_in): Ditto.
10010 (is_superset_of): Ditto.
10011
10012 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10013
10014 Update copyright years.
10015
10016 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10017
10018 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10019 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10020 config/arc/arc.md, config/arc/arc.opt,
10021 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10022 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10023 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10024 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10025 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10026 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10027 vtable-verify.c, vtable-verify.h: Use the standard form for the
10028 copyright notice.
10029
10030 2014-01-02 Tobias Burnus <burnus@net-b.de>
10031
10032 * gcc.c (process_command): Update copyright notice dates.
10033 * gcov-dump.c: Ditto.
10034 * gcov.c: Ditto.
10035 * doc/cpp.texi: Bump @copying's copyright year.
10036 * doc/cppinternals.texi: Ditto.
10037 * doc/gcc.texi: Ditto.
10038 * doc/gccint.texi: Ditto.
10039 * doc/gcov.texi: Ditto.
10040 * doc/install.texi: Ditto.
10041 * doc/invoke.texi: Ditto.
10042
10043 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10044
10045 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10046
10047 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10048
10049 * config/i386/sse.md (*mov<mode>_internal): Guard
10050 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10051
10052 PR rtl-optimization/59647
10053 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10054 new_rtx into UNSIGNED_FLOAT rtxes.
10055 \f
10056 Copyright (C) 2014 Free Software Foundation, Inc.
10057
10058 Copying and distribution of this file, with or without modification,
10059 are permitted in any medium without royalty provided the copyright
10060 notice and this notice are preserved.