1 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
3 * config/rs6000/rs6000.md (type): Add "logical". Delete
6 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
7 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
8 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
9 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
10 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
11 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
12 *mov<mode>_internal2, and 10 anonymous define_insns): Use
14 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
16 * config/rs6000/40x.md: (ppc403-integer, ppc403-compare): Adjust.
17 * config/rs6000/440.md: (ppc440-integer, ppc440-compare): Adjust.
18 * config/rs6000/476.md: (ppc476-simple-integer, ppc476-compare):
20 * config/rs6000/603.md: (ppc603-integer, ppc603-compare): Adjust.
21 * config/rs6000/6xx.md: (ppc604-integer, ppc604-compare): Adjust.
22 * config/rs6000/7450.md: (ppc7450-integer, ppc7450-compare):
24 * config/rs6000/7xx.md: (ppc750-integer, ppc750-compare): Adjust.
25 * config/rs6000/8540.md: (ppc8540_su): Adjust.
26 * config/rs6000/cell.md: (cell-integer, cell-fast-cmp,
27 cell-cmp-microcoded): Adjust.
28 * config/rs6000/e300c2c3.md: (ppce300c3_cmp, ppce300c3_iu):
30 * config/rs6000/e500mc.md: (e500mc_su): Adjust.
31 * config/rs6000/e500mc64.md: (e500mc64_su, e500mc64_su2): Adjust.
32 * config/rs6000/e5500.md: (e5500_sfx, e5500_sfx2): Adjust.
33 * config/rs6000/e6500.md: (e6500_sfx, e6500_sfx2): Adjust.
34 * config/rs6000/mpc.md: (mpccore-integer, mpccore-compare):
36 * config/rs6000/power4.md: (power4-integer, power4-cmp): Adjust.
37 * config/rs6000/power5.md: (power5-integer, power5-cmp): Adjust.
38 * config/rs6000/power6.md: (power6-integer, power6-fast-compare):
40 * config/rs6000/power7.md: (power7-integer, power7-cmp): Adjust.
41 * config/rs6000/power8.md: (power8-1cyc, power8-fast-compare):
42 Adjust. Adjust comment.
43 * config/rs6000/rs64.md: (rs64a-integer, rs64a-compare): Adjust.
44 * config/rs6000/titan.md: (titan_fxu_adder, titan_fxu_alu):
47 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
49 * config/rs6000/rs6000.md (type): Add "add".
50 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
51 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
52 define_insns): Use it.
53 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
55 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
56 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
57 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare):
59 * config/rs6000/601.md (ppc601-integer): Adjust.
60 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
61 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
62 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
63 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
64 * config/rs6000/8540.md (ppc8540_su): Adjust.
65 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
66 cell-cmp-microcoded): Adjust.
67 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
68 * config/rs6000/e500mc.md (e500mc_su): Adjust.
69 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
70 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
71 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
72 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
73 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
74 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
75 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
77 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
78 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
80 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
81 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
83 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
85 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
86 "delayed_compare", "var_delayed_compare".
87 (var_shift): New attribute.
89 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
90 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
91 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
92 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
93 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
94 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
95 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
96 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
97 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
98 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
99 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
100 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
101 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
102 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
103 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
104 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
105 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
106 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
107 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
108 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
109 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
110 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
111 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
112 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
113 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
115 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
116 * config/rs6000/440.md (ppc440-integer): Adjust.
117 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare):
119 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
120 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
121 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
122 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare):
124 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
125 * config/rs6000/8540.md (ppc8540_su): Adjust.
126 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
127 cell-cmp-microcoded): Adjust.
128 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
129 * config/rs6000/e500mc.md (e500mc_su): Adjust.
130 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
131 e500mc64_delayed): Adjust.
132 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
133 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
134 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
135 * config/rs6000/power4.md (power4-integer, power4-compare):
137 * config/rs6000/power5.md (power5-integer, power5-compare):
139 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
140 power6-delayed-compare, power6-var-delayed-compare): Adjust.
141 * config/rs6000/power7.md (power7-integer, power7-compare):
143 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
145 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
146 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
148 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
150 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add
152 (bits): New mode_attr.
153 (idiv_ldiv): Delete mode_attr.
154 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
155 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
156 rs6000_adjust_priority, is_nonpipeline_insn,
157 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
159 * config/rs6000/40x.md (ppc403-idiv): Adjust.
160 * config/rs6000/440.md (ppc440-idiv): Adjust.
161 * config/rs6000/476.md (ppc476-idiv): Adjust.
162 * config/rs6000/601.md (ppc601-idiv): Adjust.
163 * config/rs6000/603.md (ppc603-idiv): Adjust.
164 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
165 ppc620-ldiv): Adjust.
166 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
167 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
168 * config/rs6000/8540.md (ppc8540_divide): Adjust.
169 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
170 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
171 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
172 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
173 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
174 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
175 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
176 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
177 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
178 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
179 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
180 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
181 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
182 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
183 * config/rs6000/titan.md (titan_fxu_div): Adjust.
185 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
187 * config/rs6000/rs6000.md (type): Delete "insert_word",
188 "insert_dword". Add "insert".
189 (size): Update comment.
190 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
191 insn_must_be_first_in_group): Adjust.
192 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
193 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
194 *insvsi_internal6, insvdi_internal): Adjust.
196 * config/rs6000/40x.md (ppc403-integer): Adjust.
197 * config/rs6000/440.md (ppc440-integer): Adjust.
198 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
199 * config/rs6000/601.md (ppc601-integer): Adjust.
200 * config/rs6000/603.md (ppc603-integer): Adjust.
201 * config/rs6000/6xx.md (ppc604-integer): Adjust.
202 * config/rs6000/7450.md (ppc7450-integer): Adjust.
203 * config/rs6000/7xx.md (ppc750-integer): Adjust.
204 * config/rs6000/8540.md (ppc8540_su): Adjust.
205 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
206 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
207 * config/rs6000/e500mc.md (e500mc_su): Adjust.
208 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
209 * config/rs6000/e5500.md (e5500_sfx): Adjust.
210 * config/rs6000/e6500.md (e6500_sfx): Adjust.
211 * config/rs6000/mpc.md (mpccore-integer): Adjust.
212 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
213 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
214 * config/rs6000/power6.md (power6-insert, power6-insert-dword):
216 * config/rs6000/power7.md (power7-integer): Adjust.
217 * config/rs6000/power8.md (power8-1cyc): Adjust.
218 * config/rs6000/rs64.md (rs64a-integer): Adjust.
219 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
221 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
223 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
224 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
225 (size): New attribute.
226 (dot): New attribute.
227 (cell_micro): Adjust.
228 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
229 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
230 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
231 umuldi3_highpart): Adjust.
232 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
233 rs6000_adjust_priority, is_nonpipeline_insn,
234 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
236 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
237 ppc405-imul3): Adjust.
238 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
239 * config/rs6000/476.md (ppc476-imul): Adjust.
240 * config/rs6000/601.md (ppc601-imul): Adjust.
241 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
242 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
243 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
244 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
245 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
247 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
248 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
249 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
251 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
252 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
253 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
254 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
255 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
256 * config/rs6000/mpc.md (mpccore-imul): Adjust.
257 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
258 power4-lmul, power4-imul, power4-imul3): Adjust.
259 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
260 power5-lmul, power5-imul, power5-imul3): Adjust.
261 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
262 power6-lmul, power6-imul, power6-imul3): Adjust.
263 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
264 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
266 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
268 * config/rs6000/titan.md (titan_imul): Adjust.
270 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
272 * config/rs6000/rs6000.md (type): Add new value "halfmul".
273 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
274 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
275 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
276 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
277 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu):
279 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
280 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
281 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
282 * config/rs6000/titan.md: Delete nonsensical comment.
283 (titan_imul): Add type imul3.
284 (titan_mulhw): Remove type imul3; add type halfmul.
286 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
288 * config/rs6000/rs6000.md (type): Reorder, reformat.
290 2014-05-23 Martin Jambor <mjambor@suse.cz>
292 PR tree-optimization/53787
293 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
294 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
295 analysis_done, update all uses.
296 * ipa-prop.c: Include domwalk.h
297 (param_analysis_info): Removed.
298 (param_aa_status): New type.
299 (ipa_bb_info): Likewise.
300 (func_body_info): Likewise.
301 (ipa_get_bb_info): New function.
302 (aa_overwalked): Likewise.
303 (find_dominating_aa_status): Likewise.
304 (parm_bb_aa_status_for_bb): Likewise.
305 (parm_preserved_before_stmt_p): Changed to use new param AA info.
306 (load_from_unmodified_param): Accept func_body_info as a parameter
307 instead of parms_ainfo.
308 (parm_ref_data_preserved_p): Changed to use new param AA info.
309 (parm_ref_data_pass_through_p): Likewise.
310 (ipa_load_from_parm_agg_1): Likewise. Update callers.
311 (compute_complex_assign_jump_func): Changed to use new param AA info.
312 (compute_complex_ancestor_jump_func): Likewise.
313 (ipa_compute_jump_functions_for_edge): Likewise.
314 (ipa_compute_jump_functions): Removed.
315 (ipa_compute_jump_functions_for_bb): New function.
316 (ipa_analyze_indirect_call_uses): Likewise, moved variable
318 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
319 and info, moved variable declarations down.
320 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
322 (ipa_analyze_stmt_uses): Likewise.
323 (ipa_analyze_params_uses): Removed.
324 (ipa_analyze_params_uses_in_bb): New function.
325 (ipa_analyze_controlled_uses): Likewise.
326 (free_ipa_bb_info): Likewise.
327 (analysis_dom_walker): New class.
328 (ipa_analyze_node): Handle node-specific forbidden analysis,
329 initialize and free func_body_info, use dominator walker.
330 (ipcp_modif_dom_walker): New class.
331 (ipcp_transform_function): Create and free func_body_info, use
332 ipcp_modif_dom_walker, moved a lot of functionality there.
334 2014-05-23 Marek Polacek <polacek@redhat.com>
335 Jakub Jelinek <jakub@redhat.com>
337 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
338 * gcc.c (sanitize_spec_function): Likewise.
339 * convert.c (convert_to_integer): Include "ubsan.h". Add
340 floating-point to integer instrumentation.
341 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
342 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
344 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
345 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
346 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
347 * ubsan.c: Include "realmpfr.h" and "dfp.h".
348 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
349 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
350 float/double/long double.
351 (ubsan_instrument_float_cast): New function.
352 * ubsan.h (ubsan_instrument_float_cast): Declare.
354 2014-05-23 Jiong Wang <jiong.wang@arm.com>
356 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
358 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
359 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
360 Adjust for tailcalling through registers.
361 * config/aarch64/aarch64.h (enum reg_class): New caller save
363 (REG_CLASS_NAMES): Likewise.
364 (REG_CLASS_CONTENTS): Likewise.
365 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
366 Allow tailcalling without decls.
368 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
370 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
371 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
373 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
374 gsi, and variables v_* to v*.
376 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
378 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
380 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
382 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
383 * omp-low.c: Update accordingly.
385 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
386 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
387 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
388 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
389 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
390 GF_OMP_TARGET_KIND_UPDATE.
392 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
393 Explicitly enumerate the expected region types.
395 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
398 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
399 documentation; the old documentation didn't clearly state the
400 constraints on the contents of the pointed-to storage.
402 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
404 Fix bootstrap error on ia64
405 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
406 Return default value.
408 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
410 PR tree-optimization/54733
411 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
413 (find_bswap_or_nop_load): New.
414 (find_bswap_1): Renamed to ...
415 (find_bswap_or_nop_1): This. Also add support for memory source.
416 (find_bswap): Renamed to ...
417 (find_bswap_or_nop): This. Also add support for memory source and
418 detection of bitwise operations equivalent to load in host endianness.
419 (execute_optimize_bswap): Likewise. Also move its leading comment back
420 in place and split statement transformation into ...
421 (bswap_replace): This.
423 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
425 PR rtl-optimization/61215
426 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
427 simplify_gen_subreg until final substitution.
429 2014-05-23 Alan Modra <amodra@gmail.com>
432 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
433 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
434 Use "Y" constraint rather than "m".
436 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
438 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
440 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
441 New function declaration.
442 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
443 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
444 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
445 (aarch64_init_builtins) : Initialize builtins
446 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
447 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
448 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
449 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
450 and __builtins_aarch64_set_fpsr.
451 (aarch64_atomic_assign_expand_fenv): New function.
452 * config/aarch64/aarch64.md (set_fpcr): New pattern.
453 (get_fpcr) : Likewise.
454 (set_fpsr) : Likewise.
455 (get_fpsr) : Likewise.
456 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
457 and UNSPECV_SET_FPSR.
458 * doc/extend.texi (AARCH64 Built-in Functions) : Document
459 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
460 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
462 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
464 PR rtl-optimization/60969
465 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
466 constraints. Set up mem cost for NO_REGS case.
468 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
470 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
472 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
474 * config/darwin.c: Include "lto-section-names.h".
475 (LTO_SEGMENT_NAME): Don't define.
476 * config/i386/winnt.c: Include "lto-section-names.h".
477 * lto-streamer.c: Include "lto-section-names.h".
478 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
479 * lto-wrapper.c: Include "lto-section-names.h".
480 (LTO_SECTION_NAME_PREFIX): Don't define.
481 * lto-section-names.h: New file.
482 * cgraphunit.c: Include "lto-section-names.h".
484 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
486 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
488 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
491 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
493 2014-05-22 Nick Clifton <nickc@redhat.com>
495 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
498 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
500 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
501 -> (T)A transformation to integer types.
503 2014-05-22 Teresa Johnson <tejohnson@google.com>
505 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
506 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
507 (gcov_rewrite): Use gcov_nonruntime_assert.
509 (gcov_write_words): Ditto.
510 (gcov_write_length): Ditto.
511 (gcov_read_words): Use gcov_nonruntime_assert, and remove
512 gcc_assert from IN_LIBGCOV code.
513 (gcov_read_summary): Use gcov_error to flag profile corruption.
514 (gcov_sync): Use gcov_nonruntime_assert.
515 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
516 (gcov_histo_index): Use gcov_nonruntime_assert.
517 (static void gcov_histogram_merge): Ditto.
518 (compute_working_sets): Ditto.
519 * gcov-io.h (gcov_nonruntime_assert): Define.
520 (gcov_error): Define for !IN_LIBGCOV
522 2014-05-22 Richard Biener <rguenther@suse.de>
524 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
525 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
526 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
527 and deallocation site.
528 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
529 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
530 passing through the incoming points-to set.
531 (handle_lhs_call): Use flags argument instead of recomputing it.
532 (find_func_aliases_for_call): Call handle_lhs_call with proper
535 2014-05-22 Jakub Jelinek <jakub@redhat.com>
537 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
538 all padding bits in REAL_VALUE_TYPE are cleared.
540 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
542 Cleanup and improve multipass_dfa_lookahead_guard
543 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
544 (core2i7_first_cycle_multipass_begin,)
545 (core2i7_first_cycle_multipass_issue,)
546 (core2i7_first_cycle_multipass_backtrack): Update signature.
548 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
549 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
550 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
552 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
553 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
555 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
557 * doc/tm.texi: Regenerate.
559 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
560 * haifa-sched.c (ready_try): Make signed to allow negative values.
561 (rebug_ready_list_1): Update.
562 (choose_ready): Simplify.
563 (sched_extend_ready_list): Update.
565 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
567 Remove IA64 speculation tweaking flags
568 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
569 speculation tuning flags.
570 (msched-prefer-non-data-spec-insns,)
571 (msched-prefer-non-control-spec-insns): Obsolete options.
572 * haifa-sched.c (choose_ready): Remove handling of
573 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
574 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
575 and PREFER_NON_DATA_SPEC.
576 * sel-sched.c (process_spec_exprs): Remove handling of
577 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
579 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
581 Improve scheduling debug output
582 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
583 (advance_one_cycle): Update.
584 (schedule_insn, queue_to_ready): Add debug printouts.
585 (debug_ready_list_1): New static function.
586 (debug_ready_list): Update.
587 (max_issue): Add debug printouts.
588 (dump_insn_stream): New static function.
589 (schedule_block): Use it. Also better indent printouts.
591 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
593 Fix sched_insn debug counter
594 * haifa-sched.c (schedule_insn): Update.
595 (struct haifa_saved_data): Add nonscheduled_insns_begin.
596 (save_backtrack_point, restore_backtrack_point): Update.
597 (first_nonscheduled_insn): New static function.
598 (queue_to_ready, choose_ready): Use it.
599 (schedule_block): Init nonscheduled_insns_begin.
600 (sched_emit_insn): Update.
603 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
605 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
607 (aarch64_secondary_reload) : LikeWise.
608 (aarch64_class_max_nregs) : Remove CORE_REGS.
609 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
610 (REG_CLASS_NAMES) : Likewise.
611 (REG_CLASS_CONTENTS) : LikeWise.
612 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
614 2014-05-21 Guozhi Wei <carrot@google.com>
617 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
619 (vqdmulhq_n_s16): Likewise.
621 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
623 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
625 2014-05-21 Marek Polacek <polacek@redhat.com>
628 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
630 2014-05-21 Martin Jambor <mjambor@suse.cz>
632 * doc/invoke.texi (Optimize Options): Document parameters
633 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
634 ipa-cp-array-index-hint-bonus.
636 2014-05-21 Mark Wielaard <mjw@redhat.com>
639 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
640 version >= 3 or not strict DWARF.
641 * langhooks.h (struct lang_hooks_for_types): Add
642 enum_underlying_base_type.
643 * langhooks.c (lhd_enum_underlying_base_type): New function.
644 * gcc/langhooks.h (struct lang_hooks_for_types): Add
645 enum_underlying_base_type.
646 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
647 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
648 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
650 2014-05-21 Richard Biener <rguenther@suse.de>
652 * doc/invoke.texi (-flto-partition=): Document one and
655 2014-05-21 John Marino <gnugcc@marino.st>
657 * config.gcc (*-*-dragonfly*): New target.
658 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
659 * configure: Regenerate.
660 * config/dragonfly-stdint.h: New.
661 * config/dragonfly.h: New.
662 * config/dragonfly.opt: New.
663 * config/i386/dragonfly.h: New.
664 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
666 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
668 * tree.def (VOID_CST): New.
669 * tree-core.h (TI_VOID): New.
670 * tree.h (void_node): New.
671 * tree.c (tree_node_structure_for_code, tree_code_size)
672 (iterative_hash_expr): Handle VOID_CST.
673 (build_common_tree_nodes): Initialize void_node.
675 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
677 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
679 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
681 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
684 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
685 flag_reorder_blocks_and_partition.
686 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
688 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
691 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
693 (*addc_r_t): Add new insn_and_split.
695 2014-05-21 Jakub Jelinek <jakub@redhat.com>
698 * omp-low.c (handle_simd_reference): New function.
699 (lower_rec_input_clauses): Use it. Defer adding reference
700 initialization even for reduction without placeholder if in simd,
701 handle it properly later on.
703 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
705 PR tree-optimization/60899
706 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
707 assume all static symbols will have definition wile parsing and
708 check the do have definition later in compilation; check that
709 variable referring symbol will be output before concluding that
710 reference is safe; be conservative for referring local statics;
711 be more precise about when comdat is output in other partition.
713 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
716 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
717 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
718 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
720 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
722 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
723 set writeonly flag only for vars actually written to.
725 2014-05-20 Dehao Chen <dehao@google.com>
727 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
728 and callee count to get clone count.
729 * tree-inline.c (expand_call_inline): Use callee count instead of bb
732 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
734 PR rtl-optimization/61243
735 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
737 2014-05-20 Xinliang David Li <davidxl@google.com>
739 * cgraphunit.c (walk_polymorphic_call_targets): Add
740 dbgcnt and fopt-info support.
741 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
742 * ipa-devirt.c (ipa_devirt): Ditto.
743 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
745 * ipa.c (walk_polymorphic_call_targets): Ditto.
746 * gimple-fold.c (fold_gimple_assign): Ditto.
747 (gimple_fold_call): Ditto.
748 * dbgcnt.def: New counter.
750 2014-05-20 DJ Delorie <dj@redhat.com>
752 * config/msp430/msp430.md (split): Don't allow subregs when
753 splitting SImode adds.
754 (andneghi): Fix subtraction logic.
755 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
757 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
759 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
761 * except.c (switch_to_exception_section, resolve_unique_section,
762 get_named_text_section, default_function_rodata_section,
763 align_variable, get_block_for_decl, default_section_type_flags):
764 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
765 * symtab.c (symtab_add_to_same_comdat_group,
766 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
767 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
769 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
770 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
771 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
772 (c6x_function_in_section_p): Likewise.
773 * config/darwin.c (machopic_select_section): Likewise.
774 * config/arm/arm.c (arm_function_in_section_p): Likewise.
775 * config/mips/mips.c (mips_function_rodata_section): Likewise.
776 * config/mep/mep.c (mep_select_section): LIkewise.
777 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
779 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
781 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
782 EH region of calls to pure functions that can throw an exception.
783 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
784 (copy_reference_ops_from_call): Also copy the EH region of the call if
785 it can throw an exception.
787 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
789 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
790 nested VEC_SELECTs that are inverses of each other.
792 2014-05-20 Richard Biener <rguenther@suse.de>
794 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
796 (extract_and_process_scc_for_name): not here.
797 (cond_dom_walker::before_dom_children): Only process
798 stmts that end the BB in interesting ways.
799 (run_scc_vn): Mark param uses as visited.
801 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
803 * config/arm/arm.md (arith_shiftsi): Do not predicate for
806 2014-05-20 Nick Clifton <nickc@redhat.com>
808 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
809 (msp430_gimplify_va_arg_expr): New function.
810 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
812 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
813 operand 0 in order to prevent confusion about the number of
816 2014-05-20 Richard Biener <rguenther@suse.de>
818 PR tree-optimization/61221
819 * tree-ssa-pre.c (el_to_update): Remove.
820 (eliminate_dom_walker::before_dom_children): Handle released
821 VDEFs by value-numbering them to the associated VUSE. Update
822 stmt immediately for substituted call address.
823 (eliminate): Remove delayed stmt updating code.
824 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
825 possibly late re-numbered vuses.
826 (vn_reference_lookup_2): Adjust.
827 (vn_reference_lookup_pieces): Likewise.
828 (vn_reference_lookup): Likewise.
830 2014-05-20 Richard Biener <rguenther@suse.de>
832 * config.gcc: Remove need_64bit_hwint.
833 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
834 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
836 * config.in: Regenerate.
837 * configure: Likewise.
839 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
841 * doc/extend.texi: Create Label Attributes section,
842 move all label attributes into it and reference it.
844 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
846 * arm.c (thumb1_reorg): When scanning backwards skip anything
847 that's not a proper insn.
849 2014-05-19 Richard Biener <rguenther@suse.de>
851 PR tree-optimization/61221
852 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
853 Do nothing for unreachable blocks.
854 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
855 Improve unreachability detection.
857 2014-05-19 Richard Biener <rguenther@suse.de>
859 PR tree-optimization/61209
860 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
862 2014-05-19 Nick Clifton <nickc@redhat.com>
864 * except.c (init_eh): Fix computation of builtin setjmp buffer
865 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
867 2014-05-19 Richard Biener <rguenther@suse.de>
869 PR tree-optimization/61184
870 * tree-vrp.c (is_negative_overflow_infinity): Use
871 TREE_OVERFLOW_P and do that check first.
872 (is_positive_overflow_infinity): Likewise.
873 (is_overflow_infinity): Likewise.
874 (vrp_operand_equal_p): Properly treat operands with
875 differing overflow as not equal.
877 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
879 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
880 shift simplification where it was intended.
882 2014-05-19 Christian Bruel <christian.bruel@st.com>
885 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
887 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
890 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
893 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
895 * reg-notes.def (CROSSING_JUMP): Likewise.
896 * rtl.h (rtx_def): Update comment for jump flag.
897 (CROSSING_JUMP_P): Define.
898 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
899 of a REG_CROSSING_JUMP note.
900 * cfghooks.c (tidy_fallthru_edges): Likewise.
901 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
902 * emit-rtl.c (try_split): Likewise.
903 * haifa-sched.c (sched_create_recovery_edges): Likewise.
904 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
905 * jump.c (redirect_jump_2): Likewise.
906 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
907 (relax_delay_slots): Likewise.
908 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
910 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
911 * config/sh/sh.md (jump_compact): Likewise.
912 * bb-reorder.c (rotate_loop): Likewise.
913 (pass_duplicate_computed_gotos::execute): Likewise.
914 (add_reg_crossing_jump_notes): Rename to...
915 (update_crossing_jump_flags): ...this.
916 (pass_partition_blocks::execute): Update accordingly.
918 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
920 * tree.h: Remove extraneous template <>.
922 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
924 * ipa.c (symtab_remove_unreachable_nodes): Remove
925 symbol from comdat group if its body was eliminated.
926 (comdat_can_be_unshared_p_1): Static symbols can always
928 * symtab.c (symtab_remove_from_same_comdat_group): Break out
930 (symtab_unregister_node): ... this one.
931 (verify_symtab_base): More strict checking of comdats.
932 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
934 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
936 * tree-pass.h (make_pass_ipa_comdats): New pass.
937 * timevar.def (TV_IPA_COMDATS): New timevar.
938 * passes.def (pass_ipa_comdats): Add.
939 * Makefile.in (OBJS): Add ipa-comdats.o
940 * ipa-comdats.c: New file.
942 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
944 * ipa.c (update_visibility_by_resolution_info): New function.
945 (function_and_variable_visibility): Use it.
947 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
949 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
951 (FOR_EACH_DEFINED_SYMBOL): New macro.
952 (varpool_first_static_initializer, varpool_next_static_initializer,
953 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
954 (symtab_in_same_comdat_p): Correctly deal with inline functions.
956 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
958 * ggc-page.c (ggc_handle_finalizers): Add comment.
960 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
962 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
963 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
964 (ggc_internal_cleared_alloc): Likewise.
965 * ggc-page.c (finalizer): New class.
966 (vec_finalizer): Likewise.
967 (globals::finalizers): New member.
968 (globals::vec_finalizers): Likewise.
969 (ggc_internal_alloc): Record the finalizer if any for the block being
971 (ggc_handle_finalizers): New function.
972 (ggc_collect): Call ggc_handle_finalizers.
973 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
975 (ggc_internal_cleared_alloc): Likewise.
976 (finalize): New function.
977 (need_finalization_p): Likewise.
978 (ggc_alloc): Install the type's destructor as the finalizer if it
980 (ggc_cleared_alloc): Likewise.
981 (ggc_vec_alloc): Likewise.
982 (ggc_cleared_vec_alloc): Likewise.
984 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
986 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
988 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
990 * alias.c (record_alias_subset): Adjust.
991 * bitmap.c (bitmap_element_allocate): Likewise.
992 (bitmap_gc_alloc_stat): Likewise.
993 * cfg.c (init_flow): Likewise.
994 (alloc_block): Likewise.
995 (unchecked_make_edge): Likewise.
996 * cfgloop.c (alloc_loop): Likewise.
997 (flow_loops_find): Likewise.
998 (rescan_loop_exit): Likewise.
999 * cfgrtl.c (init_rtl_bb_info): Likewise.
1000 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1001 (cgraph_allocate_node): Likewise.
1002 (cgraph_create_edge_1): Likewise.
1003 (cgraph_allocate_init_indirect_info): Likewise.
1004 * cgraphclones.c (cgraph_clone_edge): Likewise.
1005 * cgraphunit.c (add_asm_node): Likewise.
1006 (init_lowered_empty_function): Likewise.
1007 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1008 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1009 (alpha_use_linkage): Likewise.
1010 * config/arc/arc.c (arc_init_machine_status): Likewise.
1011 * config/arm/arm.c (arm_init_machine_status): Likewise.
1012 * config/avr/avr.c (avr_init_machine_status): Likewise.
1013 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1014 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1015 * config/cris/cris.c (cris_init_machine_status): Likewise.
1016 * config/darwin.c (machopic_indirection_name): Likewise.
1017 (darwin_build_constant_cfstring): Likewise.
1018 (darwin_enter_string_into_cfstring_table): Likewise.
1019 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1020 * config/frv/frv.c (frv_init_machine_status): Likewise.
1021 * config/i386/i386.c (get_dllimport_decl): Likewise.
1022 (ix86_init_machine_status): Likewise.
1023 (assign_386_stack_local): Likewise.
1024 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1025 (i386_pe_maybe_record_exported_symbol): Likewise.
1026 (i386_pe_record_stub): Likewise.
1027 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1028 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1029 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1030 (m32c_note_pragma_address): Likewise.
1031 * config/mep/mep.c (mep_init_machine_status): Likewise.
1032 (mep_note_pragma_flag): Likewise.
1033 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1034 (mips16_local_alias): Likewise.
1035 (mips_init_machine_status): Likewise.
1036 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1037 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1038 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1039 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1040 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1041 * config/pa/pa.c (pa_init_machine_status): Likewise.
1042 (pa_get_deferred_plabel): Likewise.
1043 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1044 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1045 (rs6000_init_machine_status): Likewise.
1046 (output_toc): Likewise.
1047 * config/s390/s390.c (s390_init_machine_status): Likewise.
1048 * config/score/score.c (score_output_external): Likewise.
1049 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1050 * config/spu/spu.c (spu_init_machine_status): Likewise.
1051 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1052 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1053 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1054 * coverage.c (coverage_end_function): Likewise.
1055 * dbxout.c (dbxout_init): Likewise.
1056 * doc/gty.texi: Don't mention variable_size attribute.
1057 * dwarf2cfi.c (new_cfi): Adjust.
1058 (new_cfi_row): Likewise.
1059 (copy_cfi_row): Likewise.
1060 (create_cie_data): Likewise.
1061 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1062 (new_loc_descr): Likewise.
1063 (find_AT_string_in_table): Likewise.
1064 (add_addr_table_entry): Likewise.
1065 (new_die): Likewise.
1066 (add_var_loc_to_decl): Likewise.
1067 (clone_die): Likewise.
1068 (clone_as_declaration): Likewise.
1069 (break_out_comdat_types): Likewise.
1070 (new_loc_list): Likewise.
1071 (add_loc_descr_to_each): Likewise.
1072 (add_location_or_const_value_attribute): Likewise.
1073 (add_linkage_name): Likewise.
1074 (lookup_filename): Likewise.
1075 (dwarf2out_var_location): Likewise.
1076 (new_line_info_table): Likewise.
1077 (dwarf2out_init): Likewise.
1078 (mem_loc_descriptor): Likewise.
1079 (loc_descriptor): Likewise.
1080 (add_const_value_attribute): Likewise.
1081 (tree_add_const_value_attribute): Likewise.
1082 (comp_dir_string): Likewise.
1083 (dwarf2out_vms_debug_main_pointer): Likewise.
1084 (string_cst_pool_decl): Likewise.
1085 * emit-rtl.c (set_mem_attrs): Likewise.
1086 (get_reg_attrs): Likewise.
1087 (start_sequence): Likewise.
1088 (init_emit): Likewise.
1089 (init_emit_regs): Likewise.
1090 * except.c (init_eh_for_function): Likewise.
1091 (gen_eh_region): Likewise.
1092 (gen_eh_region_catch): Likewise.
1093 (gen_eh_landing_pad): Likewise.
1094 (add_call_site): Likewise.
1095 * function.c (add_frame_space): Likewise.
1096 (insert_temp_slot_address): Likewise.
1097 (assign_stack_temp_for_type): Likewise.
1098 (get_hard_reg_initial_val): Likewise.
1099 (allocate_struct_function): Likewise.
1100 (prepare_function_start): Likewise.
1101 (types_used_by_var_decl_insert): Likewise.
1102 * gengtype.c (variable_size_p): Remove function.
1103 (enum alloc_quantity): Remove enum.
1104 (write_typed_alloc_def): Remove function.
1105 (write_typed_struct_alloc_def): Likewise.
1106 (write_typed_typedef_alloc_def): Likewise.
1107 (write_typed_alloc_defns): Likewise.
1109 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1110 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1111 * ggc.h (ggc_alloc): new function.
1112 (ggc_cleared_alloc): Likewise.
1113 (ggc_vec_alloc): Template on type of vector element, and remove
1114 element size argument.
1115 (ggc_cleared_vec_alloc): Likewise.
1116 * gimple.c (gimple_build_omp_for): Adjust.
1117 (gimple_copy): Likewise.
1118 * ipa-cp.c (get_replacement_map): Likewise.
1119 (find_aggregate_values_for_callers_subset): Likewise.
1120 (known_aggs_to_agg_replacement_list): Likewise.
1121 * ipa-devirt.c (get_odr_type): Likewise.
1122 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1123 (read_agg_replacement_chain): Likewise.
1124 * loop-iv.c (get_simple_loop_desc): Likewise.
1125 * lto-cgraph.c (input_node_opt_summary): Likewise.
1126 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1127 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1128 (input_eh_region): Likewise.
1129 (input_eh_lp): Likewise.
1130 (input_cfg): Likewise.
1131 * optabs.c (set_optab_libfunc): Likewise.
1132 (init_tree_optimization_optabs): Likewise.
1133 (set_conv_libfunc): Likewise.
1134 * passes.c (do_per_function_toporder): Likewise.
1135 * rtl.h: Don't use variable_size gty attribute.
1136 * sese.c (if_region_set_false_region): Adjust.
1137 * stringpool.c (gt_pch_save_stringpool): Likewise.
1138 * target-globals.c (save_target_globals): Likewise.
1139 * toplev.c (general_init): Likewise.
1140 * trans-mem.c (record_tm_replacement): Likewise.
1141 (split_bb_make_tm_edge): Likewise.
1142 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1143 * tree-data-ref.h (lambda_vector_new): Likewise.
1144 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1145 * tree-iterator.c (tsi_link_before): Likewise.
1146 (tsi_link_after): Likewise.
1147 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1148 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1149 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1150 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1151 * tree-ssa.c (init_tree_ssa): Adjust.
1152 * tree-ssanames.c (set_range_info): Likewise.
1153 (get_ptr_info): Likewise.
1154 (duplicate_ssa_name_ptr_info): Likewise.
1155 (duplicate_ssa_name_range_info): Likewise.
1156 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1157 (unpack_ts_fixed_cst_value_fields): Likewise.
1158 * tree.c (build_fixed): Likewise.
1159 (build_real): Likewise.
1160 (build_string): Likewise.
1161 (decl_priority_info): Likewise.
1162 (decl_debug_expr_insert): Likewise.
1163 (decl_value_expr_insert): Likewise.
1164 (decl_debug_args_insert): Likewise.
1165 (type_hash_add): Likewise.
1166 (build_omp_clause): Likewise.
1167 * ubsan.c (decl_for_type_insert): Likewise.
1168 * varasm.c (get_unnamed_section): Likewise.
1169 (get_noswitch_section): Likewise.
1170 (get_section): Likewise.
1171 (get_block_for_section): Likewise.
1172 (create_block_symbol): Likewise.
1173 (build_constant_desc): Likewise.
1174 (create_constant_pool): Likewise.
1175 (force_const_mem): Likewise.
1176 (record_tm_clone_pair): Likewise.
1177 * varpool.c (varpool_create_empty_node): Likewise.
1179 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1181 * dwarf2out.c (tree_add_const_value_attribute): Call
1182 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1183 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1184 instead of ggc_internal_<x>alloc_stat.
1185 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1186 (ggc_realloc): Likewise.
1187 * ggc-none.c (ggc_internal_alloc): Likewise.
1188 (ggc_internal_cleared_alloc): Likewise.
1189 * ggc-page.c: Likewise.
1190 * ggc.h (ggc_internal_alloc_stat): Likewise.
1191 (ggc_internal_alloc): Remove macro.
1192 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1193 (ggc_internal_cleared_alloc): Remove macro.
1194 (GGC_RESIZEVEC): Adjust.
1195 (ggc_resizevar): Remove macro.
1196 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1197 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1198 (ggc_internal_vec_cleared_alloc): Remove macro.
1199 (ggc_alloc_atomic_stat): Drop _stat suffix.
1200 (ggc_alloc_atomic): Remove macro.
1201 (ggc_alloc_cleared_atomic): Remove macro.
1202 (ggc_alloc_string_stat): Drop _stat suffix.
1203 (ggc_alloc_string): Remove macro.
1204 (ggc_alloc_rtx_def_stat): Adjust.
1205 (ggc_alloc_tree_node_stat): Likewise.
1206 (ggc_alloc_cleared_tree_node_stat): Likewise.
1207 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1208 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1209 * gimple.c (gimple_build_omp_for): Likewise.
1210 (gimple_copy): Likewise.
1211 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1212 * toplev.c (realloc_for_line_map): Adjust.
1213 * tree-data-ref.h (lambda_vector_new): Likewise.
1214 * tree-phinodes.c (allocate_phi_node): Likewise.
1215 * tree.c (grow_tree_vec_stat): Likewise.
1216 * vec.h (va_gc::reserve): Adjust.
1218 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1220 * config/microblaze/microblaze.c (break_handler): New Declaration.
1221 (microblaze_break_function_p,microblaze_is_break_handler): New.
1222 (compute_frame_size): Use microblaze_break_function_p.
1223 Add the test of break_handler.
1224 (microblaze_function_prologue) : Add the test of variable
1225 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1226 (microblaze_function_epilogue) : Add the test of break_handler.
1227 (microblaze_globalize_label) : Add the test of break_handler.
1228 Check the name by BREAK_HANDLER_NAME.
1230 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1232 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1233 microblaze_is_break_handler test.
1234 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1235 Use SYMBOL_REF_DECL.
1237 * config/microblaze/microblaze-protos.h
1238 (microblaze_break_function_p,microblaze_is_break_handler):
1241 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1242 new MicroBlaze break_handler functions.
1244 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1246 * doc/extend.texi (Size of an asm): Move node text according
1247 to its @menu entry position.
1249 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1251 PR tree-optimization/61140
1252 PR tree-optimization/61150
1253 PR tree-optimization/61197
1254 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1256 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1258 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1260 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1262 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1263 __SIZEOF_INT128__ is defined.
1265 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1267 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1268 (rs6000_delegitimize_address): Use it.
1270 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1272 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1273 inplace argument. Store the new address in the original MEM when true.
1274 * emit-rtl.c (change_address_1): Likewise.
1275 (adjust_address_1, adjust_automodify_address_1, offset_address):
1277 * rtl.h (plus_constant): Add an inplace argument.
1278 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1279 when true. Avoid generating (plus X (const_int 0)).
1280 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1281 in-place. Pass true to plus_constant.
1282 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1284 2014-05-16 Dehao Chen <dehao@google.com>
1286 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1288 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1291 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1293 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1295 2014-05-16 Dehao Chen <dehao@google.com>
1297 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1298 optimize_function_for_size_p.
1299 * regs.h (REG_FREQ_FROM_BB): Likewise.
1301 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1304 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1305 negt_reg_operand cases.
1306 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1308 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1310 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1312 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1315 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1318 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1320 * tree-cfg.c (dump_function_to_file): Dump the return type of
1321 functions, in a line to itself before the function body, mimicking
1322 the layout of a C function.
1324 2014-05-16 Dehao Chen <dehao@google.com>
1326 * cfghooks.c (make_forwarder_block): Use direct computation to
1327 get fall-through edge's count and frequency.
1329 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1331 * config/arc/arc.c (arc_init): Fix typo in error message.
1332 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1333 (split_stack_prologue_scratch_regno): Likewise.
1334 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1335 word from error message.
1337 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1339 * ira-costs.c: Fix typo in comment.
1341 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1343 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1345 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1347 * varpool.c (dump_varpool_node): Dump write-only flag.
1348 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1350 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1351 write-only variables.
1352 * ipa.c (process_references): New function.
1353 (set_readonly_bit): New function.
1354 (set_writeonly_bit): New function.
1355 (clear_addressable_bit): New function.
1356 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1357 fix handling of aliases.
1358 * cgraph.h (struct varpool_node): Add writeonly flag.
1360 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1362 PR rtl-optimization/60969
1363 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1364 Calculate costs for this case.
1366 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1368 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1369 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1371 2014-05-16 Richard Biener <rguenther@suse.de>
1373 PR tree-optimization/61194
1374 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1375 bool patterns ending in a COND_EXPR.
1377 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1379 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1381 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1383 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1384 where we were unable to cost an RTX.
1386 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1388 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1391 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1392 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1394 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1396 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1397 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1399 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1400 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1402 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1403 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1405 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1408 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1409 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1411 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1414 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1415 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1417 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1418 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1420 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1421 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1423 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1426 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1427 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1429 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1430 ZERO_EXTEND and SIGN_EXTEND better.
1432 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1433 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1435 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1438 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1439 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1441 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1442 costs when costing loads and stores to memory.
1444 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1445 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1447 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1450 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1452 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1454 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1455 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1457 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1459 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1461 (aarch64_rtx_mult_cost): New.
1462 (aarch64_rtx_costs): Use it, refactor as appropriate.
1464 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1465 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1467 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1468 emit instructions, return number of instructions which would
1470 (aarch64_add_constant): Update call to aarch64_build_constant.
1471 (aarch64_output_mi_thunk): Likewise.
1472 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1475 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1477 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1478 (TARGET_RTX_COSTS): Call it.
1480 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1482 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1483 (cortexa57_vector_cost): Likewise.
1484 (cortexa57_tunings): Use them.
1486 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1488 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1489 (cpu_addrcost_table): Use it.
1490 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1491 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1494 2014-05-16 Richard Biener <rguenther@suse.de>
1496 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1497 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1498 (visit_phi): Ignore edges marked as not executable.
1499 (class cond_dom_walker): New.
1500 (cond_dom_walker::before_dom_children): Value-number
1501 control statements and mark successor edges as not
1502 executable if possible.
1503 (run_scc_vn): First walk all control statements in
1504 dominator order, marking edges as not executable.
1505 * tree-inline.c (copy_edges_for_bb): Be not confused
1506 about random edge flags.
1508 2014-05-16 Richard Biener <rguenther@suse.de>
1510 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1512 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1515 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1516 (__TM_simple_begin): Use it.
1517 (__TM_begin): Likewise.
1519 2014-05-15 Martin Jambor <mjambor@suse.cz>
1522 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1523 type_preserved flag when the indirect edge is polymorphic.
1525 2014-05-15 Martin Jambor <mjambor@suse.cz>
1527 PR tree-optimization/61090
1528 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1529 build_ref_for_model.
1531 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1533 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1534 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1536 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1538 PR tree-optimization/61158
1539 * fold-const.c (fold_binary_loc): If X is zero-extended and
1540 shiftc >= prec, make sure zerobits is all ones instead of
1541 invoking undefined behavior.
1543 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1545 * regcprop.h: New file.
1546 * regcprop.c (skip_debug_insn_p): New decl.
1547 (replace_oldest_value_reg): Check skip_debug_insn_p.
1548 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1549 * shrink-wrap.c: Include regcprop.h.
1550 (prepare_shrink_wrap): Call
1551 copyprop_hardreg_forward_bb_without_debug_insn.
1553 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1555 * shrink-wrap.h: Update comment.
1556 * shrink-wrap.c: Update comment.
1557 (next_block_for_reg): Rename to live_edge_for_reg.
1558 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1559 (move_insn_for_shrink_wrap): Split live_edge.
1560 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1562 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1564 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1566 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1567 * config/sparc/sparc.md (fptype_ut699): New attribute.
1568 (in_branch_delay): Return false if -mfix-ut699 is specified and
1569 fptype_ut699 is set to single.
1570 (truncdfsf2): Add fptype_ut699 attribute.
1571 (fix_truncdfsi2): Likewise.
1572 (floatsisf2): Change fptype attribute.
1573 (fix_truncsfsi2): Likewise.
1574 (negtf2_notv9): Delete.
1575 (negtf2_v9): Likewise.
1576 (negtf2_hq): New instruction.
1577 (negtf2): New instruction and splitter.
1578 (negdf2_notv9): Rewrite.
1579 (abstf2_notv9): Delete.
1580 (abstf2_hq_v9): Likewise.
1581 (abstf2_v9): Likewise.
1582 (abstf2_hq): New instruction.
1583 (abstf2): New instruction and splitter.
1584 (absdf2_notv9): Rewrite.
1586 2014-05-14 Cary Coutant <ccoutant@google.com>
1589 * opts.c (common_handle_option): Don't special-case "-g".
1590 (set_debug_level): Default to at least level 2 with "-g".
1592 2014-05-14 DJ Delorie <dj@redhat.com>
1594 * config/msp430/msp430.c (msp430_builtin): Add
1595 MSP430_BUILTIN_DELAY_CYCLES.
1596 (msp430_init_builtins): Register void __delay_cycles(long long).
1597 (msp430_builtin_decl): Add it.
1598 (cg_magic_constant): New.
1599 (msp430_expand_delay_cycles): New.
1600 (msp430_expand_builtin): Call it.
1601 (msp430_print_operand_raw): Change integer printing from "int" to
1603 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1604 (delay_cycles_start): New.
1605 (delay_cycles_end): New.
1606 (delay_cycles_32): New.
1607 (delay_cycles_32x): New.
1608 (delay_cycles_16): New.
1609 (delay_cycles_16x): New.
1610 (delay_cycles_2): New.
1611 (delay_cycles_1): New.
1612 * doc/extend.texi: Document __delay_cycles().
1614 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1616 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1617 length attribute computation.
1619 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1622 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1624 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1627 * config/sparc/sparc.md: Fix types of low and high in DI constant
1628 splitter. Use gen_int_mode in some other splitters.
1630 2014-05-14 Martin Jambor <mjambor@suse.cz>
1633 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1635 2014-05-14 James Norris <jnorris@codesourcery.com>
1637 * omp-low.c (expand_parallel_call): Remove shadow variable.
1638 (expand_omp_taskreg): Likewise.
1640 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1642 * common/config/i386/i386-common.c
1643 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1644 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1645 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1646 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1647 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1648 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1649 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1650 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1651 xsavecintrin.h, xsavesintrin.h.
1652 (x86_64-*-*): Ditto.
1653 * config/i386/clflushoptintrin.h: New.
1654 * config/i386/xsavecintrin.h: Ditto.
1655 * config/i386/xsavesintrin.h: Ditto.
1656 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1657 (bit_XSAVES): Ditto.
1658 (bit_XSAVES): Ditto.
1659 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1660 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1662 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1663 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1664 OPTION_MASK_ISA_XSAVES.
1665 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1667 (PTA_CLFLUSHOPT) Define.
1668 (PTA_XSAVEC): Ditto.
1669 (PTA_XSAVES): Ditto.
1670 (ix86_option_override_internal): Handle new options.
1671 (ix86_valid_target_attribute_inner_p): Ditto.
1672 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1673 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1674 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1675 (bdesc_special_args): Add __builtin_ia32_xsaves,
1676 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1677 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1678 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1679 (ix86_expand_builtin): Handle new builtins.
1680 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1681 (TARGET_CLFLUSHOPT_P): Ditto.
1682 (TARGET_XSAVEC): Ditto.
1683 (TARGET_XSAVEC_P): Ditto.
1684 (TARGET_XSAVES): Ditto.
1685 (TARGET_XSAVES_P): Ditto.
1686 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1687 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1688 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1690 (ANY_XRSTOR64): Ditto.
1692 (xrstor): Change into <xrstor>.
1693 (xrstor_rex64): Change into <xrstor>_rex64.
1694 (xrstor64): Change into <xrstor>64
1696 * config/i386/i386.opt (mclflushopt): New.
1699 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1701 * doc/invoke.texi: Document new options.
1703 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1705 PR rtl-optimization/60866
1706 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1707 Default it to -1. Pass it down to init_simplejump_data.
1708 (init_simplejump_data): New parameter old_seqno. Pass it down
1709 to get_seqno_for_a_jump.
1710 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1711 initializing new jump seqno as a last resort. Add comment.
1712 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1713 jump and pass it down to sel_init_new_insn.
1714 (sel_redirect_edge_and_branch_force): Likewise.
1716 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1718 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1719 shifted values to avoid build warning.
1721 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1723 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1724 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1725 (cfg_layout_merge_blocks): Likewise.
1726 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1728 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1730 PR rtl-optimization/60901
1731 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1732 bb predecessor belongs to the same scheduling region. Adjust comment.
1734 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1736 * doc/sourcebuild.texi: (dfp_hw): Document.
1737 (p8vector_hw): Likewise.
1738 (powerpc_eabi_ok): Likewise.
1739 (powerpc_elfv2): Likewise.
1740 (powerpc_htm_ok): Likewise.
1741 (ppc_recip_hw): Likewise.
1744 2014-05-13 Cary Coutant <ccoutant@google.com>
1746 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1748 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1750 * gengtype-parse.c (require3): Eliminate in favor of...
1752 (require_template_declaration): Update to support optional single *
1755 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1756 (create_user_defined_type): Handle a single level of explicit
1757 pointerness within template arguments.
1758 (struct write_types_data): Add field "kind".
1759 (filter_type_name): Handle "*" character.
1760 (write_user_func_for_structure_ptr): Require a write_types_data
1761 rather than just a prefix string, so that we can look up the kind
1762 of the wtd and use it as an index into wrote_user_func_for_ptr,
1763 ensuring that such functions are written at most once. Support
1764 subclasses by invoking the marking function of the ultimate base class.
1765 (write_user_func_for_structure_body): Require a write_types_data
1766 rather than just a prefix string, so that we can pass this to
1767 write_user_func_for_structure_ptr.
1768 (write_func_for_structure): Likewise.
1769 (ggc_wtd): Add initializer of new "kind" field.
1770 (pch_wtd): Likewise.
1772 * gengtype.h (enum write_types_kinds): New.
1773 (struct type): Add field wrote_user_func_for_ptr to the "s"
1776 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1778 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1779 instead of const_binop.
1780 (fold_binary_loc): Likewise.
1782 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1784 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1785 calculation to match get_ref_base_and_extent.
1787 2014-05-13 Catherine Moore <clm@codesourcery.com>
1788 Sandra Loosemore <sandra@codesourcery.com>
1790 * configure.ac: Fix assembly for explicit JALR relocation check.
1791 * configure: Regenerate.
1793 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1795 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1796 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1797 Remove associated type declarations and initialisations.
1798 (arm_expand_neon_builtin): Likewise.
1799 (neon_emit_pair_result_insn): Delete.
1800 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1801 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1802 (neon_vzip<mode>): Likewise.
1803 (neon_vuzp<mode>): Likewise.
1805 2014-05-13 Richard Biener <rguenther@suse.de>
1808 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1809 it needs revisiting whether the call still may be tail-called.
1811 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1813 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1814 * rtl.h (block_symbol): Reduce number of fields to 2.
1815 (rtx_def): Add u2.symbol_ref_flags.
1816 (SYMBOL_REF_FLAGS): Use it.
1817 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1818 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1819 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1820 Lower index of SYMBOL_REF_DATA.
1821 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1822 Print SYMBOL_REF_FLAGS at the same time.
1823 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1825 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1827 * rtl.def (VAR_LOCATION): Remove "i" field.
1828 * rtl.h (rtx_def): Add u2.var_location_status.
1829 (PAT_VAR_LOCATION_STATUS): Use it.
1830 (gen_rtx_VAR_LOCATION): Declare.
1831 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1832 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1833 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1835 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1837 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1838 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1840 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1842 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1843 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1844 * rtl.h (rtx_def): Add insn_uid to u2 field.
1845 (RTX_FLAG_CHECK8): Delete in favor of...
1846 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1847 (INSN_DELETED_P): Update accordingly.
1848 (INSN_UID): Use u2.insn_uid.
1849 (INSN_CHAIN_CODE_P): Define.
1850 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1851 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1852 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1853 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1854 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1855 indices accordingly.
1856 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1857 Update indices for insn-chain rtxes.
1858 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1859 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1860 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1861 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1862 * combine.c (try_combine): Likewise.
1863 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1865 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1867 * rtl.def (REG): Remove middle field.
1868 * rtl.h (rtx_def): Add orignal_regno to u2.
1869 (ORIGINAL_REGNO): Use it instead of field 1.
1870 (REG_ATTRS): Lower field index accordingly.
1871 * gengtype.c (adjust_field_rtx_def): Remove handling of
1872 ORIGINAL_REGNO. Move REG_ATTRS index down.
1873 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1874 code that prints the REGNO.
1876 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1878 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1881 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1883 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1885 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1887 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1888 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1890 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1892 * config/aarch64/aarch64-protos.h
1893 (aarch64_hard_regno_caller_save_mode): New prototype.
1894 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1896 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1898 2014-05-13 Christian Bruel <christian.bruel@st.com>
1900 * target.def (mode_switching): New hook vector.
1901 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1902 (mode_exit, modepriority_to_mode): Likewise.
1903 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1904 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1905 * target.h: Include tm.h and hard-reg-set.h.
1906 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1907 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1908 * doc/tm.texi Regenerate.
1909 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1910 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1911 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1912 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1913 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1914 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1915 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1916 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1917 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1918 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1919 (ix86_emit_mode_set): Hookify.
1920 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1922 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1923 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1924 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1925 (epiphany_mode_priority_to_mode): Remove declaration.
1926 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1927 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1928 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1930 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1931 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1932 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1934 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1937 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1938 is const0_rtx, return immediately. Don't test count == 0 when
1941 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1943 * Makefile.in: add shrink-wrap.o.
1944 * config/i386/i386.c: include "shrink-wrap.h"
1945 * function.c: Likewise.
1946 (requires_stack_frame_p, next_block_for_reg,
1947 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1948 dup_block_and_redirect): Move to shrink-wrap.c
1949 (thread_prologue_and_epilogue_insns): Extract three code segments
1950 as functions in shrink-wrap.c
1951 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1953 * shrink-wrap.c: New file.
1954 * shrink-wrap.h: New file.
1956 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1958 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1959 reference to Solaris.
1961 2014-05-12 Mike Stump <mikestump@comcast.net>
1964 * genattrtab.c (filename): Add.
1965 (convert_set_attr_alternative): Improve error message.
1966 (check_defs): Restore read_md_filename for error messages.
1967 (gen_insn): Save filename.
1969 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1971 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1972 -fno-local-ivars and -fivar-visibility.
1973 * c-family/c.opt: Make -Wshadow also implicitly enable
1976 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1978 * doc/tm.texi: Remove reference to deleted macro.
1979 * doc/tm.texi.in: Likewise.
1981 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1984 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1987 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1989 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1990 * config/arm/aout.h (License): Same.
1991 * config/arm/bpabi.h (License): Same.
1992 * config/arm/elf.h (License): Same.
1993 * config/arm/linux-elf.h (License): Same.
1994 * config/arm/linux-gas.h (License): Same.
1995 * config/arm/netbsd-elf.h (License): Same.
1996 * config/arm/uclinux-eabi.h (License): Same.
1997 * config/arm/uclinux-elf.h (License): Same.
1998 * config/arm/vxworks.h (License): Same.
2000 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2002 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2003 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2004 number of operands to 3.
2005 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2006 * tree-nested.c (convert_nonlocal_omp_clauses,
2007 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2008 * gimplify.c (gimplify_scan_omp_clauses): Handle
2009 OMP_CLAUSE_LINEAR_STMT.
2010 * omp-low.c (lower_rec_input_clauses): Fix typo.
2011 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2012 cast between Fortran boolean_type_node and C _Bool if
2015 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2017 PR tree-optimization/61136
2018 * wide-int.h (multiple_of_p): Define a version that doesn't return
2020 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2021 integer_zerop/const_binop pair.
2022 (multiple_of_p): Likewise, converting both operands to widest_int
2025 2014-05-09 Teresa Johnson <tejohnson@google.com>
2027 * cgraphunit.c (analyze_functions): Use correct dump file.
2029 2014-05-09 Florian Weimer <fweimer@redhat.com>
2031 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2033 (stack_protect_return_slot_p): New function.
2034 (expand_used_vars): Call stack_protect_decl_p and
2035 stack_protect_return_slot_p for -fstack-protector-strong.
2037 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2040 * fold-const.c (fold_binary_loc): Changed width of mask.
2042 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2044 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2045 unsigned int initializers for regno_in, regno_out.
2047 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2050 * config/avr/avr.md (cc): Add new attribute set_vzn.
2051 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2052 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2053 with INC, DEC or NEG.
2054 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2055 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2056 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2058 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2061 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2063 * wide-int.cc (UTItype): Define.
2064 (UDWtype): Define for appropriate W_TYPE_SIZE.
2066 2014-05-09 Richard Biener <rguenther@suse.de>
2068 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2069 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2070 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2071 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2072 ssa_propagate): Adjust.
2074 2014-05-08 Jeff Law <law@redhat.com>
2076 PR tree-optimization/61009
2077 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2078 tri-state rather than a boolean. When a block is too big to
2079 thread through, inform caller via negative return value.
2080 (thread_across_edge): If a block was too big for normal threading,
2081 then it's too big for a joiner too, so remove temporary equivalences
2082 and return immediately.
2084 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2085 Matthias Klose <doko@ubuntu.com>
2088 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2090 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2093 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2095 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2098 * config/alpha/alpha.c: Include gimple-iterator.h.
2099 (alpha_gimple_fold_builtin): New function. Move
2100 ALPHA_BUILTIN_UMULH folding from ...
2101 (alpha_fold_builtin): ... here.
2102 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2104 2014-05-08 Wei Mi <wmi@google.com>
2107 * config/i386/i386.c (ix86_compute_frame_layout): Update
2108 preferred_stack_boundary for call, expanded from tls descriptor.
2109 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2110 to depend on SP register.
2111 (*tls_local_dynamic_base_32_gnu): Ditto.
2112 (*tls_local_dynamic_32_once): Ditto.
2113 (tls_global_dynamic_64_<mode>): Set
2114 ix86_tls_descriptor_calls_expanded_in_cfun.
2115 (tls_local_dynamic_base_64_<mode>): Ditto.
2116 (tls_global_dynamic_32): Set
2117 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2118 to depend on SP register.
2119 (tls_local_dynamic_base_32): Ditto.
2121 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2123 * config/arm/arm_neon.h: Update comment.
2124 * config/arm/neon-docgen.ml: Delete.
2125 * config/arm/neon-gen.ml: Delete.
2126 * doc/arm-neon-intrinsics.texi: Update comment.
2128 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2130 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2132 (vand, vorr, veor, vorn, vbic): Remove.
2133 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2135 (neon_vsub_unspec): Likewise.
2136 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2138 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2140 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2141 (vadd_s16): Likewise.
2142 (vadd_s32): Likewise.
2143 (vadd_f32): Likewise.
2144 (vadd_u8): Likewise.
2145 (vadd_u16): Likewise.
2146 (vadd_u32): Likewise.
2147 (vadd_s64): Likewise.
2148 (vadd_u64): Likewise.
2149 (vaddq_s8): Likewise.
2150 (vaddq_s16): Likewise.
2151 (vaddq_s32): Likewise.
2152 (vaddq_s64): Likewise.
2153 (vaddq_f32): Likewise.
2154 (vaddq_u8): Likewise.
2155 (vaddq_u16): Likewise.
2156 (vaddq_u32): Likewise.
2157 (vaddq_u64): Likewise.
2158 (vmul_s8): Likewise.
2159 (vmul_s16): Likewise.
2160 (vmul_s32): Likewise.
2161 (vmul_f32): Likewise.
2162 (vmul_u8): Likewise.
2163 (vmul_u16): Likewise.
2164 (vmul_u32): Likewise.
2165 (vmul_p8): Likewise.
2166 (vmulq_s8): Likewise.
2167 (vmulq_s16): Likewise.
2168 (vmulq_s32): Likewise.
2169 (vmulq_f32): Likewise.
2170 (vmulq_u8): Likewise.
2171 (vmulq_u16): Likewise.
2172 (vmulq_u32): Likewise.
2173 (vsub_s8): Likewise.
2174 (vsub_s16): Likewise.
2175 (vsub_s32): Likewise.
2176 (vsub_f32): Likewise.
2177 (vsub_u8): Likewise.
2178 (vsub_u16): Likewise.
2179 (vsub_u32): Likewise.
2180 (vsub_s64): Likewise.
2181 (vsub_u64): Likewise.
2182 (vsubq_s8): Likewise.
2183 (vsubq_s16): Likewise.
2184 (vsubq_s32): Likewise.
2185 (vsubq_s64): Likewise.
2186 (vsubq_f32): Likewise.
2187 (vsubq_u8): Likewise.
2188 (vsubq_u16): Likewise.
2189 (vsubq_u32): Likewise.
2190 (vsubq_u64): Likewise.
2191 (vand_s8): Likewise.
2192 (vand_s16): Likewise.
2193 (vand_s32): Likewise.
2194 (vand_u8): Likewise.
2195 (vand_u16): Likewise.
2196 (vand_u32): Likewise.
2197 (vand_s64): Likewise.
2198 (vand_u64): Likewise.
2199 (vandq_s8): Likewise.
2200 (vandq_s16): Likewise.
2201 (vandq_s32): Likewise.
2202 (vandq_s64): Likewise.
2203 (vandq_u8): Likewise.
2204 (vandq_u16): Likewise.
2205 (vandq_u32): Likewise.
2206 (vandq_u64): Likewise.
2207 (vorr_s8): Likewise.
2208 (vorr_s16): Likewise.
2209 (vorr_s32): Likewise.
2210 (vorr_u8): Likewise.
2211 (vorr_u16): Likewise.
2212 (vorr_u32): Likewise.
2213 (vorr_s64): Likewise.
2214 (vorr_u64): Likewise.
2215 (vorrq_s8): Likewise.
2216 (vorrq_s16): Likewise.
2217 (vorrq_s32): Likewise.
2218 (vorrq_s64): Likewise.
2219 (vorrq_u8): Likewise.
2220 (vorrq_u16): Likewise.
2221 (vorrq_u32): Likewise.
2222 (vorrq_u64): Likewise.
2223 (veor_s8): Likewise.
2224 (veor_s16): Likewise.
2225 (veor_s32): Likewise.
2226 (veor_u8): Likewise.
2227 (veor_u16): Likewise.
2228 (veor_u32): Likewise.
2229 (veor_s64): Likewise.
2230 (veor_u64): Likewise.
2231 (veorq_s8): Likewise.
2232 (veorq_s16): Likewise.
2233 (veorq_s32): Likewise.
2234 (veorq_s64): Likewise.
2235 (veorq_u8): Likewise.
2236 (veorq_u16): Likewise.
2237 (veorq_u32): Likewise.
2238 (veorq_u64): Likewise.
2239 (vbic_s8): Likewise.
2240 (vbic_s16): Likewise.
2241 (vbic_s32): Likewise.
2242 (vbic_u8): Likewise.
2243 (vbic_u16): Likewise.
2244 (vbic_u32): Likewise.
2245 (vbic_s64): Likewise.
2246 (vbic_u64): Likewise.
2247 (vbicq_s8): Likewise.
2248 (vbicq_s16): Likewise.
2249 (vbicq_s32): Likewise.
2250 (vbicq_s64): Likewise.
2251 (vbicq_u8): Likewise.
2252 (vbicq_u16): Likewise.
2253 (vbicq_u32): Likewise.
2254 (vbicq_u64): Likewise.
2255 (vorn_s8): Likewise.
2256 (vorn_s16): Likewise.
2257 (vorn_s32): Likewise.
2258 (vorn_u8): Likewise.
2259 (vorn_u16): Likewise.
2260 (vorn_u32): Likewise.
2261 (vorn_s64): Likewise.
2262 (vorn_u64): Likewise.
2263 (vornq_s8): Likewise.
2264 (vornq_s16): Likewise.
2265 (vornq_s32): Likewise.
2266 (vornq_s64): Likewise.
2267 (vornq_u8): Likewise.
2268 (vornq_u16): Likewise.
2269 (vornq_u32): Likewise.
2270 (vornq_u64): Likewise.
2272 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2274 * wide-int.cc (UTItype): Define.
2275 (UDWtype): Define for appropriate W_TYPE_SIZE.
2277 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2279 PR tree-optimization/59100
2280 * tree-ssa-phiopt.c: Include tree-inline.h.
2281 (neutral_element_p, absorbing_element_p): New functions.
2282 (value_replacement): Handle conditional binary operations with a
2283 neutral or absorbing element.
2285 2014-05-08 Richard Biener <rguenther@suse.de>
2287 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2288 folding the expression.
2289 (valueize_expr): Remove.
2290 (visit_reference_op_load): Do not valueize the result of
2292 (simplify_binary_expression): Likewise.
2293 (simplify_unary_expression): Likewise.
2295 2014-05-08 Richard Biener <rguenther@suse.de>
2297 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2298 looking at TYPE_ARG_TYPES.
2300 2014-05-08 Richard Biener <rguenther@suse.de>
2302 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2303 pointer propagation special-case.
2305 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2307 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2308 core part of address expressions.
2310 2014-05-08 Alan Modra <amodra@gmail.com>
2313 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2314 loads and stores when -mno-strict-align at any alignment.
2315 (expand_block_clear): Similarly. Also correct calculation of
2318 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2321 * tree-complex.c (expand_complex_move): Keep line info when expanding
2323 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2324 of complex expression. Use new argument to display correct location
2325 for values coming from phi statement.
2326 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2327 (warn_uninitialized_phi): Pass location of phi argument to
2329 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2330 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2332 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2334 * config/rs6000/predicates.md (indexed_address_mem): New.
2335 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2336 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2337 fpstore_ux, fpstore_u.
2338 (sign_extend, indexed, update): New.
2339 (cell_micro): Adjust.
2340 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2341 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2342 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2343 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2344 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2345 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2346 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2347 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2348 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2349 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2350 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2351 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2352 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2353 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2354 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2356 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2357 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2358 *vsx_extract_<mode>_store): Adjust.
2359 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2360 is_cracked_insn, insn_must_be_first_in_group,
2361 insn_must_be_last_in_group): Adjust.
2363 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2365 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2366 ppc440-fpstore): Adjust.
2367 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2368 ppc476-fpstore): Adjust.
2369 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2370 ppc601-fpstore): Adjust.
2371 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2373 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2375 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2376 ppc7450-fpstore): Adjust.
2377 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2378 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2379 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2381 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2382 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2383 cell-fpstore, cell-fpstore-update): Adjust.
2384 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2385 ppce300c3_store, ppce300c3_fpstore): Adjust.
2386 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2387 e500mc_fpstore): Adjust.
2388 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2389 e500mc64_store, e500mc64_fpstore): Adjust.
2390 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2391 e5500_fpstore): Adjust.
2392 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2393 e6500_fpstore): Adjust.
2394 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2396 * config/rs6000/power4.md (power4-load, power4-load-ext,
2397 power4-load-ext-update, power4-load-ext-update-indexed,
2398 power4-load-update-indexed, power4-load-update, power4-fpload,
2399 power4-fpload-update, power4-store, power4-store-update,
2400 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2402 * config/rs6000/power5.md (power5-load, power5-load-ext,
2403 power5-load-ext-update, power5-load-ext-update-indexed,
2404 power5-load-update-indexed, power5-load-update, power5-fpload,
2405 power5-fpload-update, power5-store, power5-store-update,
2406 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2408 * config/rs6000/power6.md (power6-load, power6-load-ext,
2409 power6-load-update, power6-load-update-indexed,
2410 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2411 power6-fpload-update, power6-store, power6-store-update,
2412 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2414 * config/rs6000/power7.md (power7-load, power7-load-ext,
2415 power7-load-update, power7-load-update-indexed,
2416 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2417 power7-fpload-update, power7-store, power7-store-update,
2418 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2420 * config/rs6000/power8.md (power8-load, power8-load-update,
2421 power8-load-ext, power8-load-ext-update, power8-fpload,
2422 power8-fpload-update, power8-store, power8-store-update-indexed,
2423 power8-fpstore, power8-fpstore-update): Adjust.
2424 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2426 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2427 titan_lsu_store, titan_lsu_fpstore): Adjust.
2428 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2430 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2433 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2434 unrolled byte insns. Emit address increments after move insns.
2436 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2438 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2439 const_gimple, rather than a gimple.
2440 (gimple_call_builtin_p): Likewise, for the three variants.
2442 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2443 (gimple_call_builtin_p): Likewise, for the three variants.
2445 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2447 PR tree-optimization/61095
2448 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2450 2014-05-07 Richard Biener <rguenther@suse.de>
2452 PR tree-optimization/61034
2453 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2454 (maybe_skip_until): Use translate to take into account
2455 lattices when trying to do disambiguations.
2456 (get_continuation_for_phi_1): Likewise.
2457 (get_continuation_for_phi): Adjust for added translate arguments.
2458 (walk_non_aliased_vuses): Likewise.
2459 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2460 (walk_non_aliased_vuses): Likewise.
2461 (call_may_clobber_ref_p_1): Declare.
2462 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2463 calls. Stop early if we are only supposed to disambiguate.
2464 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2466 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2468 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2469 Emit an error when the function has arguments.
2471 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2473 * cfgloop.h (unswitch_loops): Remove.
2474 * doc/passes.texi: Remove references to loop-unswitch.c
2475 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2477 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2479 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2480 check for loads group of length 3.
2481 (vect_permute_load_chain): New permutations for loads group of
2483 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2484 of vec_perm_shuffle for the new permutations.
2486 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2488 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2489 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2490 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2491 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2492 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2493 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2494 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2495 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2497 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2499 * loop-unswitch.c: Delete.
2501 2014-05-07 Richard Biener <rguenther@suse.de>
2503 * config.gcc: Always set need_64bit_hwint to yes.
2505 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2507 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2508 of using optimize_size.
2510 2014-05-06 Mike Stump <mikestump@comcast.net>
2512 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2514 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2516 * config/i386/sse.md (*mov<mode>_internal)
2517 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2518 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2519 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2520 (*<code><mode>3, *andnot<mode>3<mask_name>)
2521 (<mask_codefor><code><mode>3<mask_name>): Only consider
2522 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2524 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2527 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2529 * lra-constraints.c (valid_address_p): Move earlier in file.
2530 Add a constraint argument to the address_info version.
2531 (satisfies_memory_constraint_p): New function.
2532 (satisfies_address_constraint_p): Likewise.
2533 (process_alt_operands, curr_insn_transform): Use them.
2534 (process_address): Pass the constraint to valid_address_p when
2535 checking address operands.
2537 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2539 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2540 to their respective blocks. Fix inadvertent use of "node".
2542 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2544 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2546 (init_emit_once): ...here.
2547 * rtl.h (init_derived_machine_modes): Declare.
2548 * toplev.c (do_compile): Call it even if no_backend.
2550 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2551 Mike Stump <mikestump@comcast.net>
2552 Richard Sandiford <rdsandiford@googlemail.com>
2553 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2555 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2556 (rtx_equal_for_memref_p): Update comment.
2557 (adjust_offset_for_component_ref): Use wide-int interfaces.
2558 * builtins.c (get_object_alignment_2): Likewise.
2559 (c_readstr): Likewise.
2560 (target_char_cast): Add comment.
2561 (determine_block_size): Use wide-int interfaces.
2562 (expand_builtin_signbit): Likewise.
2563 (fold_builtin_int_roundingfn): Likewise.
2564 (fold_builtin_bitop): Likewise.
2565 (fold_builtin_bswap): Likewise.
2566 (fold_builtin_logarithm): Use signop.
2567 (fold_builtin_pow): Likewise.
2568 (fold_builtin_memory_op): Use wide-int interfaces.
2569 (fold_builtin_object_size): Likewise.
2570 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2571 nb_iterations_estimate.
2572 (record_niter_bound): Use wide-int interfaces.
2573 (get_estimated_loop_iterations_int): Likewise.
2574 (get_estimated_loop_iterations): Likewise.
2575 (get_max_loop_iterations): Likewise.
2576 * cfgloop.h: Include wide-int.h.
2577 (struct nb_iter_bound): Change bound to widest_int.
2578 (struct loop): Change nb_iterations_upper_bound and
2579 nb_iterations_estimate to widest_int.
2580 (record_niter_bound): Switch to use widest_int.
2581 (get_estimated_loop_iterations): Likewise.
2582 (get_max_loop_iterations): Likewise.
2583 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2584 update for wide-int.
2585 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2586 * combine.c (try_combine): Likewise.
2587 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2588 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2590 (aarch64_float_const_representable_p): Likewise.
2591 * config/arc/arc.c: Include wide-int.h.
2592 (arc_can_use_doloop_p): Use wide-int interfaces.
2593 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2594 (vfp3_const_double_index): Likewise.
2595 * config/avr/avr.c (avr_out_round): Likewise.
2596 (avr_fold_builtin): Likewise.
2597 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2598 (bfin_can_use_doloop_p): Likewise.
2599 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2600 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2601 * config/i386/i386.c: Include wide-int.h.
2602 (ix86_data_alignment): Use wide-int interfaces.
2603 (ix86_local_alignment): Likewise.
2604 (ix86_emit_swsqrtsf): Update real_from_integer.
2605 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2606 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2607 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2608 (zero_constant): Likewise.
2609 (input_operand): Likewise.
2610 (splat_input_operand): Likewise.
2611 (non_logical_cint_operand): Change const_double to const_wide_int.
2612 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2613 (easy_altivec_constant): Remove comment.
2614 (paired_expand_vector_init): Use CONSTANT_P.
2615 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2616 (rs6000_emit_move): Update checks.
2617 (rs6000_aggregate_candidate): Use wide-int interfaces.
2618 (rs6000_expand_ternop_builtin): Likewise.
2619 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2620 (rs6000_assemble_integer): Likewise.
2621 (rs6000_hash_constant): Likewise.
2622 (output_toc): Likewise.
2623 (rs6000_rtx_costs): Likewise.
2624 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2625 * config/rs6000/rs6000-c.c: Include wide-int.h.
2626 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2627 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2628 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2629 Handle CONST_WIDE_INT.
2630 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2631 Use tree_fits_uhwi_p.
2632 * config/sparc/sparc.c: Include wide-int.h.
2633 (sparc_fold_builtin): Use wide-int interfaces.
2634 * config/vax/vax.c: Include wide-int.h.
2635 (vax_float_literal): Use real_from_integer.
2636 * coretypes.h (struct hwivec_def): New.
2638 (const_hwivec): New.
2639 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2640 (equiv_constant): Handle CONST_WIDE_INT.
2641 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2642 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2643 * dbxout.c (stabstr_U): Use wide-int interfaces.
2644 (dbxout_type): Update to use cst_fits_shwi_p.
2645 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2646 (TARGET_SUPPORTS_WIDE_INT): Add default.
2647 * dfp.c: Include wide-int.h.
2648 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2649 decimal_real_to_integer.
2650 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2651 decimal_real_to_integer.
2652 * doc/generic.texi (Constant expressions): Update for wide_int.
2653 * doc/rtl.texi (const_double): Likewise.
2654 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2655 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2656 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2657 (REAL_VALUE_FROM_INT): Remove.
2658 (TARGET_SUPPORTS_WIDE_INT): New.
2659 * doc/tm.texi: Regenerate.
2660 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2661 * double-int.h: Include wide-int.h.
2662 (struct wi::int_traits): New.
2663 * dwarf2out.c (get_full_len): New.
2664 (dw_val_equal_p): Add case dw_val_class_wide_int.
2665 (size_of_loc_descr): Likewise.
2666 (output_loc_operands): Likewise.
2667 (insert_double): Remove.
2668 (insert_wide_int): New.
2670 (print_die): Add case dw_val_class_wide_int.
2671 (attr_checksum): Likewise.
2672 (attr_checksum_ordered): Likewise.
2673 (same_dw_val_p): Likewise.
2674 (size_of_die): Likewise.
2675 (value_format): Likewise.
2676 (output_die): Likewise.
2677 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2679 (clz_loc_descriptor): Use wide-int interfaces.
2680 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2681 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2682 (round_up_to_align): Use wide-int interfaces.
2683 (field_byte_offset): Likewise.
2684 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2685 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2686 CONST_DOUBLE handling. Use wide-int interfaces.
2687 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2688 (gen_enumeration_type_die): Use add_AT_wide.
2689 (hash_loc_operands): Add case dw_val_class_wide_int.
2690 (compare_loc_operands): Likewise.
2691 * dwarf2out.h: Include wide-int.h.
2692 (wide_int_ptr): New.
2693 (enum dw_val_class): Add dw_val_class_wide_int.
2694 (struct dw_val_struct): Add val_wide.
2695 * emit-rtl.c (const_wide_int_htab): New.
2696 (const_wide_int_htab_hash): New.
2697 (const_wide_int_htab_eq): New.
2698 (lookup_const_wide_int): New.
2699 (const_double_htab_hash): Use wide-int interfaces.
2700 (const_double_htab_eq): Likewise.
2701 (rtx_to_double_int): Conditionally compile for wide-int.
2702 (immed_double_int_const): Rename to immed_wide_int_const and
2703 update for wide-int.
2704 (immed_double_const): Conditionally compile for wide-int.
2705 (init_emit_once): Use wide-int interfaces.
2706 * explow.c (plus_constant): Likewise.
2707 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2708 (lshift_value): Use wide-int interfaces.
2709 (expand_mult): Likewise.
2710 (choose_multiplier): Likewise.
2711 (expand_smod_pow2): Likewise.
2712 (make_tree): Likewise.
2713 * expr.c (convert_modes): Consolidate handling of constants.
2714 Use wide-int interfaces.
2715 (emit_group_load_1): Add note.
2716 (store_expr): Update comment.
2717 (get_inner_reference): Use wide-int interfaces.
2718 (expand_constructor): Update comment.
2719 (expand_expr_real_2): Use wide-int interfaces.
2720 (expand_expr_real_1): Likewise.
2721 (reduce_to_bit_field_precision): Likewise.
2722 (const_vector_from_tree): Likewise.
2723 * final.c: Include wide-int-print.h.
2724 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2725 * fixed-value.c: Include wide-int.h.
2726 (fixed_from_string): Use wide-int interfaces.
2727 (fixed_to_decimal): Likewise.
2728 (fixed_convert_from_real): Likewise.
2729 (real_convert_from_fixed): Likewise.
2730 * fold-const.h (mem_ref_offset): Return an offset_int.
2731 (div_if_zero_remainder): Remove code parameter.
2732 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2733 Use wide-int interfaces.
2734 (may_negate_without_overflow_p): Use wide-int interfaces.
2735 (negate_expr_p): Likewise.
2736 (fold_negate_expr): Likewise.
2737 (int_const_binop_1): Likewise.
2738 (const_binop): Likewise.
2739 (fold_convert_const_int_from_int): Likewise.
2740 (fold_convert_const_int_from_real): Likewise.
2741 (fold_convert_const_int_from_fixed): Likewise.
2742 (fold_convert_const_fixed_from_int): Likewise.
2743 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2744 (sign_bit_p): Use wide-int interfaces.
2745 (make_range_step): Likewise.
2746 (build_range_check): Likewise. Pass an integer of the correct type
2747 instead of using integer_one_node.
2748 (range_predecessor): Pass an integer of the correct type instead
2749 of using integer_one_node.
2750 (range_successor): Likewise.
2751 (merge_ranges): Likewise.
2752 (unextend): Use wide-int interfaces.
2753 (extract_muldiv_1): Likewise.
2754 (fold_div_compare): Likewise.
2755 (fold_single_bit_test): Likewise.
2756 (fold_sign_changed_comparison): Likewise.
2757 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2758 (fold_plusminus_mult_expr): Use wide-int interfaces.
2759 (native_encode_int): Likewise.
2760 (native_interpret_int): Likewise.
2761 (fold_unary_loc): Likewise.
2762 (pointer_may_wrap_p): Likewise.
2763 (size_low_cst): Likewise.
2764 (mask_with_tz): Likewise.
2765 (fold_binary_loc): Likewise.
2766 (fold_ternary_loc): Likewise.
2767 (multiple_of_p): Likewise.
2768 (tree_call_nonnegative_warnv_p): Update calls to
2769 tree_int_cst_min_precision and real_from_integer.
2770 (fold_negate_const): Use wide-int interfaces.
2771 (fold_abs_const): Likewise.
2772 (fold_relational_const): Use tree_int_cst_lt.
2773 (round_up_loc): Use wide-int interfaces.
2774 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2775 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2776 * gengtype.c: Remove include of double-int.h.
2777 (do_typedef): Use wide-int interfaces.
2778 (open_base_files): Add wide-int.h.
2779 (main): Add offset_int and widest_int typedefs.
2780 * gengtype-lex.l: Handle "^".
2781 (CXX_KEYWORD): Add "static".
2782 * gengtype-parse.c (require3): New.
2783 (require_template_declaration): Handle constant template arguments
2784 and nested templates.
2785 * gengtype-state.c: Don't include "double-int.h".
2786 * genpreds.c (write_one_predicate_function): Update comment.
2787 (write_tm_constrs_h): Add check for hval and lval use in
2789 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2790 (add_to_sequence): Likewise.
2791 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2792 and const_double_operand.
2793 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2795 * gimple-fold.c (get_base_constructor): Likewise.
2796 (fold_array_ctor_reference): Likewise.
2797 (fold_nonarray_ctor_reference): Likewise.
2798 (fold_const_aggregate_ref_1): Likewise.
2799 (gimple_val_nonnegative_real_p): Likewise.
2800 (gimple_fold_indirect_ref): Likewise.
2801 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2802 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2803 (struct slsr_cand_d): Change index to be widest_int.
2804 (struct incr_info_d): Change incr to be widest_int.
2805 (alloc_cand_and_find_basis): Use wide-int interfaces.
2806 (slsr_process_phi): Likewise.
2807 (backtrace_base_for_ref): Likewise. Return a widest_int.
2808 (restructure_reference): Take a widest_int instead of a double_int.
2809 (slsr_process_ref): Use wide-int interfaces.
2810 (create_mul_ssa_cand): Likewise.
2811 (create_mul_imm_cand): Likewise.
2812 (create_add_ssa_cand): Likewise.
2813 (create_add_imm_cand): Take a widest_int instead of a double_int.
2814 (slsr_process_add): Use wide-int interfaces.
2815 (slsr_process_cast): Likewise.
2816 (slsr_process_copy): Likewise.
2817 (dump_candidate): Likewise.
2818 (dump_incr_vec): Likewise.
2819 (replace_ref): Likewise.
2820 (cand_increment): Likewise. Return a widest_int.
2821 (cand_abs_increment): Likewise.
2822 (replace_mult_candidate): Take a widest_int instead of a double_int.
2823 (replace_unconditional_candidate): Use wide-int interfaces.
2824 (incr_vec_index): Take a widest_int instead of a double_int.
2825 (create_add_on_incoming_edge): Likewise.
2826 (create_phi_basis): Use wide-int interfaces.
2827 (replace_conditional_candidate): Likewise.
2828 (record_increment): Take a widest_int instead of a double_int.
2829 (record_phi_increments): Use wide-int interfaces.
2830 (phi_incr_cost): Take a widest_int instead of a double_int.
2831 (lowest_cost_path): Likewise.
2832 (total_savings): Likewise.
2833 (analyze_increments): Use wide-int interfaces.
2834 (ncd_with_phi): Take a widest_int instead of a double_int.
2835 (ncd_of_cand_and_phis): Likewise.
2836 (nearest_common_dominator_for_cands): Likewise.
2837 (insert_initializers): Use wide-int interfaces.
2838 (all_phi_incrs_profitable): Likewise.
2839 (replace_one_candidate): Likewise.
2840 (replace_profitable_candidates): Likewise.
2841 * godump.c: Include wide-int-print.h.
2842 (go_output_typedef): Use wide-int interfaces.
2843 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2844 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2845 (build_loop_iteration_domains): Likewise.
2846 * hooks.h: Include wide-int.h rather than double-int.h.
2847 (hook_bool_dint_dint_uint_bool_true): Delete.
2848 (hook_bool_wint_wint_uint_bool_true): Declare.
2849 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2850 (hook_bool_wint_wint_uint_bool_true): New.
2851 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2853 (ubsan_expand_si_overflow_mul_check): Likewise.
2854 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2855 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2856 (get_ancestor_addr_info): Likewise.
2857 (ipa_modify_call_arguments): Likewise.
2858 * loop-doloop.c (doloop_modify): Likewise.
2859 (doloop_optimize): Likewise.
2860 * loop-iv.c (iv_number_of_iterations): Likewise.
2861 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2862 (unroll_loop_constant_iterations): Likewise.
2863 (decide_unroll_runtime_iterations): Likewise.
2864 (unroll_loop_runtime_iterations): Likewise.
2865 (decide_peel_simple): Likewise.
2866 (decide_unroll_stupid): Likewise.
2867 * lto-streamer-in.c (streamer_read_wi): Add.
2868 (input_cfg): Use wide-int interfaces.
2869 (lto_input_tree_1): Likewise.
2870 * lto-streamer-out.c (streamer_write_wi): Add.
2871 (hash_tree): Use wide-int interfaces.
2872 (output_cfg): Likewise.
2873 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2874 (GTFILES): Add wide-int.h and signop.h.
2875 (TAGS): Look for .cc files too.
2876 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2877 * optabs.c (expand_subword_shift): Likewise.
2878 (expand_doubleword_shift): Likewise.
2879 (expand_absneg_bit): Likewise.
2880 (expand_copysign_absneg): Likewise.
2881 (expand_copysign_bit): Likewise.
2882 * postreload.c (reload_cse_simplify_set): Likewise.
2883 * predict.c (predict_iv_comparison): Likewise.
2884 * pretty-print.h: Include wide-int-print.h.
2886 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2887 * print-tree.c: Include wide-int-print.h.
2888 (print_node_brief): Use wide-int interfaces.
2889 (print_node): Likewise.
2890 * read-rtl.c (validate_const_wide_int): New.
2891 (read_rtx_code): Add CONST_WIDE_INT case.
2892 * real.c: Include wide-int.h.
2893 (real_to_integer2): Delete.
2894 (real_to_integer): New function, returning a wide_int.
2895 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2896 (ten_to_ptwo): Update call to real_from_integer.
2897 (real_digit): Likewise.
2898 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2899 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2900 (REAL_VALUE_TO_INT): Delete.
2901 (real_to_integer): Declare a wide-int form.
2902 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2903 * recog.c (const_int_operand): Improve comment.
2904 (const_scalar_int_operand): New.
2905 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2906 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2907 (split_double): Likewise.
2908 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2909 (rtx_size): Likewise.
2910 (rtx_alloc_stat_v): New.
2911 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2912 (cwi_output_hex): New.
2913 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2914 (cwi_check_failed_bounds): New.
2915 * rtl.def (CONST_WIDE_INT): New.
2916 * rtl.h: Include <utility> and wide-int.h.
2917 (struct hwivec_def): New.
2918 (CWI_GET_NUM_ELEM): New.
2919 (CWI_PUT_NUM_ELEM): New.
2920 (struct rtx_def): Add num_elem and hwiv.
2921 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2922 (CASE_CONST_UNIQUE): Likewise.
2923 (CASE_CONST_ANY): Likewise.
2924 (CONST_SCALAR_INT_P): Likewise.
2925 (CONST_WIDE_INT_P): New.
2927 (HWIVEC_CHECK): New.
2928 (cwi_check_failed_bounds): New.
2930 (HWIVEC_CHECK): New.
2931 (CONST_WIDE_INT_VEC) New.
2932 (CONST_WIDE_INT_NUNITS) New.
2933 (CONST_WIDE_INT_ELT) New.
2934 (rtx_mode_t): New type.
2935 (wi::int_traits <rtx_mode_t>): New.
2937 (wi::min_value): New.
2938 (wi::max_value): New.
2940 (const_wide_int_alloc): New.
2941 (immed_wide_int_const): New.
2942 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2943 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2944 * signop.h: New file.
2945 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2946 (simplify_const_unary_operation): Use wide-int interfaces.
2947 (simplify_binary_operation_1): Likewise.
2948 (simplify_const_binary_operation): Likewise.
2949 (simplify_const_relational_operation): Likewise.
2950 (simplify_immed_subreg): Likewise.
2951 * stmt.c (expand_case): Likewise.
2952 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2953 signop rather than a bool.
2954 * stor-layout.c (layout_type): Use wide-int interfaces.
2955 (initialize_sizetypes): Update calls to
2956 set_min_and_max_values_for_integral_type.
2957 (set_min_and_max_values_for_integral_type): Take a signop rather
2958 than a bool. Use wide-int interfaces.
2959 (fixup_signed_type): Update accordingly. Remove
2960 HOST_BITS_PER_DOUBLE_INT limit.
2961 (fixup_unsigned_type): Likewise.
2962 * system.h (STATIC_CONSTANT_P): New.
2963 (STATIC_ASSERT): New.
2964 * target.def (can_use_doloop_p): Take widest_ints rather than
2966 * target.h: Include wide-int.h rather than double-int.h.
2967 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2969 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2970 rather than INT_CST_LT_UNSIGNED.
2971 (can_use_doloop_if_innermost): Take widest_ints rather than
2973 * tree-affine.c: Include wide-int-print.h.
2974 (double_int_ext_for_comb): Delete.
2975 (wide_int_ext_for_comb): New.
2976 (aff_combination_zero): Use wide-int interfaces.
2977 (aff_combination_const): Take a widest_int instead of a double_int.
2978 (aff_combination_elt): Use wide-int interfaces.
2979 (aff_combination_scale): Take a widest_int instead of a double_int.
2980 (aff_combination_add_elt): Likewise.
2981 (aff_combination_add_cst): Likewise.
2982 (aff_combination_add): Use wide-int interfaces.
2983 (aff_combination_convert): Likewise.
2984 (tree_to_aff_combination): Likewise.
2985 (add_elt_to_tree): Take a widest_int instead of a double_int.
2986 (aff_combination_to_tree): Use wide-int interfaces.
2987 (aff_combination_remove_elt): Likewise.
2988 (aff_combination_add_product): Take a widest_int instead of
2990 (aff_combination_mult): Use wide-int interfaces.
2991 (aff_combination_expand): Likewise.
2992 (double_int_constant_multiple_p): Delete.
2993 (wide_int_constant_multiple_p): New.
2994 (aff_combination_constant_multiple_p): Take a widest_int pointer
2995 instead of a double_int pointer.
2996 (print_aff): Use wide-int interfaces.
2997 (get_inner_reference_aff): Take a widest_int pointer
2998 instead of a double_int pointer.
2999 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3000 * tree-affine.h: Include wide-int.h.
3001 (struct aff_comb_elt): Change type of coef to widest_int.
3002 (struct affine_tree_combination): Change type of offset to widest_int.
3003 (double_int_ext_for_comb): Delete.
3004 (wide_int_ext_for_comb): New.
3005 (aff_combination_const): Use widest_int instead of double_int.
3006 (aff_combination_scale): Likewise.
3007 (aff_combination_add_elt): Likewise.
3008 (aff_combination_constant_multiple_p): Likewise.
3009 (get_inner_reference_aff): Likewise.
3010 (aff_comb_cannot_overlap_p): Likewise.
3011 (aff_combination_zero_p): Use wide-int interfaces.
3012 * tree.c: Include tree.h.
3013 (init_ttree): Use make_int_cst.
3014 (tree_code_size): Removed code for INTEGER_CST case.
3015 (tree_size): Add INTEGER_CST case.
3016 (make_node_stat): Update comment.
3017 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3018 (build_int_cst_type): Use wide-int interfaces.
3019 (double_int_to_tree): Likewise.
3020 (double_int_fits_to_tree_p): Delete.
3021 (force_fit_type_double): Delete.
3022 (force_fit_type): New.
3023 (int_cst_hash_hash): Use wide-int interfaces.
3024 (int_cst_hash_eq): Likewise.
3025 (build_int_cst_wide): Delete.
3026 (wide_int_to_tree): New.
3027 (cache_integer_cst): Use wide-int interfaces.
3028 (build_low_bits_mask): Likewise.
3029 (cst_and_fits_in_hwi): Likewise.
3030 (real_value_from_int_cst): Likewise.
3031 (make_int_cst_stat): New.
3032 (integer_zerop): Use wide_int interfaces.
3033 (integer_onep): Likewise.
3034 (integer_all_onesp): Likewise.
3035 (integer_pow2p): Likewise.
3036 (integer_nonzerop): Likewise.
3037 (tree_log2): Likewise.
3038 (tree_floor_log2): Likewise.
3039 (tree_ctz): Likewise.
3040 (int_size_in_bytes): Likewise.
3041 (mem_ref_offset): Return an offset_int rather than a double_int.
3042 (build_type_attribute_qual_variant): Use wide_int interfaces.
3043 (type_hash_eq): Likewise
3044 (tree_int_cst_equal): Likewise.
3045 (tree_int_cst_lt): Delete.
3046 (tree_int_cst_compare): Likewise.
3047 (tree_fits_shwi_p): Use wide_int interfaces.
3048 (tree_fits_uhwi_p): Likewise.
3049 (tree_int_cst_sign_bit): Likewise.
3050 (tree_int_cst_sgn): Likewise.
3051 (tree_int_cst_min_precision): Take a signop rather than a bool.
3052 (simple_cst_equal): Use wide_int interfaces.
3053 (compare_tree_int): Likewise.
3054 (iterative_hash_expr): Likewise.
3055 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3057 (get_type_static_bounds): Use wide_int interfaces.
3058 (tree_int_cst_elt_check_failed): New.
3059 (build_common_tree_nodes): Reordered to set prec before filling in
3061 (int_cst_value): Check cst_and_fits_in_hwi.
3062 (widest_int_cst_value): Use wide_int interfaces.
3063 (upper_bound_in_type): Likewise.
3064 (lower_bound_in_type): Likewise.
3065 (num_ending_zeros): Likewise.
3066 (drop_tree_overflow): Likewise.
3067 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3068 (gen_conditions_for_pow_cst_base): Likewise.
3069 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3070 (group_case_labels_stmt): Use wide-int interfaces.
3071 (verify_gimple_assign_binary): Likewise.
3072 (print_loop): Likewise.
3073 * tree-chrec.c (tree_fold_binomial): Likewise.
3074 * tree-core.h (struct tree_base): Add int_length.
3075 (struct tree_int_cst): Change rep of value.
3076 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3077 (dr_may_alias_p): Likewise.
3078 (max_stmt_executions_tree): Likewise.
3079 * tree.def (INTEGER_CST): Update comment.
3080 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3081 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3082 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3083 (dequeue_and_dump): Use wide-int interfaces.
3084 * tree.h: Include wide-int.h.
3085 (NULL_TREE): Moved to earlier loc in file.
3086 (TREE_INT_CST_ELT_CHECK): New.
3087 (tree_int_cst_elt_check_failed): New.
3089 (TREE_INT_CST): Delete.
3090 (TREE_INT_CST_LOW): Use wide-int interfaces.
3091 (TREE_INT_CST_HIGH): Delete.
3092 (TREE_INT_CST_NUNITS): New.
3093 (TREE_INT_CST_EXT_NUNITS): Likewise.
3094 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3095 (TREE_INT_CST_ELT): Likewise.
3096 (INT_CST_LT): Delete.
3097 (tree_int_cst_elt_check): New (two forms).
3098 (type_code_size): Update comment.
3099 (make_int_cst_stat, make_int_cst): New.
3100 (tree_to_double_int): Delete.
3101 (double_int_fits_to_tree_p): Delete.
3102 (force_fit_type_double): Delete.
3103 (build_int_cstu): Replace with out-of-line function.
3104 (build_int_cst_wide): Delete.
3105 (tree_int_cst_lt): Define inline.
3106 (tree_int_cst_le): New.
3107 (tree_int_cst_compare): Define inline.
3108 (tree_int_cst_min_precision): Take a signop rather than a bool.
3109 (wi::int_traits <const_tree>): New.
3110 (wi::int_traits <tree>): New.
3111 (wi::extended_tree): New.
3112 (wi::int_traits <wi::extended_tree>): New.
3113 (wi::to_widest): New.
3114 (wi::to_offset): New.
3115 (wi::fits_to_tree_p): New.
3116 (wi::min_value): New.
3117 (wi::max_value): New.
3118 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3119 (copy_tree_body_r): Likewise.
3120 * tree-object-size.c (compute_object_offset): Likewise.
3121 (addr_object_size): Likewise.
3122 * tree-predcom.c: Include wide-int-print.h.
3123 (struct dref_d): Change type of offset to widest_int.
3124 (dump_dref): Call wide-int printer.
3125 (aff_combination_dr_offset): Use wide-int interfaces.
3126 (determine_offset): Take a widest_int pointer rather than a
3128 (split_data_refs_to_components): Use wide-int interfaces.
3129 (suitable_component_p): Likewise.
3130 (order_drefs): Likewise.
3131 (add_ref_to_chain): Likewise.
3132 (valid_initializer_p): Likewise.
3133 (determine_roots_comp): Likewise.
3134 * tree-pretty-print.c: Include wide-int-print.h.
3135 (dump_generic_node): Use wide-int interfaces.
3136 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3137 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3138 (move_fixed_address_to_symbol): Likewise.
3139 (move_hint_to_base): Likewise.
3140 (move_pointer_to_base): Likewise.
3141 (move_variant_to_index): Likewise.
3142 (most_expensive_mult_to_index): Likewise.
3143 (addr_to_parts): Likewise.
3144 (copy_ref_info): Likewise.
3145 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3146 (indirect_refs_may_alias_p): Likewise.
3147 (stmt_kills_ref_p_1): Likewise.
3148 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3149 * tree-ssa-ccp.c: Update comment at top of file. Include
3151 (struct prop_value_d): Change type of mask to widest_int.
3152 (extend_mask): New function.
3153 (dump_lattice_value): Use wide-int interfaces.
3154 (get_default_value): Likewise.
3155 (set_constant_value): Likewise.
3156 (set_value_varying): Likewise.
3157 (valid_lattice_transition): Likewise.
3158 (set_lattice_value): Likewise.
3159 (value_to_double_int): Delete.
3160 (value_to_wide_int): New.
3161 (get_value_from_alignment): Use wide-int interfaces.
3162 (get_value_for_expr): Likewise.
3163 (do_dbg_cnt): Likewise.
3164 (ccp_finalize): Likewise.
3165 (ccp_lattice_meet): Likewise.
3166 (bit_value_unop_1): Use widest_ints rather than double_ints.
3167 (bit_value_binop_1): Likewise.
3168 (bit_value_unop): Use wide-int interfaces.
3169 (bit_value_binop): Likewise.
3170 (bit_value_assume_aligned): Likewise.
3171 (evaluate_stmt): Likewise.
3172 (ccp_fold_stmt): Likewise.
3173 (visit_cond_stmt): Likewise.
3174 (ccp_visit_stmt): Likewise.
3175 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3176 (constant_pointer_difference): Likewise.
3177 (associate_pointerplus): Likewise.
3178 (combine_conversions): Likewise.
3179 * tree-ssa-loop.h: Include wide-int.h.
3180 (struct tree_niter_desc): Change type of max to widest_int.
3181 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3182 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3183 (remove_redundant_iv_tests): Likewise.
3184 (canonicalize_loop_induction_variables): Likewise.
3185 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3186 (constant_multiple_of): Take a widest_int pointer instead of
3187 a double_int pointer.
3188 (get_computation_aff): Use wide-int interfaces.
3189 (ptr_difference_cost): Likewise.
3190 (difference_cost): Likewise.
3191 (get_loop_invariant_expr_id): Likewise.
3192 (get_computation_cost_at): Likewise.
3193 (iv_elimination_compare_lt): Likewise.
3194 (may_eliminate_iv): Likewise.
3195 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3196 instead of double_int.
3197 (max_loop_iterations): Likewise.
3198 (max_stmt_executions): Likewise.
3199 (estimated_stmt_executions): Likewise.
3200 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3201 (split_to_var_and_offset): Use wide-int interfaces.
3202 (determine_value_range): Likewise.
3203 (bound_difference_of_offsetted_base): Likewise.
3204 (bounds_add): Take a widest_int instead of a double_int.
3205 (number_of_iterations_ne_max): Use wide-int interfaces.
3206 (number_of_iterations_ne): Likewise.
3207 (number_of_iterations_lt_to_ne): Likewise.
3208 (assert_loop_rolls_lt): Likewise.
3209 (number_of_iterations_lt): Likewise.
3210 (number_of_iterations_le): Likewise.
3211 (number_of_iterations_cond): Likewise.
3212 (number_of_iterations_exit): Likewise.
3213 (finite_loop_p): Likewise.
3214 (derive_constant_upper_bound_assign): Likewise.
3215 (derive_constant_upper_bound): Return a widest_int.
3216 (derive_constant_upper_bound_ops): Likewise.
3217 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3218 (record_estimate): Take a widest_int rather than a double_int.
3219 (record_nonwrapping_iv): Use wide-int interfaces.
3220 (double_int_cmp): Delete.
3221 (wide_int_cmp): New.
3222 (bound_index): Take a widest_int rather than a double_int.
3223 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3224 (maybe_lower_iteration_bound): Likewise.
3225 (estimate_numbers_of_iterations_loop): Likewise.
3226 (estimated_loop_iterations): Take a widest_int pointer than than
3227 a double_int pointer.
3228 (estimated_loop_iterations_int): Use wide-int interfaces.
3229 (max_loop_iterations): Take a widest_int pointer than than
3230 a double_int pointer.
3231 (max_loop_iterations_int): Use wide-int interfaces.
3232 (max_stmt_executions): Take a widest_int pointer than than
3233 a double_int pointer.
3234 (estimated_stmt_executions): Likewise.
3235 (n_of_executions_at_most): Use wide-int interfaces.
3236 (scev_probably_wraps_p): Likewise.
3237 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3239 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3241 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3242 double_ints. Adjust for trailing_wide_ints <3> representation.
3243 (set_nonzero_bits): Likewise.
3244 (get_range_info): Return wide_ints rather than double_ints.
3245 Adjust for trailing_wide_ints <3> representation.
3246 (get_nonzero_bits): Likewise.
3247 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3249 * tree-ssanames.h (struct range_info_def): Replace min, max and
3250 nonzero_bits with a trailing_wide_ints <3>.
3251 (set_range_info): Use wide_int_refs rather than double_ints.
3252 (set_nonzero_bits): Likewise.
3253 (get_range_info): Return wide_ints rather than double_ints.
3254 (get_nonzero_bits): Likewise.
3255 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3256 * tree-ssa-pre.c (phi_translate_1): Likewise.
3257 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3258 (acceptable_pow_call): Likewise.
3259 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3261 (vn_reference_fold_indirect): Likewise.
3262 (vn_reference_maybe_forwprop_address): Likewise.
3263 (valueize_refs_1): Likewise.
3264 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3265 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3266 tree_int_cst_lt and tree_int_cst_le.
3267 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3269 (streamer_alloc_tree): Likewise.
3270 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3271 (streamer_write_tree_header): Likewise.
3272 (streamer_write_integer_cst): Likewise.
3273 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3274 (build_constructors): Likewise.
3275 (array_value_type): Likewise.
3276 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3277 (vect_check_gather): Likewise.
3278 * tree-vect-generic.c (build_replicated_const): Likewise.
3279 (expand_vector_divmod): Likewise.
3280 * tree-vect-loop.c (vect_transform_loop): Likewise.
3281 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3282 (vect_do_peeling_for_alignment): Likewise.
3283 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3284 * tree-vrp.c: Include wide-int.h.
3285 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3286 (extract_range_from_assert): Use wide-int interfaces.
3287 (vrp_int_const_binop): Likewise.
3288 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3289 double_int pointers.
3290 (ranges_from_anti_range): Use wide-int interfaces.
3291 (quad_int_cmp): Delete.
3292 (quad_int_pair_sort): Likewise.
3293 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3294 (extract_range_from_unary_expr_1): Likewise.
3295 (adjust_range_with_scev): Likewise.
3296 (masked_increment): Take and return wide_ints rather than double_ints.
3297 (register_edge_assert_for_2): Use wide-int interfaces.
3298 (check_array_ref): Likewise.
3299 (search_for_addr_array): Likewise.
3300 (maybe_set_nonzero_bits): Likewise.
3301 (union_ranges): Pass an integer of the correct type instead of
3302 using integer_one_node.
3303 (intersect_ranges): Likewise.
3304 (simplify_truth_ops_using_ranges): Likewise.
3305 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3306 (range_fits_type_p): Likewise.
3307 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3309 (simplify_conversion_using_ranges): Use wide-int interfaces.
3310 (simplify_float_conversion_using_ranges): Likewise.
3311 (vrp_finalize): Likewise.
3312 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3313 (gimple_stringops_transform): Likewise.
3314 * varasm.c (decode_addr_const): Likewise.
3315 (const_hash_1): Likewise.
3316 (const_rtx_hash_1): Likewise
3317 (output_constant): Likewise.
3318 (array_size_for_constructor): Likewise.
3319 (output_constructor_regular_field): Likewise.
3320 (output_constructor_bitfield): Likewise.
3321 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3322 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3324 * gencheck.c: Define BITS_PER_UNIT.
3327 * wide-int-print.cc: New.
3328 * wide-int-print.h: New.
3330 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3332 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3334 2014-05-06 Richard Biener <rguenther@suse.de>
3336 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3337 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3338 (TODO_verify_all): Adjust.
3339 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3340 TODO_verify_stmts and TODO_verify_rtl_sharing.
3341 * bb-reorder.c: Likewise.
3342 * cfgexpand.c: Likewise.
3343 * cprop.c: Likewise.
3345 * function.c: Likewise.
3346 * fwprop.c: Likewise.
3348 * gimple-ssa-isolate-paths.c: Likewise.
3349 * gimple-ssa-strength-reduction.c: Likewise.
3350 * ipa-split.c: Likewise.
3351 * loop-init.c: Likewise.
3352 * loop-unroll.c: Likewise.
3353 * lower-subreg.c: Likewise.
3354 * modulo-sched.c: Likewise.
3355 * postreload-gcse.c: Likewise.
3356 * predict.c: Likewise.
3357 * recog.c: Likewise.
3358 * sched-rgn.c: Likewise.
3359 * store-motion.c: Likewise.
3360 * tracer.c: Likewise.
3361 * trans-mem.c: Likewise.
3362 * tree-call-cdce.c: Likewise.
3363 * tree-cfg.c: Likewise.
3364 * tree-cfgcleanup.c: Likewise.
3365 * tree-complex.c: Likewise.
3366 * tree-eh.c: Likewise.
3367 * tree-emutls.c: Likewise.
3368 * tree-if-conv.c: Likewise.
3369 * tree-into-ssa.c: Likewise.
3370 * tree-loop-distribution.c: Likewise.
3371 * tree-object-size.c: Likewise.
3372 * tree-parloops.c: Likewise.
3373 * tree-pass.h: Likewise.
3374 * tree-sra.c: Likewise.
3375 * tree-ssa-ccp.c: Likewise.
3376 * tree-ssa-copy.c: Likewise.
3377 * tree-ssa-copyrename.c: Likewise.
3378 * tree-ssa-dce.c: Likewise.
3379 * tree-ssa-dom.c: Likewise.
3380 * tree-ssa-dse.c: Likewise.
3381 * tree-ssa-forwprop.c: Likewise.
3382 * tree-ssa-ifcombine.c: Likewise.
3383 * tree-ssa-loop-ch.c: Likewise.
3384 * tree-ssa-loop-ivcanon.c: Likewise.
3385 * tree-ssa-loop.c: Likewise.
3386 * tree-ssa-math-opts.c: Likewise.
3387 * tree-ssa-phiopt.c: Likewise.
3388 * tree-ssa-phiprop.c: Likewise.
3389 * tree-ssa-pre.c: Likewise.
3390 * tree-ssa-reassoc.c: Likewise.
3391 * tree-ssa-sink.c: Likewise.
3392 * tree-ssa-strlen.c: Likewise.
3393 * tree-ssa-tail-merge.c: Likewise.
3394 * tree-ssa-uncprop.c: Likewise.
3395 * tree-switch-conversion.c: Likewise.
3396 * tree-tailcall.c: Likewise.
3397 * tree-vect-generic.c: Likewise.
3398 * tree-vectorizer.c: Likewise.
3399 * tree-vrp.c: Likewise.
3401 * var-tracking.c: Likewise.
3402 * bt-load.c: Likewise.
3403 * cfgcleanup.c: Likewise.
3404 * combine-stack-adj.c: Likewise.
3405 * combine.c: Likewise.
3406 * compare-elim.c: Likewise.
3407 * config/epiphany/resolve-sw-modes.c: Likewise.
3408 * config/i386/i386.c: Likewise.
3409 * config/mips/mips.c: Likewise.
3410 * config/s390/s390.c: Likewise.
3411 * config/sh/sh_treg_combine.cc: Likewise.
3412 * config/sparc/sparc.c: Likewise.
3415 * final.c: Likewise.
3416 * ifcvt.c: Likewise.
3417 * mode-switching.c: Likewise.
3418 * passes.c: Likewise.
3419 * postreload.c: Likewise.
3421 * reg-stack.c: Likewise.
3422 * regcprop.c: Likewise.
3423 * regrename.c: Likewise.
3426 2014-05-06 Richard Biener <rguenther@suse.de>
3429 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3430 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3432 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3435 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3437 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3438 Tom de Vries <tom@codesourcery.com>
3440 * target.def (call_fusage_contains_non_callee_clobbers): New
3442 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3444 (@node Miscellaneous Register Hooks): New node.
3445 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3446 * doc/tm.texi: Regenerate.
3448 2014-05-05 Marek Polacek <polacek@redhat.com>
3451 * opts.c (common_handle_option): Call error_at instead of warning_at.
3453 2014-05-05 Richard Biener <rguenther@suse.de>
3455 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3456 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3457 under the TODO_verify_il umbrella.
3459 2014-05-05 Richard Biener <rguenther@suse.de>
3461 * passes.c (execute_function_todo): Move TODO_verify_flow under
3462 the TODO_verify_ul umbrella.
3464 2014-05-05 Richard Biener <rguenther@suse.de>
3467 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3468 X & CST away from a CST that is the mask of a mode.
3470 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3472 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3473 int argument to enum machine_mode.
3474 (picochip_class_max_nregs): Ditto.
3475 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3476 (picochip_class_max_nregs): Ditto.
3478 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3480 * target.def: Add new target hook.
3481 * doc/tm.texi: Regenerate.
3482 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3483 * targhooks.c (default_keep_leaf_when_profiled): New function.
3485 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3486 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3488 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3490 PR tree-optimization/60363
3491 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3492 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3493 (update_destination_phis): New parameter.
3494 (create_edge_and_update_destination_phis): Ditto.
3495 (ssa_fix_duplicate_block_edges): Pass new arguments.
3496 (thread_single_edge): Ditto.
3498 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3500 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3501 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3502 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3503 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3504 Use RS6000_BTM_HARD_FLOAT.
3505 (BU_MISC_2): Likewise.
3506 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3507 RS6000_BTM_HARD_FLOAT.
3508 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3510 (rs6000_invalid_builtin): Add hard floating builtin support.
3511 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3512 hard float builtins.
3513 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3515 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3517 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3518 Add missing function* argument.
3520 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3522 * lra-constraints.c (valid_address_p): Move earlier in file.
3523 Add a constraint argument to the address_info version.
3524 (satisfies_memory_constraint_p): New function.
3525 (satisfies_address_constraint_p): Likewise.
3526 (process_alt_operands, curr_insn_transform): Use them.
3527 (process_address): Pass the constraint to valid_address_p when
3528 checking address operands.
3530 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3532 * config/mips/mips.c (mips_isa_rev): New variable.
3533 (mips_set_architecture): Set it.
3534 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3536 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3537 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3538 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3539 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3540 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3541 conditions in terms of mips_isa_rev.
3542 (mips_isa_rev): Declare.
3544 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3546 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3547 (prob_unlikely, prob_likely): Make variables const.
3549 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3551 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3553 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3555 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3557 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3559 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3560 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3561 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3563 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3564 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3566 Replace usage of ROUND_REG with sh_round_reg.
3567 Use CEIL instead of ROUND_ADVANCE.
3569 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3572 * config/sh/sh.c: Include stdlib headers before everything else.
3574 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3576 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3577 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3578 (gimplify_adjust_omp_clauses): Simd region is never
3579 directly nested in combined parallel. Instead, for linear
3580 with copyin/copyout, if in combined for simd loop, make decl
3581 firstprivate/lastprivate on OMP_FOR.
3582 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3583 expand_omp_for_static_chunk): When setting endvar, also set
3584 fd->loop.v to the same value.
3586 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3588 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3590 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3592 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3595 2014-05-02 Marek Polacek <polacek@redhat.com>
3597 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3599 2014-05-02 Kito Cheng <kito@0xlab.org>
3601 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3602 to a C expression marco.
3603 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3604 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3605 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3606 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3607 HONOR_REG_ALLOC_ORDER.
3608 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3610 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3612 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3614 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3616 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3618 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3620 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3621 (convert_scalar_cond_reduction): Likewise.
3622 (predicate_scalar_phi): Add recognition and transformation
3623 of simple conditioanl reduction to be vectorizable.
3625 2014-05-01 Marek Polacek <polacek@redhat.com>
3628 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3630 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3632 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3633 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3634 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3635 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3636 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3637 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3638 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3639 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3641 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3643 * config/arc/arc.opt (mlra): Move comment above option name
3644 to avoid mis-parsing as language options.
3646 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3648 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3649 * config/sol2.h: ... here.
3650 * config/sol2-10.h: Remove.
3652 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3653 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3654 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3655 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3656 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3657 * config/sol2.h: ... here.
3658 (SECTION_NAME_FORMAT): Don't redefine.
3659 (STARTFILE_ARCH32_SPEC): Rename to ...
3660 (STARTFILE_ARCH_SPEC): ... this.
3661 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3662 * config/sparc/sol2.h: ... here.
3663 (SECTION_NAME_FORMAT): Don't undef.
3664 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3665 (SUBTARGET_EXTRA_SPECS): Remove.
3666 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3668 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3669 (MD_STARTFILE_PREFIX): Remove.
3670 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3671 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3672 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3673 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3674 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3675 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3676 * config/i386/sol2.h: ... here.
3677 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3678 * config/i386/sol2-bi.h: Remove.
3679 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3680 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3682 * config/i386/t-sol2-64: Rename to ...
3683 * config/i386/t-sol2: ... this.
3684 * config/sparc/t-sol2-64: Rename to ...
3685 * config/sparc/t-sol2: ... this.
3687 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3688 sol2_tm_file_head, sol2_tm_file_tail.
3689 Include ${cpu_type}/sol2.h before sol2.h.
3691 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3692 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3693 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3694 Reflect i386/t-sol2-64 renaming.
3695 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3696 Reflect sparc/t-sol2-64 renaming.
3698 2014-04-30 Richard Biener <rguenther@suse.de>
3700 * passes.c (execute_function_todo): Move TODO_verify_stmts
3701 and TODO_verify_ssa under the TODO_verify_il umbrella.
3702 * tree-ssa.h (verify_ssa): Adjust prototype.
3703 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3704 we should verify SSA operands.
3705 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3706 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3707 whether we should verify whether not throwing stmts have EH info.
3708 * graphite-scop-detection.c (create_sese_edges): Adjust.
3709 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3710 * tree-eh.c (lower_try_finally_switch): Do not add the
3711 default case label twice.
3713 2014-04-30 Marek Polacek <polacek@redhat.com>
3715 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3716 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3717 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3718 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3720 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3722 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3723 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3724 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3725 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3726 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3727 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3728 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3729 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3731 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3733 * tree-cfg.c (dump_function_to_file): Dump the return type of
3734 functions, in a line to itself before the function body, mimicking
3735 the layout of a C function.
3737 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3739 PR tree-optimization/60971
3740 * tree-tailcall.c (process_assignment): Reject conversions which
3743 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3745 * calls.c (initialize_argument_information): Always treat
3746 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3747 (expand_call): Likewise.
3748 (emit_library_call_calue_1): Likewise.
3749 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3750 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3753 2014-04-29 Nick Clifton <nickc@redhat.com>
3755 * config/msp430/msp430.md (umulsidi): Fix typo.
3756 (mulhisi3): Enable even inside interrupt handlers.
3757 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3758 bigger return address pushed in large mode.
3760 2014-04-29 Nick Clifton <nickc@redhat.com>
3762 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3763 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3765 * config/m32r/m32r.c (init_reg_tables): Likewise.
3766 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3767 enum to hold the modes.
3769 2014-04-29 Richard Biener <rguenther@suse.de>
3771 * dominance.c (free_dominance_info): Add overload with
3773 (dom_info_state): Likewise.
3774 (dom_info_available_p): Likewise.
3775 * basic-block.h (free_dominance_info, dom_info_state,
3776 dom_info_available_p): Declare overloads.
3777 * passes.c (execute_function_todo): Verify that verifiers
3778 don't change dominator info state. Drop dominator info
3779 for IPA pass invocations.
3780 * cgraph.c (release_function_body): Restore asserts that
3781 dominator information is released.
3783 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3785 * doc/invoke.texi: Fix typo.
3786 * tree-vrp.c: Fix typos.
3787 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3789 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3791 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3793 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3795 * config/aarch64/aarch64-builtins.c
3796 (aarch64_types_storestruct_lane_qualifiers): New.
3797 (TYPES_STORESTRUCT_LANE): Likewise.
3798 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3799 (st3_lane): Likewise.
3800 (st4_lane): Likewise.
3801 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3802 (vec_store_lanesci_lane<mode>): Likewise.
3803 (vec_store_lanesxi_lane<mode>): Likewise.
3804 (aarch64_st2_lane<VQ:mode>): Likewise.
3805 (aarch64_st3_lane<VQ:mode>): Likewise.
3806 (aarch64_st4_lane<VQ:mode>): Likewise.
3807 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3808 * config/aarch64/arm_neon.h
3809 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3810 use new macro arguments.
3811 (__ST3_LANE_FUNC): Likewise.
3812 (__ST4_LANE_FUNC): Likewise.
3813 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3814 (V_THREE_ELEM): Likewise.
3815 (V_FOUR_ELEM): Likewise.
3817 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3819 * doc/gimple.texi: Replace the description of the now-defunct
3820 union gimple_statement_d with a diagram showing the
3821 gimple_statement_base class hierarchy and its relationships to
3822 the GSS_ and GIMPLE_ enums.
3824 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3826 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3827 * config/aarch64/aarch64.c
3828 (aarch64_cannot_change_mode_class): Weaken conditions.
3829 (aarch64_modes_tieable_p): New.
3830 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3832 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3834 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3835 (loadsync_<mode>): Change mode.
3836 (load_quadpti, store_quadpti): New.
3837 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3838 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3840 2014-04-28 Martin Jambor <mjambor@suse.cz>
3842 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3843 same alias type as the original statement.
3844 (subreplacement_assignment_data): New type.
3845 (handle_unscalarized_data_in_subtree): New type of parameter,
3846 generate new memory accesses with same alias type as the original
3848 (load_assign_lhs_subreplacements): Likewise.
3849 (sra_modify_constructor_assign): Generate new memory accesses with
3850 same alias type as the original statement.
3852 2014-04-28 Richard Biener <rguenther@suse.de>
3854 * tree-pass.h (TODO_verify_il): Define.
3855 (TODO_verify_all): Complete properly.
3856 * passes.c (execute_function_todo): Move existing loop-closed
3857 SSA verification under TODO_verify_il.
3858 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3859 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3860 Fix tree sharing issue.
3862 2014-04-28 Richard Biener <rguenther@suse.de>
3865 * builtins.def (DEF_C11_BUILTIN): Add.
3866 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3867 * coretypes.h (enum function_class): Add function_c11_misc.
3868 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3869 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3870 (call_may_clobber_ref_p_1): Likewise.
3871 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3872 (mark_all_reaching_defs_necessary_1): Likewise.
3873 (propagate_necessity): Likewise.
3874 (eliminate_unnecessary_stmts): Likewise.
3875 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3877 2014-04-28 Richard Biener <rguenther@suse.de>
3879 * tree-vrp.c (vrp_var_may_overflow): Remove.
3880 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3881 with overflow immediately bump to one before that value and
3882 let iteration figure out overflow status.
3884 2014-04-28 Richard Biener <rguenther@suse.de>
3886 * configure.ac: Do valgrind header checks unconditionally.
3887 Add --enable-valgrind-annotations.
3888 * system.h: Guard valgrind header inclusion with
3889 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3890 * alloc-pool.c (pool_alloc, pool_free): Use
3891 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3892 to guard possibly dead code.
3893 * config.in: Regenerated.
3894 * configure: Likewise.
3896 2014-04-28 Jeff Law <law@redhat.com>
3898 PR tree-optimization/60902
3899 * tree-ssa-threadedge.c
3900 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3901 over real defs when invalidating outputs from statements that do not
3902 produce useful outputs for threading.
3904 2014-04-28 Richard Biener <rguenther@suse.de>
3906 PR tree-optimization/60979
3907 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3908 SCOPs that end in a block with a successor with abnormal
3911 2014-04-28 Richard Biener <rguenther@suse.de>
3913 * tree-pass.h (execute_pass_list): Adjust prototype.
3914 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3915 (do_per_function): Change callback signature, push all actual
3916 work to the callbals.
3917 (do_per_function_toporder): Likewise.
3918 (execute_function_dump): Adjust.
3919 (execute_function_todo): Likewise.
3920 (clear_last_verified): Likewise.
3921 (verify_curr_properties): Likewise.
3922 (update_properties_after_pass): Likewise.
3923 (execute_pass_list_1): Split out from ...
3924 (execute_pass_list): ... here. Adjust.
3925 (execute_ipa_pass_list): Likewise.
3926 * cgraphunit.c (cgraph_add_new_function): Adjust.
3927 (analyze_function): Likewise.
3928 (expand_function): Likewise.
3929 * cgraph.c (release_function_body): Free dominance info
3930 here instead of asserting it was magically freed elsewhere.
3932 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3934 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3935 * configure: Regenerate.
3936 * config/sparc/sparc.opt (muser-mode): New option.
3937 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3939 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3940 * doc/invoke.texi (SPARC options): Document -muser-mode.
3942 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3944 * cselib.c (find_slot_memmode): Delete.
3945 (cselib_hasher): Change compare_type to a struct.
3946 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3948 (preserve_constants_and_equivs): Adjust for new compare_type.
3949 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3950 (wrap_constant): Delete.
3951 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3953 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3955 * doc/install.texi (Building with profile feedback): Remove
3958 2014-04-26 Tom de Vries <tom@codesourcery.com>
3960 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3963 2014-04-25 Cary Coutant <ccoutant@google.com>
3966 * dwarf2out.c (should_move_die_to_comdat): A type definition
3967 can contain a subprogram definition, but don't move it to a
3969 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3970 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3972 (clone_tree_hash): Rename to...
3973 (clone_tree_partial): ...this; change callers. Copy
3974 DW_TAG_subprogram DIEs as declarations.
3975 (copy_decls_walk): Don't copy children of a declaration into a
3978 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3981 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3984 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3986 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3987 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3989 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3992 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3994 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3996 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3998 PR tree-optimization/60930
3999 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4000 creating a multiply candidate by folding two constant
4001 multiplicands when the result overflows.
4003 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4005 PR tree-optimization/60960
4006 * tree-vect-generic.c (expand_vector_operation): Only call
4007 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4009 2014-04-25 Tom de Vries <tom@codesourcery.com>
4011 * expr.c (clobber_reg_mode): New function.
4012 * expr.h (clobber_reg): New function.
4014 2014-04-25 Tom de Vries <tom@codesourcery.com>
4016 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4019 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4020 Tom de Vries <tom@codesourcery.com>
4022 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4024 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4025 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4026 new argument to find_all_hard_reg_sets call.
4028 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4030 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4031 Use HOST_WIDE_INT_C for mask literal.
4032 (aarch_rev16_shleft_mask_imm_p): Likewise.
4034 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4037 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4039 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4041 PR preprocessor/56540
4042 * config/i386/i386-c.c (ix86_target_macros): Define
4043 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4045 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4047 * configure.ac (tga_func): Remove.
4048 (LIB_TLS_SPEC): Remove.
4049 * configure: Regenerate.
4050 * config.in: Regenerate.
4051 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4053 2014-04-25 Richard Biener <rguenther@suse.de>
4056 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4057 call stmt use/clobber sets during stmt walk instead of
4058 walking the possibly incomplete set of caller edges.
4060 2014-04-25 Richard Biener <rguenther@suse.de>
4063 * passes.c (apply_ipa_transforms): Inline into only caller ...
4064 (execute_one_pass): ... here. Properly bring in function
4065 bodies for nodes we want to apply IPA transforms to.
4067 2014-04-24 Cong Hou <congh@google.com>
4069 PR tree-optimization/60896
4070 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4071 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4072 (vect_mark_pattern_stmts): Set the def type of all statements in
4073 PATTERN_DEF_SEQ as vect_internal_def.
4075 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4077 * doc/extend.texi (PowerPC Built-in Functions): Document new
4078 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4079 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4081 * config/rs6000/predicates.md (const_0_to_3_operand): New
4082 predicate to match 0..3 integer constants.
4084 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4085 to support adding miscellaneous builtin functions.
4086 (BU_DFP_MISC_2): Likewise.
4087 (BU_P7_MISC_1): Likewise.
4088 (BU_P7_MISC_2): Likewise.
4089 (BU_P8V_MISC_3): Likewise.
4090 (BU_MISC_1): Likewise.
4091 (BU_MISC_2): Likewise.
4092 (DIVWE): Add extended divide builtin functions.
4095 (DIVWEUO): Likewise.
4099 (DIVDEUO): Likewise.
4100 (DXEX): Add decimal floating-point builtin functions.
4103 (DDEDPDQ): Likewise.
4105 (DENBCDQ): Likewise.
4112 (CDTBCD): Add new BCD builtin functions.
4116 (BCDADD_LT): Likewise.
4117 (BCDADD_EQ): Likewise.
4118 (BCDADD_GT): Likewise.
4119 (BCDADD_OV): Likewise.
4121 (BCDSUB_LT): Likewise.
4122 (BCDSUB_EQ): Likewise.
4123 (BCDSUB_GT): Likewise.
4124 (BCDSUB_OV): Likewise.
4125 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4126 (UNPACK_TD): Likewise.
4127 (PACK_TF): Likewise.
4128 (UNPACK_TF): Likewise.
4129 (UNPACK_TF_0): Likewise.
4130 (UNPACK_TF_1): Likewise.
4131 (PACK_V1TI): Likewise.
4132 (UNPACK_V1TI): Likewise.
4134 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4135 support for decimal floating point builtin functions.
4136 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4137 functions that take constant arguments.
4138 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4139 (rs6000_init_builtins): Setup long double, _Decimal64, and
4140 _Decimal128 types for new builtin functions.
4141 (builtin_function_type): Set the unsigned flags appropriately for
4142 the new builtin functions.
4143 (rs6000_opt_masks): Add support for decimal floating point builtin
4146 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4147 floating point builtin functions.
4148 (RS6000_BTM_COMMON): Likewise.
4149 (RS6000_BTI_long_double): Likewise.
4150 (RS6000_BTI_dfloat64): Likewise.
4151 (RS6000_BTI_dfloat128): Likewise.
4152 (long_double_type_internal_node): Likewise.
4153 (dfloat64_type_internal_node): Likewise.
4154 (dfloat128_type_internal_node): Likewise.
4156 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4157 2.07 bcd arithmetic instructions.
4158 (UNSPEC_BCDSUB): Likewise.
4159 (UNSPEC_BCD_OVERFLOW): Likewise.
4160 (UNSPEC_BCD_ADD_SUB): Likewise.
4161 (bcd_add_sub): Likewise.
4162 (BCD_TEST): Likewise.
4163 (bcd<bcd_add_sub>): Likewise.
4164 (bcd<bcd_add_sub>_test): Likewise.
4165 (bcd<bcd_add_sub>_test2): Likewise.
4166 (bcd<bcd_add_sub>_<code>): Likewise.
4167 (peephole2 for combined bcd ops): Likewise.
4169 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4170 decimal floating point builtin functions.
4171 (UNSPEC_DENBCD): Likewise.
4172 (UNSPEC_DXEX): Likewise.
4173 (UNSPEC_DIEX): Likewise.
4174 (UNSPEC_DSCLI): Likewise.
4175 (UNSPEC_DSCRI): Likewise.
4176 (D64_D128): Likewise.
4177 (dfp_suffix): Likewise.
4178 (dfp_ddedpd_<mode>): Likewise.
4179 (dfp_denbcd_<mode>): Likewise.
4180 (dfp_dxex_<mode>): Likewise.
4181 (dfp_diex_<mode>): Likewise.
4182 (dfp_dscli_<mode>): Likewise.
4183 (dfp_dscri_<mode>): Likewise.
4185 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4187 (UNSPEC_CDTBCD): Likewise.
4188 (UNSPEC_CBCDTD): Likewise.
4189 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4190 (UNSPEC_DIVEO): Likewise.
4191 (UNSPEC_DIVEU): Likewise.
4192 (UNSPEC_DIVEUO): Likewise.
4193 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4194 pack/unpack 128-bit types.
4195 (UNSPEC_PACK_128BIT): Likewise.
4196 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4197 (udiv<mode>3): Use idiv_ldiv mode attribute.
4198 (div<mode>3): Likewise.
4199 (addg6s): Add new BCD builtin functions.
4202 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4203 (div_extend): Likewise.
4204 (div<div_extend>_<mode>"): Likewise.
4205 (FP128_64): Add support for new builtin functions to pack/unpack
4207 (unpack<mode>): Likewise.
4208 (unpacktf_0): Likewise.
4209 (unpacktf_1): Likewise.
4210 (unpack<mode>_dm): Likewise.
4211 (unpack<mode>_nodm): Likewise.
4212 (pack<mode>): Likewise.
4213 (unpackv1ti): Likewise.
4214 (packv1ti): Likewise.
4216 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4218 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4221 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4223 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4224 * gimplify.c (omp_is_private): Change last argument's type to int.
4225 Only diagnose lastprivate if the simd argument is 1, only diagnose
4226 linear if the simd argument is 2.
4227 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4228 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4229 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4230 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4231 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4232 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4233 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4234 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4235 * tree-nested.c (convert_nonlocal_omp_clauses,
4236 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4238 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4241 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4244 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4246 * flag-types.h (enum ivar_visibility): Add.
4248 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4250 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4251 function * argument.
4253 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4255 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4257 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4258 Tom de Vries <tom@codesourcery.com>
4260 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4261 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4263 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4264 * emit-rtl.c (try_split): Same.
4266 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4267 Tom de Vries <tom@codesourcery.com>
4269 * common.opt (fuse-caller-save): New option.
4271 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4273 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4274 elements for big-endian.
4276 2014-04-24 Richard Biener <rguenther@suse.de>
4278 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4279 during TER and instead use the sepops interface for expanding
4280 non-GIMPLE_SINGLE_RHS.
4282 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4284 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4285 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4287 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4289 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4290 assembler 64-bit option.
4291 * configure: Regenerate.
4293 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4295 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4296 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4297 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4298 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4299 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4301 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4303 * config/aarch64/aarch64-builtins.c
4304 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4305 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4306 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4307 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4309 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4310 (Vrevsuff): New mode attribute.
4312 2014-04-24 Terry Guo <terry.guo@arm.com>
4314 * config/arm/arm.h (machine_function): Define variable
4315 after_arm_reorg here.
4316 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4317 (arm_split_constant): Update the way to access variable
4320 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4322 2014-04-23 Tom de Vries <tom@codesourcery.com>
4324 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4326 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4328 * is-a.h: Update comments to reflect the following changes to the
4329 "pointerness" of the API, making the template parameter match the
4330 return type, allowing use of is-a.h with typedefs of pointers.
4331 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4332 that the return type matches the parameter to the is_a_helper.
4334 (dyn_cast): Likewise.
4336 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4337 pointer from the is-a.h API.
4339 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4340 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4342 (is_a_helper <varpool_node>::test): Likewise, convert to...
4343 (is_a_helper <varpool_node *>::test): ...this.
4345 (varpool_first_variable): Update for removal of implicit pointer
4346 from the is-a.h API.
4347 (varpool_next_variable): Likewise.
4348 (varpool_first_static_initializer): Likewise.
4349 (varpool_next_static_initializer): Likewise.
4350 (varpool_first_defined_variable): Likewise.
4351 (varpool_next_defined_variable): Likewise.
4352 (cgraph_first_defined_function): Likewise.
4353 (cgraph_next_defined_function): Likewise.
4354 (cgraph_first_function): Likewise.
4355 (cgraph_next_function): Likewise.
4356 (cgraph_first_function_with_gimple_body): Likewise.
4357 (cgraph_next_function_with_gimple_body): Likewise.
4358 (cgraph_alias_target): Likewise.
4359 (varpool_alias_target): Likewise.
4360 (cgraph_function_or_thunk_node): Likewise.
4361 (varpool_variable_node): Likewise.
4362 (symtab_real_symbol_p): Likewise.
4363 * cgraphunit.c (referred_to_p): Likewise.
4364 (analyze_functions): Likewise.
4365 (handle_alias_pairs): Likewise.
4366 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4367 * gimple-ssa.h (gimple_vuse_op): Likewise.
4368 (gimple_vdef_op): Likewise.
4369 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4370 * gimple.c (gimple_build_asm_1): Likewise.
4371 (gimple_build_try): Likewise.
4372 (gimple_build_resx): Likewise.
4373 (gimple_build_eh_dispatch): Likewise.
4374 (gimple_build_omp_for): Likewise.
4375 (gimple_omp_for_set_clauses): Likewise.
4377 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4378 (is_a_helper <gimple_statement_asm *>::test): ...this.
4379 (is_a_helper <gimple_statement_bind>::test): Convert to...
4380 (is_a_helper <gimple_statement_bind *>::test): ...this.
4381 (is_a_helper <gimple_statement_call>::test): Convert to...
4382 (is_a_helper <gimple_statement_call *>::test): ...this.
4383 (is_a_helper <gimple_statement_catch>::test): Convert to...
4384 (is_a_helper <gimple_statement_catch *>::test): ...this.
4385 (is_a_helper <gimple_statement_resx>::test): Convert to...
4386 (is_a_helper <gimple_statement_resx *>::test): ...this.
4387 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4388 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4389 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4390 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4391 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4392 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4393 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4394 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4395 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4396 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4397 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4398 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4399 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4400 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4401 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4402 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4403 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4404 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4405 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4406 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4407 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4408 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4409 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4410 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4411 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4412 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4413 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4414 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4415 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4416 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4417 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4418 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4419 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4420 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4421 (is_a_helper <gimple_statement_phi>::test): Convert to...
4422 (is_a_helper <gimple_statement_phi *>::test): ...this.
4423 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4424 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4425 (is_a_helper <gimple_statement_try>::test): Convert to...
4426 (is_a_helper <gimple_statement_try *>::test): ...this.
4427 (is_a_helper <gimple_statement_wce>::test): Convert to...
4428 (is_a_helper <gimple_statement_wce *>::test): ...this.
4429 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4430 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4431 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4432 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4433 (is_a_helper <const gimple_statement_call>::test): Convert to...
4434 (is_a_helper <const gimple_statement_call *>::test): ...this.
4435 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4436 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4437 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4438 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4439 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4440 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4441 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4442 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4443 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4445 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4447 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4449 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4451 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4452 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4453 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4455 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4456 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4458 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4459 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4460 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4461 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4462 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4463 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4465 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4466 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4467 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4468 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4470 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4471 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4472 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4473 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4474 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4475 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4476 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4477 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4478 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4479 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4480 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4481 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4482 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4483 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4484 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4485 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4487 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4489 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4490 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4492 (gimple_use_ops): Update for removal of implicit pointer from the
4494 (gimple_set_use_ops): Likewise.
4495 (gimple_vuse): Likewise.
4496 (gimple_vdef): Likewise.
4497 (gimple_vuse_ptr): Likewise.
4498 (gimple_vdef_ptr): Likewise.
4499 (gimple_set_vuse): Likewise.
4500 (gimple_set_vdef): Likewise.
4501 (gimple_omp_return_set_lhs): Likewise.
4502 (gimple_omp_return_lhs): Likewise.
4503 (gimple_omp_return_lhs_ptr): Likewise.
4504 (gimple_call_fntype): Likewise.
4505 (gimple_call_set_fntype): Likewise.
4506 (gimple_call_set_internal_fn): Likewise.
4507 (gimple_call_use_set): Likewise.
4508 (gimple_call_clobber_set): Likewise.
4509 (gimple_bind_vars): Likewise.
4510 (gimple_bind_set_vars): Likewise.
4511 (gimple_bind_body_ptr): Likewise.
4512 (gimple_bind_set_body): Likewise.
4513 (gimple_bind_add_stmt): Likewise.
4514 (gimple_bind_block): Likewise.
4515 (gimple_bind_set_block): Likewise.
4516 (gimple_asm_ninputs): Likewise.
4517 (gimple_asm_noutputs): Likewise.
4518 (gimple_asm_nclobbers): Likewise.
4519 (gimple_asm_nlabels): Likewise.
4520 (gimple_asm_input_op): Likewise.
4521 (gimple_asm_input_op_ptr): Likewise.
4522 (gimple_asm_output_op): Likewise.
4523 (gimple_asm_output_op_ptr): Likewise.
4524 (gimple_asm_set_output_op): Likewise.
4525 (gimple_asm_clobber_op): Likewise.
4526 (gimple_asm_set_clobber_op): Likewise.
4527 (gimple_asm_label_op): Likewise.
4528 (gimple_asm_set_label_op): Likewise.
4529 (gimple_asm_string): Likewise.
4530 (gimple_catch_types): Likewise.
4531 (gimple_catch_types_ptr): Likewise.
4532 (gimple_catch_handler_ptr): Likewise.
4533 (gimple_catch_set_types): Likewise.
4534 (gimple_catch_set_handler): Likewise.
4535 (gimple_eh_filter_types): Likewise.
4536 (gimple_eh_filter_types_ptr): Likewise.
4537 (gimple_eh_filter_failure_ptr): Likewise.
4538 (gimple_eh_filter_set_types): Likewise.
4539 (gimple_eh_filter_set_failure): Likewise.
4540 (gimple_eh_must_not_throw_fndecl): Likewise.
4541 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4542 (gimple_eh_else_n_body_ptr): Likewise.
4543 (gimple_eh_else_e_body_ptr): Likewise.
4544 (gimple_eh_else_set_n_body): Likewise.
4545 (gimple_eh_else_set_e_body): Likewise.
4546 (gimple_try_eval_ptr): Likewise.
4547 (gimple_try_cleanup_ptr): Likewise.
4548 (gimple_try_set_eval): Likewise.
4549 (gimple_try_set_cleanup): Likewise.
4550 (gimple_wce_cleanup_ptr): Likewise.
4551 (gimple_wce_set_cleanup): Likewise.
4552 (gimple_phi_capacity): Likewise.
4553 (gimple_phi_num_args): Likewise.
4554 (gimple_phi_result): Likewise.
4555 (gimple_phi_result_ptr): Likewise.
4556 (gimple_phi_set_result): Likewise.
4557 (gimple_phi_arg): Likewise.
4558 (gimple_phi_set_arg): Likewise.
4559 (gimple_resx_region): Likewise.
4560 (gimple_resx_set_region): Likewise.
4561 (gimple_eh_dispatch_region): Likewise.
4562 (gimple_eh_dispatch_set_region): Likewise.
4563 (gimple_omp_critical_name): Likewise.
4564 (gimple_omp_critical_name_ptr): Likewise.
4565 (gimple_omp_critical_set_name): Likewise.
4566 (gimple_omp_for_clauses): Likewise.
4567 (gimple_omp_for_clauses_ptr): Likewise.
4568 (gimple_omp_for_set_clauses): Likewise.
4569 (gimple_omp_for_collapse): Likewise.
4570 (gimple_omp_for_index): Likewise.
4571 (gimple_omp_for_index_ptr): Likewise.
4572 (gimple_omp_for_set_index): Likewise.
4573 (gimple_omp_for_initial): Likewise.
4574 (gimple_omp_for_initial_ptr): Likewise.
4575 (gimple_omp_for_set_initial): Likewise.
4576 (gimple_omp_for_final): Likewise.
4577 (gimple_omp_for_final_ptr): Likewise.
4578 (gimple_omp_for_set_final): Likewise.
4579 (gimple_omp_for_incr): Likewise.
4580 (gimple_omp_for_incr_ptr): Likewise.
4581 (gimple_omp_for_set_incr): Likewise.
4582 (gimple_omp_for_pre_body_ptr): Likewise.
4583 (gimple_omp_for_set_pre_body): Likewise.
4584 (gimple_omp_parallel_clauses): Likewise.
4585 (gimple_omp_parallel_clauses_ptr): Likewise.
4586 (gimple_omp_parallel_set_clauses): Likewise.
4587 (gimple_omp_parallel_child_fn): Likewise.
4588 (gimple_omp_parallel_child_fn_ptr): Likewise.
4589 (gimple_omp_parallel_set_child_fn): Likewise.
4590 (gimple_omp_parallel_data_arg): Likewise.
4591 (gimple_omp_parallel_data_arg_ptr): Likewise.
4592 (gimple_omp_parallel_set_data_arg): Likewise.
4593 (gimple_omp_task_clauses): Likewise.
4594 (gimple_omp_task_clauses_ptr): Likewise.
4595 (gimple_omp_task_set_clauses): Likewise.
4596 (gimple_omp_task_child_fn): Likewise.
4597 (gimple_omp_task_child_fn_ptr): Likewise.
4598 (gimple_omp_task_set_child_fn): Likewise.
4599 (gimple_omp_task_data_arg): Likewise.
4600 (gimple_omp_task_data_arg_ptr): Likewise.
4601 (gimple_omp_task_set_data_arg): Likewise.
4602 (gimple_omp_taskreg_clauses): Likewise.
4603 (gimple_omp_taskreg_clauses_ptr): Likewise.
4604 (gimple_omp_taskreg_set_clauses): Likewise.
4605 (gimple_omp_taskreg_child_fn): Likewise.
4606 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4607 (gimple_omp_taskreg_set_child_fn): Likewise.
4608 (gimple_omp_taskreg_data_arg): Likewise.
4609 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4610 (gimple_omp_taskreg_set_data_arg): Likewise.
4611 (gimple_omp_task_copy_fn): Likewise.
4612 (gimple_omp_task_copy_fn_ptr): Likewise.
4613 (gimple_omp_task_set_copy_fn): Likewise.
4614 (gimple_omp_task_arg_size): Likewise.
4615 (gimple_omp_task_arg_size_ptr): Likewise.
4616 (gimple_omp_task_set_arg_size): Likewise.
4617 (gimple_omp_task_arg_align): Likewise.
4618 (gimple_omp_task_arg_align_ptr): Likewise.
4619 (gimple_omp_task_set_arg_align): Likewise.
4620 (gimple_omp_single_clauses): Likewise.
4621 (gimple_omp_single_clauses_ptr): Likewise.
4622 (gimple_omp_single_set_clauses): Likewise.
4623 (gimple_omp_target_clauses): Likewise.
4624 (gimple_omp_target_clauses_ptr): Likewise.
4625 (gimple_omp_target_set_clauses): Likewise.
4626 (gimple_omp_target_child_fn): Likewise.
4627 (gimple_omp_target_child_fn_ptr): Likewise.
4628 (gimple_omp_target_set_child_fn): Likewise.
4629 (gimple_omp_target_data_arg): Likewise.
4630 (gimple_omp_target_data_arg_ptr): Likewise.
4631 (gimple_omp_target_set_data_arg): Likewise.
4632 (gimple_omp_teams_clauses): Likewise.
4633 (gimple_omp_teams_clauses_ptr): Likewise.
4634 (gimple_omp_teams_set_clauses): Likewise.
4635 (gimple_omp_sections_clauses): Likewise.
4636 (gimple_omp_sections_clauses_ptr): Likewise.
4637 (gimple_omp_sections_set_clauses): Likewise.
4638 (gimple_omp_sections_control): Likewise.
4639 (gimple_omp_sections_control_ptr): Likewise.
4640 (gimple_omp_sections_set_control): Likewise.
4641 (gimple_omp_for_set_cond): Likewise.
4642 (gimple_omp_for_cond): Likewise.
4643 (gimple_omp_atomic_store_set_val): Likewise.
4644 (gimple_omp_atomic_store_val): Likewise.
4645 (gimple_omp_atomic_store_val_ptr): Likewise.
4646 (gimple_omp_atomic_load_set_lhs): Likewise.
4647 (gimple_omp_atomic_load_lhs): Likewise.
4648 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4649 (gimple_omp_atomic_load_set_rhs): Likewise.
4650 (gimple_omp_atomic_load_rhs): Likewise.
4651 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4652 (gimple_omp_continue_control_def): Likewise.
4653 (gimple_omp_continue_control_def_ptr): Likewise.
4654 (gimple_omp_continue_set_control_def): Likewise.
4655 (gimple_omp_continue_control_use): Likewise.
4656 (gimple_omp_continue_control_use_ptr): Likewise.
4657 (gimple_omp_continue_set_control_use): Likewise.
4658 (gimple_transaction_body_ptr): Likewise.
4659 (gimple_transaction_label): Likewise.
4660 (gimple_transaction_label_ptr): Likewise.
4661 (gimple_transaction_set_body): Likewise.
4662 (gimple_transaction_set_label): Likewise.
4664 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4665 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4666 * ipa-ref.c (ipa_record_reference): Likewise.
4667 * ipa-reference.c (analyze_function): Likewise.
4668 (ipa_reference_write_optimization_summary): Likewise.
4669 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4670 (address_taken_from_non_vtable_p): Likewise.
4671 (comdat_can_be_unshared_p_1): Likewise.
4672 * lto-cgraph.c (lto_output_ref): Likewise.
4673 (add_references): Likewise.
4674 (compute_ltrans_boundary): Likewise.
4675 (output_symtab): Likewise.
4676 (input_ref): Likewise.
4677 (input_cgraph_1): Likewise.
4678 (output_cgraph_opt_summary): Likewise.
4679 * lto-streamer-out.c (lto_output): Likewise.
4680 (output_symbol_p): Likewise.
4681 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4682 (lsei_start_function_in_partition): Likewise.
4683 (lsei_next_variable_in_partition): Likewise.
4684 (lsei_start_variable_in_partition): Likewise.
4685 * symtab.c (insert_to_assembler_name_hash): Likewise.
4686 (unlink_from_assembler_name_hash): Likewise.
4687 (symtab_unregister_node): Likewise.
4688 (symtab_remove_node): Likewise.
4689 (dump_symtab_node): Likewise.
4690 (verify_symtab_base): Likewise.
4691 (verify_symtab_node): Likewise.
4692 (symtab_make_decl_local): Likewise.
4693 (symtab_alias_ultimate_target): Likewise.
4694 (symtab_resolve_alias): Likewise.
4695 (symtab_get_symbol_partitioning_class): Likewise.
4696 * tree-phinodes.c (allocate_phi_node): Likewise.
4697 (reserve_phi_args_for_new_edge): Likewise.
4698 (remove_phi_args): Likewise.
4699 * varpool.c (varpool_node_for_asm): Likewise.
4700 (varpool_remove_unreferenced_decls): Likewise.
4702 2014-04-23 Jeff Law <law@redhat.com>
4704 PR tree-optimization/60902
4705 * tree-ssa-threadedge.c
4706 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4707 invalidate outputs from statements that do not produce useful
4708 outputs for threading.
4710 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4712 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4713 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4714 machine descriptions for Stack Smashing Protector.
4716 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4718 * aarch64.md (<optab>_rol<mode>3): New pattern.
4719 (<optab>_rolsi3_uxtw): Likewise.
4720 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4722 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4724 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4725 (arm_cortex_a12_tune): Likewise.
4727 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4729 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4731 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4733 * config/arm/arm.md (arm_rev16si2): New pattern.
4734 (arm_rev16si2_alt): Likewise.
4735 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4737 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4739 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4740 (rev16<mode>2_alt): Likewise.
4741 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4742 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4743 (aarch_rev16_shleft_mask_imm_p): Likewise.
4744 (aarch_rev16_p_1): Likewise.
4745 (aarch_rev16_p): Likewise.
4746 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4747 (aarch_rev16_shright_mask_imm_p): Likewise.
4748 (aarch_rev16_shleft_mask_imm_p): Likewise.
4750 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4752 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4753 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4755 (cortex_a53_extra_costs): Likewise.
4756 (cortex_a57_extra_costs): Likewise.
4757 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4758 (cortexa7_extra_costs): Likewise.
4759 (cortexa8_extra_costs): Likewise.
4760 (cortexa12_extra_costs): Likewise.
4761 (cortexa15_extra_costs): Likewise.
4762 (v7m_extra_costs): Likewise.
4763 (arm_new_rtx_costs): Handle BSWAP.
4765 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4767 * config/arm/arm.c (cortexa8_extra_costs): New table.
4768 (arm_cortex_a8_tune): New tuning struct.
4769 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4771 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4773 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4775 2014-04-23 Richard Biener <rguenther@suse.de>
4777 * Makefile.in (OBJS): Remove loop-unswitch.o.
4778 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4779 * passes.def (pass_rtl_unswitch): Likewise.
4780 * loop-init.c (gate_rtl_unswitch): Likewise.
4781 (rtl_unswitch): Likewise.
4782 (pass_data_rtl_unswitch): Likewise.
4783 (pass_rtl_unswitch): Likewise.
4784 (make_pass_rtl_unswitch): Likewise.
4785 * rtl.h (reversed_condition): Likewise.
4786 (compare_and_jump_seq): Likewise.
4787 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4789 * loop-unroll.c (compare_and_jump_seq): Likewise.
4791 2014-04-23 Richard Biener <rguenther@suse.de>
4793 PR tree-optimization/60903
4794 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4795 commented code block.
4796 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4797 loop flags to newly created BBs and edges.
4799 2014-04-23 Nick Clifton <nickc@redhat.com>
4801 * config/msp430/msp430.c (msp430_handle_option): Move function
4803 (msp430_option_override): Simplify mcu and mcpu option handling.
4804 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4805 support for -mhwmult command line option.
4806 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4807 -mhwmult command line option.
4808 (msp430_hwmult_enabled): Delete.
4809 (msp43o_output_labelref): Add support for -mhwmult command line option.
4810 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4811 (umulsidi3): Likewise.
4812 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4813 (mcpu, mlarge, msmall): Likewise.
4814 (mhwmult): New option.
4815 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4817 (msp430_is_f5_mcu): Remove prototype.
4818 (msp430_use_f5_series_hwmult): Add prototype.
4819 * config/msp430/msp430-opts.h: New file.
4820 * common/config/msp430: New directory.
4821 * common/config/msp430/msp430-common.c: New file.
4822 * config.gcc (msp430): Remove target_has_targetm_common.
4823 * doc/invoke.texi: Document -mhwmult command line option.
4825 2014-04-23 Nick Clifton <nickc@redhat.com>
4827 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4828 default-manifest.o if it can be found in the search path.
4829 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4831 2014-04-23 Terry Guo <terry.guo@arm.com>
4833 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4835 2014-04-23 Richard Biener <rguenther@suse.de>
4838 * tree-inline.c (declare_return_variable): Use mark_addressable.
4840 2014-04-23 Richard Biener <rguenther@suse.de>
4843 * loop-init.c (loop_optimizer_init): Make sure to apply
4844 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4846 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4849 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4851 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4852 if flag_sanitize_undefined_trap_on_error.
4853 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4854 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4855 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4856 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4857 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4858 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4859 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4860 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4861 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4862 * ubsan.c (ubsan_instrument_unreachable): Return
4863 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4864 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4865 if flag_sanitize_undefined_trap_on_error and
4866 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4867 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4868 instrument_bool_enum_load): Emit __builtin_trap () if
4869 flag_sanitize_undefined_trap_on_error and
4870 __builtin_handle_*_abort () if !flag_sanitize_recover.
4871 * doc/invoke.texi (-fsanitize-recover,
4872 -fsanitize-undefined-trap-on-error): Document.
4874 2014-04-22 Christian Bruel <christian.bruel@st.com>
4876 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4877 Force immediates to SImode.
4879 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4881 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4883 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4884 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4885 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4886 (nios2_fpu_insn): Add entry for round.
4887 (N2FPU_NO_ERRNO_P): Define.
4888 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4890 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4892 2014-04-22 Richard Henderson <rth@redhat.com>
4894 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4895 (add<GPI>3_compare0): Remove leading * from name.
4896 (add<GPI>3_carryin): Likewise.
4897 (sub<GPI>3_compare0): Likewise.
4898 (sub<GPI>3_carryin): Likewise.
4899 (<su_optab>mulditi3): New expander.
4900 (multi3): New expander.
4901 (madd<GPI>): Remove leading * from name.
4903 2014-04-22 Martin Jambor <mjambor@suse.cz>
4905 * cgraphclones.c (cgraph_function_versioning): Copy
4906 ipa_transforms_to_apply instead of asserting it is empty.
4908 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4911 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4912 on count_exp to get mode.
4914 2014-04-22 Andrew Pinski <apinski@cavium.com>
4916 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4917 Handle TLS for ILP32.
4918 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4919 (tlsie_small_<mode>): this and handle PTR.
4920 (tlsie_small_sidi): New pattern.
4921 (tlsle_small): Change to an expand to handle ILP32.
4922 (tlsle_small_<mode>): New pattern.
4923 (tlsdesc_small): Rename to ...
4924 (tlsdesc_small_<mode>): this and handle PTR.
4926 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4928 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4930 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4932 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4933 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4934 (aarch64_types_signed_poly_qualifiers): Likewise.
4935 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4936 (aarch64_types_poly_signed_qualifiers): Likewise.
4937 (TYPES_REINTERP_SS): Type macro added.
4938 (TYPES_REINTERP_SU): Likewise.
4939 (TYPES_REINTERP_SP): Likewise.
4940 (TYPES_REINTERP_US): Likewise.
4941 (TYPES_REINTERP_PS): Likewise.
4942 (aarch64_fold_builtin): New expression folding added.
4943 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4944 Declarations removed.
4945 (REINTERP_SS): Declarations added.
4946 (REINTERP_US): Likewise.
4947 (REINTERP_PS): Likewise.
4948 (REINTERP_SU): Likewise.
4949 (REINTERP_SP): Likewise.
4950 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4951 (vreinterpretq_p8_f64): Likewise.
4952 (vreinterpret_p16_f64): Likewise.
4953 (vreinterpretq_p16_f64): Likewise.
4954 (vreinterpret_f32_f64): Likewise.
4955 (vreinterpretq_f32_f64): Likewise.
4956 (vreinterpret_f64_f32): Likewise.
4957 (vreinterpret_f64_p8): Likewise.
4958 (vreinterpret_f64_p16): Likewise.
4959 (vreinterpret_f64_s8): Likewise.
4960 (vreinterpret_f64_s16): Likewise.
4961 (vreinterpret_f64_s32): Likewise.
4962 (vreinterpret_f64_s64): Likewise.
4963 (vreinterpret_f64_u8): Likewise.
4964 (vreinterpret_f64_u16): Likewise.
4965 (vreinterpret_f64_u32): Likewise.
4966 (vreinterpret_f64_u64): Likewise.
4967 (vreinterpretq_f64_f32): Likewise.
4968 (vreinterpretq_f64_p8): Likewise.
4969 (vreinterpretq_f64_p16): Likewise.
4970 (vreinterpretq_f64_s8): Likewise.
4971 (vreinterpretq_f64_s16): Likewise.
4972 (vreinterpretq_f64_s32): Likewise.
4973 (vreinterpretq_f64_s64): Likewise.
4974 (vreinterpretq_f64_u8): Likewise.
4975 (vreinterpretq_f64_u16): Likewise.
4976 (vreinterpretq_f64_u32): Likewise.
4977 (vreinterpretq_f64_u64): Likewise.
4978 (vreinterpret_s64_f64): Likewise.
4979 (vreinterpretq_s64_f64): Likewise.
4980 (vreinterpret_u64_f64): Likewise.
4981 (vreinterpretq_u64_f64): Likewise.
4982 (vreinterpret_s8_f64): Likewise.
4983 (vreinterpretq_s8_f64): Likewise.
4984 (vreinterpret_s16_f64): Likewise.
4985 (vreinterpretq_s16_f64): Likewise.
4986 (vreinterpret_s32_f64): Likewise.
4987 (vreinterpretq_s32_f64): Likewise.
4988 (vreinterpret_u8_f64): Likewise.
4989 (vreinterpretq_u8_f64): Likewise.
4990 (vreinterpret_u16_f64): Likewise.
4991 (vreinterpretq_u16_f64): Likewise.
4992 (vreinterpret_u32_f64): Likewise.
4993 (vreinterpretq_u32_f64): Likewise.
4995 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4997 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4998 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4999 (vreinterpret_p8_s8): Likewise.
5000 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5001 (vreinterpret_p8_s16): Likewise.
5002 (vreinterpret_p8_s32): Likewise.
5003 (vreinterpret_p8_s64): Likewise.
5004 (vreinterpret_p8_f32): Likewise.
5005 (vreinterpret_p8_u8): Likewise.
5006 (vreinterpret_p8_u16): Likewise.
5007 (vreinterpret_p8_u32): Likewise.
5008 (vreinterpret_p8_u64): Likewise.
5009 (vreinterpret_p8_p16): Likewise.
5010 (vreinterpretq_p8_s8): Likewise.
5011 (vreinterpretq_p8_s16): Likewise.
5012 (vreinterpretq_p8_s32): Likewise.
5013 (vreinterpretq_p8_s64): Likewise.
5014 (vreinterpretq_p8_f32): Likewise.
5015 (vreinterpretq_p8_u8): Likewise.
5016 (vreinterpretq_p8_u16): Likewise.
5017 (vreinterpretq_p8_u32): Likewise.
5018 (vreinterpretq_p8_u64): Likewise.
5019 (vreinterpretq_p8_p16): Likewise.
5020 (vreinterpret_p16_s8): Likewise.
5021 (vreinterpret_p16_s16): Likewise.
5022 (vreinterpret_p16_s32): Likewise.
5023 (vreinterpret_p16_s64): Likewise.
5024 (vreinterpret_p16_f32): Likewise.
5025 (vreinterpret_p16_u8): Likewise.
5026 (vreinterpret_p16_u16): Likewise.
5027 (vreinterpret_p16_u32): Likewise.
5028 (vreinterpret_p16_u64): Likewise.
5029 (vreinterpret_p16_p8): Likewise.
5030 (vreinterpretq_p16_s8): Likewise.
5031 (vreinterpretq_p16_s16): Likewise.
5032 (vreinterpretq_p16_s32): Likewise.
5033 (vreinterpretq_p16_s64): Likewise.
5034 (vreinterpretq_p16_f32): Likewise.
5035 (vreinterpretq_p16_u8): Likewise.
5036 (vreinterpretq_p16_u16): Likewise.
5037 (vreinterpretq_p16_u32): Likewise.
5038 (vreinterpretq_p16_u64): Likewise.
5039 (vreinterpretq_p16_p8): Likewise.
5040 (vreinterpret_f32_s8): Likewise.
5041 (vreinterpret_f32_s16): Likewise.
5042 (vreinterpret_f32_s32): Likewise.
5043 (vreinterpret_f32_s64): Likewise.
5044 (vreinterpret_f32_u8): Likewise.
5045 (vreinterpret_f32_u16): Likewise.
5046 (vreinterpret_f32_u32): Likewise.
5047 (vreinterpret_f32_u64): Likewise.
5048 (vreinterpret_f32_p8): Likewise.
5049 (vreinterpret_f32_p16): Likewise.
5050 (vreinterpretq_f32_s8): Likewise.
5051 (vreinterpretq_f32_s16): Likewise.
5052 (vreinterpretq_f32_s32): Likewise.
5053 (vreinterpretq_f32_s64): Likewise.
5054 (vreinterpretq_f32_u8): Likewise.
5055 (vreinterpretq_f32_u16): Likewise.
5056 (vreinterpretq_f32_u32): Likewise.
5057 (vreinterpretq_f32_u64): Likewise.
5058 (vreinterpretq_f32_p8): Likewise.
5059 (vreinterpretq_f32_p16): Likewise.
5060 (vreinterpret_s64_s8): Likewise.
5061 (vreinterpret_s64_s16): Likewise.
5062 (vreinterpret_s64_s32): Likewise.
5063 (vreinterpret_s64_f32): Likewise.
5064 (vreinterpret_s64_u8): Likewise.
5065 (vreinterpret_s64_u16): Likewise.
5066 (vreinterpret_s64_u32): Likewise.
5067 (vreinterpret_s64_u64): Likewise.
5068 (vreinterpret_s64_p8): Likewise.
5069 (vreinterpret_s64_p16): Likewise.
5070 (vreinterpretq_s64_s8): Likewise.
5071 (vreinterpretq_s64_s16): Likewise.
5072 (vreinterpretq_s64_s32): Likewise.
5073 (vreinterpretq_s64_f32): Likewise.
5074 (vreinterpretq_s64_u8): Likewise.
5075 (vreinterpretq_s64_u16): Likewise.
5076 (vreinterpretq_s64_u32): Likewise.
5077 (vreinterpretq_s64_u64): Likewise.
5078 (vreinterpretq_s64_p8): Likewise.
5079 (vreinterpretq_s64_p16): Likewise.
5080 (vreinterpret_u64_s8): Likewise.
5081 (vreinterpret_u64_s16): Likewise.
5082 (vreinterpret_u64_s32): Likewise.
5083 (vreinterpret_u64_s64): Likewise.
5084 (vreinterpret_u64_f32): Likewise.
5085 (vreinterpret_u64_u8): Likewise.
5086 (vreinterpret_u64_u16): Likewise.
5087 (vreinterpret_u64_u32): Likewise.
5088 (vreinterpret_u64_p8): Likewise.
5089 (vreinterpret_u64_p16): Likewise.
5090 (vreinterpretq_u64_s8): Likewise.
5091 (vreinterpretq_u64_s16): Likewise.
5092 (vreinterpretq_u64_s32): Likewise.
5093 (vreinterpretq_u64_s64): Likewise.
5094 (vreinterpretq_u64_f32): Likewise.
5095 (vreinterpretq_u64_u8): Likewise.
5096 (vreinterpretq_u64_u16): Likewise.
5097 (vreinterpretq_u64_u32): Likewise.
5098 (vreinterpretq_u64_p8): Likewise.
5099 (vreinterpretq_u64_p16): Likewise.
5100 (vreinterpret_s8_s16): Likewise.
5101 (vreinterpret_s8_s32): Likewise.
5102 (vreinterpret_s8_s64): Likewise.
5103 (vreinterpret_s8_f32): Likewise.
5104 (vreinterpret_s8_u8): Likewise.
5105 (vreinterpret_s8_u16): Likewise.
5106 (vreinterpret_s8_u32): Likewise.
5107 (vreinterpret_s8_u64): Likewise.
5108 (vreinterpret_s8_p8): Likewise.
5109 (vreinterpret_s8_p16): Likewise.
5110 (vreinterpretq_s8_s16): Likewise.
5111 (vreinterpretq_s8_s32): Likewise.
5112 (vreinterpretq_s8_s64): Likewise.
5113 (vreinterpretq_s8_f32): Likewise.
5114 (vreinterpretq_s8_u8): Likewise.
5115 (vreinterpretq_s8_u16): Likewise.
5116 (vreinterpretq_s8_u32): Likewise.
5117 (vreinterpretq_s8_u64): Likewise.
5118 (vreinterpretq_s8_p8): Likewise.
5119 (vreinterpretq_s8_p16): Likewise.
5120 (vreinterpret_s16_s8): Likewise.
5121 (vreinterpret_s16_s32): Likewise.
5122 (vreinterpret_s16_s64): Likewise.
5123 (vreinterpret_s16_f32): Likewise.
5124 (vreinterpret_s16_u8): Likewise.
5125 (vreinterpret_s16_u16): Likewise.
5126 (vreinterpret_s16_u32): Likewise.
5127 (vreinterpret_s16_u64): Likewise.
5128 (vreinterpret_s16_p8): Likewise.
5129 (vreinterpret_s16_p16): Likewise.
5130 (vreinterpretq_s16_s8): Likewise.
5131 (vreinterpretq_s16_s32): Likewise.
5132 (vreinterpretq_s16_s64): Likewise.
5133 (vreinterpretq_s16_f32): Likewise.
5134 (vreinterpretq_s16_u8): Likewise.
5135 (vreinterpretq_s16_u16): Likewise.
5136 (vreinterpretq_s16_u32): Likewise.
5137 (vreinterpretq_s16_u64): Likewise.
5138 (vreinterpretq_s16_p8): Likewise.
5139 (vreinterpretq_s16_p16): Likewise.
5140 (vreinterpret_s32_s8): Likewise.
5141 (vreinterpret_s32_s16): Likewise.
5142 (vreinterpret_s32_s64): Likewise.
5143 (vreinterpret_s32_f32): Likewise.
5144 (vreinterpret_s32_u8): Likewise.
5145 (vreinterpret_s32_u16): Likewise.
5146 (vreinterpret_s32_u32): Likewise.
5147 (vreinterpret_s32_u64): Likewise.
5148 (vreinterpret_s32_p8): Likewise.
5149 (vreinterpret_s32_p16): Likewise.
5150 (vreinterpretq_s32_s8): Likewise.
5151 (vreinterpretq_s32_s16): Likewise.
5152 (vreinterpretq_s32_s64): Likewise.
5153 (vreinterpretq_s32_f32): Likewise.
5154 (vreinterpretq_s32_u8): Likewise.
5155 (vreinterpretq_s32_u16): Likewise.
5156 (vreinterpretq_s32_u32): Likewise.
5157 (vreinterpretq_s32_u64): Likewise.
5158 (vreinterpretq_s32_p8): Likewise.
5159 (vreinterpretq_s32_p16): Likewise.
5160 (vreinterpret_u8_s8): Likewise.
5161 (vreinterpret_u8_s16): Likewise.
5162 (vreinterpret_u8_s32): Likewise.
5163 (vreinterpret_u8_s64): Likewise.
5164 (vreinterpret_u8_f32): Likewise.
5165 (vreinterpret_u8_u16): Likewise.
5166 (vreinterpret_u8_u32): Likewise.
5167 (vreinterpret_u8_u64): Likewise.
5168 (vreinterpret_u8_p8): Likewise.
5169 (vreinterpret_u8_p16): Likewise.
5170 (vreinterpretq_u8_s8): Likewise.
5171 (vreinterpretq_u8_s16): Likewise.
5172 (vreinterpretq_u8_s32): Likewise.
5173 (vreinterpretq_u8_s64): Likewise.
5174 (vreinterpretq_u8_f32): Likewise.
5175 (vreinterpretq_u8_u16): Likewise.
5176 (vreinterpretq_u8_u32): Likewise.
5177 (vreinterpretq_u8_u64): Likewise.
5178 (vreinterpretq_u8_p8): Likewise.
5179 (vreinterpretq_u8_p16): Likewise.
5180 (vreinterpret_u16_s8): Likewise.
5181 (vreinterpret_u16_s16): Likewise.
5182 (vreinterpret_u16_s32): Likewise.
5183 (vreinterpret_u16_s64): Likewise.
5184 (vreinterpret_u16_f32): Likewise.
5185 (vreinterpret_u16_u8): Likewise.
5186 (vreinterpret_u16_u32): Likewise.
5187 (vreinterpret_u16_u64): Likewise.
5188 (vreinterpret_u16_p8): Likewise.
5189 (vreinterpret_u16_p16): Likewise.
5190 (vreinterpretq_u16_s8): Likewise.
5191 (vreinterpretq_u16_s16): Likewise.
5192 (vreinterpretq_u16_s32): Likewise.
5193 (vreinterpretq_u16_s64): Likewise.
5194 (vreinterpretq_u16_f32): Likewise.
5195 (vreinterpretq_u16_u8): Likewise.
5196 (vreinterpretq_u16_u32): Likewise.
5197 (vreinterpretq_u16_u64): Likewise.
5198 (vreinterpretq_u16_p8): Likewise.
5199 (vreinterpretq_u16_p16): Likewise.
5200 (vreinterpret_u32_s8): Likewise.
5201 (vreinterpret_u32_s16): Likewise.
5202 (vreinterpret_u32_s32): Likewise.
5203 (vreinterpret_u32_s64): Likewise.
5204 (vreinterpret_u32_f32): Likewise.
5205 (vreinterpret_u32_u8): Likewise.
5206 (vreinterpret_u32_u16): Likewise.
5207 (vreinterpret_u32_u64): Likewise.
5208 (vreinterpret_u32_p8): Likewise.
5209 (vreinterpret_u32_p16): Likewise.
5210 (vreinterpretq_u32_s8): Likewise.
5211 (vreinterpretq_u32_s16): Likewise.
5212 (vreinterpretq_u32_s32): Likewise.
5213 (vreinterpretq_u32_s64): Likewise.
5214 (vreinterpretq_u32_f32): Likewise.
5215 (vreinterpretq_u32_u8): Likewise.
5216 (vreinterpretq_u32_u16): Likewise.
5217 (vreinterpretq_u32_u64): Likewise.
5218 (vreinterpretq_u32_p8): Likewise.
5219 (vreinterpretq_u32_p16): Likewise.
5221 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5223 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5225 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5227 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5228 (vqnegd_s64): Likewise.
5229 (vqabs_s64): Likewise.
5230 (vqabsd_s64): Likewise.
5232 2014-04-22 Richard Henderson <rth@redhat.com>
5234 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5235 computation to the top of the loop.
5237 2014-04-22 Renlin <renlin.li@arm.com>
5238 Jiong Wang <jiong.wang@arm.com>
5240 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5241 * config/aarch64/aarch64.c (aarch64_layout_frame)
5242 (aarch64_initial_elimination_offset): Likewise.
5244 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5246 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5249 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5251 * machmode.h (bitwise_mode_for_mode): Declare.
5252 * stor-layout.h (bitwise_type_for_mode): Likewise.
5253 * stor-layout.c (bitwise_mode_for_mode): New function.
5254 (bitwise_type_for_mode): Likewise.
5255 * builtins.c (fold_builtin_memory_op): Use it instead of
5256 int_mode_for_mode and build_nonstandard_integer_type.
5258 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5260 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5261 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5262 (*-*-solaris2*): Simplify.
5263 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5264 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5265 *-*-solaris2.9* handling.
5267 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5269 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5270 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5272 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5273 * configure: Regenerate.
5275 * config/i386/sol2-9.h: Remove.
5277 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5278 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5279 Remove Solaris 9 references.
5281 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5283 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5284 (floatuns<GPI:mode><GPF:mode>2): Remove.
5285 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5286 and floatuns conversions.
5287 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5288 and floatuns conversions.
5289 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5290 (w1,w2): New mode attributes for inequal width conversions.
5292 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5294 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5295 the output asm format.
5297 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5299 * config/aarch64/aarch64-simd.md
5300 (aarch64_cm<optab>di): Always split.
5301 (*aarch64_cm<optab>di): New.
5302 (aarch64_cmtstdi): Always split.
5303 (*aarch64_cmtstdi): New.
5305 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5307 PR tree-optimization/60823
5308 * omp-low.c (ipa_simd_modify_function_body): Go through
5309 all SSA_NAMEs and for those refering to vector arguments
5310 which are going to be replaced adjust SSA_NAME_VAR and,
5311 if it is a default definition, change it into a non-default
5312 definition assigned at the beginning of function from new_decl.
5313 (ipa_simd_modify_stmt_ops): Rewritten.
5314 * tree-dfa.c (set_ssa_default_def): When removing default def,
5315 check for NULL loc instead of NULL *loc.
5317 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5319 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5320 restrictions on core registers for DImode values in Thumb2.
5322 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5324 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5325 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5327 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5329 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5330 (*iordi_notzesidi_di): Likewise.
5331 (*iordi_notsesidi_di): Likewise.
5333 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5335 * config/arm/arm-protos.h (tune_params): New struct members.
5336 * config/arm/arm.c: Initialise tune_params per processor.
5337 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5338 for speed, based on new tune_params.
5340 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5342 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5343 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5344 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5345 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5346 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5347 (vrnda_f64): Likewise.
5348 (vrndi_f64): Likewise.
5349 (vrndm_f64): Likewise.
5350 (vrndn_f64): Likewise.
5351 (vrndp_f64): Likewise.
5352 (vrndx_f64): Likewise.
5354 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5356 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5357 GET_MODE_SIZE argument is enum machine_mode.
5359 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5362 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5363 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5365 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5368 * asan.c (asan_emit_stack_protection): Force the base to align to
5369 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5370 appropriate bits if STRICT_ALIGNMENT.
5371 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5373 (expand_used_vars): Leave a space in the stack frame for alignment
5374 if STRICT_ALIGNMENT.
5376 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5378 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5380 (gimple_store_p): Likewise.
5381 (gimple_assign_load_p): Likewise.
5382 (gimple_assign_cast_p): Likewise.
5383 (gimple_clobber_p): Likewise.
5385 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5386 rather than a gimple.
5387 (gimple_assign_cast_p): Likewise.
5389 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5392 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5393 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5395 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5396 more debug information for E500 if -mdebug=reg.
5398 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5401 * config/i386/i386.c (ix86_expand_builtin)
5402 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5403 register for target RTX.
5404 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5406 2014-04-18 Cong Hou <congh@google.com>
5408 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5409 the widen-mult pattern by handling two operands with different sizes,
5410 and operands whose size is smaller than half of the result type.
5412 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5414 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5415 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5416 (do_estimate_edge_time): Compute it.
5417 * ipa-inline.c (want_inline_small_function_p): Bypass
5418 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5420 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5422 * ipa-inline.c (spec_rem): New static variable.
5423 (dump_overall_stats): New function.
5424 (dump_inline_stats): New function.
5426 2014-04-18 Richard Henderson <rth@redhat.com>
5428 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5429 to GET_MODE_SIZE, not a reg_class_t.
5431 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5433 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5434 (vsx_xxmrglw_<mode>): Likewise.
5436 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5439 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5440 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5441 (rs6000_init_hard_regno_mode_ok): Likewise.
5443 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5445 * ipa-inline.c (inline_small_functions): Account only non-cold
5447 * doc/invoke.texi (inline-unit-growth): Update documentation.
5449 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5451 * config/rs6000/rs6000.md (addti3, subti3): New.
5453 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5456 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5457 comment. Check optimize_insn_for_size_p instead of
5458 optimize_insn_for_speed_p.
5460 2014-04-17 Martin Jambor <mjambor@suse.cz>
5462 * gimple-iterator.c (gsi_start_edge): New function.
5463 * gimple-iterator.h (gsi_start_edge): Declare.
5464 * tree-sra.c (single_non_eh_succ): New function.
5465 (disqualify_ops_if_throwing_stmt): Renamed to
5466 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5467 having one non-EH successor BB.
5468 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5469 generate loads into replacements.
5470 (sra_modify_assign): Likewise and and also use the simple path for
5472 (sra_modify_function_body): Commit statements on edges.
5474 2014-04-17 Richard Biener <rguenther@suse.de>
5477 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5478 comparison results and add clarifying comment.
5480 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5482 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5483 (blank_mode): Initialize it.
5484 (emit_mode_size_inline, emit_mode_nunits_inline,
5485 emit_mode_inner_inline): New functions.
5486 (emit_insn_modes_h): Call them and surround their output with
5487 #if GCC_VERSION >= 4001 ... #endif.
5488 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5489 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5490 mode_* arrays if the argument is __builtin_constant_p.
5491 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5492 is enum machine_mode.
5494 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5496 * passes.c (opt_pass::execute): Adjust.
5497 (pass_manager::execute_pass_mode_switching): Likewise.
5498 (early_local_passes::execute): Likewise.
5499 (execute_one_pass): Pass cfun to the pass's execute method.
5500 * tree-pass.h (opt_pass::execute): Add function * argument.
5501 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5502 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5503 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5504 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5505 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5506 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5507 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5508 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5509 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5510 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5511 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5512 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5513 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5514 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5515 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5516 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5517 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5518 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5519 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5520 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5521 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5522 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5523 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5524 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5525 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5526 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5527 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5528 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5529 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5532 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5534 * passes.c (opt_pass::gate): Take function * argument.
5535 (gate_all_early_local_passes): Merge into
5536 (early_local_passes::gate): this.
5537 (gate_all_early_optimizations): Merge into
5538 (all_early_optimizations::gate): this.
5539 (gate_all_optimizations): Mege into
5540 (all_optimizations::gate): this.
5541 (gate_all_optimizations_g): Merge into
5542 (all_optimizations_g::gate): this.
5543 (gate_rest_of_compilation): Mege into
5544 (rest_of_compilation::gate): this.
5545 (gate_postreload): Merge into
5546 (postreload::gate): this.
5547 (dump_one_pass): Pass cfun to the pass's gate method.
5548 (execute_ipa_summary_passes): Likewise.
5549 (execute_one_pass): Likewise.
5550 (ipa_write_summaries_2): Likewise.
5551 (ipa_write_optimization_summaries_1): Likewise.
5552 (ipa_read_summaries_1): Likewise.
5553 (ipa_read_optimization_summaries_1): Likewise.
5554 (execute_ipa_stmt_fixups): Likewise.
5555 * tree-pass.h (opt_pass::gate): Add function * argument.
5556 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5557 combine-stack-adj.c, combine.c, compare-elim.c,
5558 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5559 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5560 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5561 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5562 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5563 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5564 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5565 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5566 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5567 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5568 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5569 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5570 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5571 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5572 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5573 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5574 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5575 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5576 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5577 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5578 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5579 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5580 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5581 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5582 var-tracking.c, vtable-verify.c, web.c: Adjust.
5584 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5586 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5587 * configure: Regenerate.
5589 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5591 * passes.c (dump_one_pass): don't check pass->has_gate.
5592 (execute_ipa_summary_passes): Likewise.
5593 (execute_one_pass): Likewise.
5594 (ipa_write_summaries_2): Likewise.
5595 (ipa_write_optimization_summaries_1): Likewise.
5596 (ipa_read_optimization_summaries_1): Likewise.
5597 (execute_ipa_stmt_fixups): Likewise.
5598 * tree-pass.h (pass_data::has_gate): Remove.
5599 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5600 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5601 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5602 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5603 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5604 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5605 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5606 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5607 gimple-low.c, gimple-ssa-isolate-paths.c,
5608 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5609 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5610 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5611 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5612 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5613 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5614 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5615 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5616 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5617 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5618 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5619 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5620 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5621 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5622 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5623 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5624 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5625 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5626 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5627 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5628 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5629 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5632 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5634 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5636 * passes.c (pass_manager::register_dump_files_1): Merge into
5637 (pass_manager::register_dump_files): this, and remove its handling of
5638 properties since the pass always has the properties anyway.
5639 (pass_manager::pass_manager): Adjust.
5641 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5643 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5644 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5645 dealing with properties.
5646 (pass_manager::register_dump_files): Adjust.
5648 2014-03-20 Mark Wielaard <mjw@redhat.com>
5650 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5651 then represent the bound as normal constant value.
5653 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5656 Forward port from 4.8 branch
5657 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5659 * config/i386/bmiintrin.h (_blsi_u32): New.
5663 (_blsmsk_u32): Ditto.
5664 (_blsmsk_u64): Ditto.
5665 (_tzcnt_u32): Ditto.
5666 (_tzcnt_u64): Ditto.
5668 2014-04-17 Kito Cheng <kito@0xlab.org>
5670 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5672 2014-04-17 Richard Biener <rguenther@suse.de>
5675 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5676 boolean results for comparisons.
5678 2014-04-17 Richard Biener <rguenther@suse.de>
5680 PR tree-optimization/60836
5681 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5682 initial PHI args to be gimple values.
5684 2014-04-17 Richard Biener <rguenther@suse.de>
5686 PR tree-optimization/60841
5687 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5688 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5689 of stmts to SLP build.
5690 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5691 (vect_analyze_slp): Likewise.
5692 (vect_analyze_slp_instance): Likewise.
5693 (vect_build_slp_tree): Limit overall SLP tree growth.
5694 * tree-vectorizer.h (vect_analyze_data_refs,
5695 vect_analyze_slp): Adjust prototypes.
5697 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5699 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5702 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5704 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5705 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5706 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5707 for TARGET_SLOW_PSHUFB
5709 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5711 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5712 * config/i386/i386.c (intel_cost): Ditto.
5714 2014-04-17 Joey Ye <joey.ye@arm.com>
5716 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5718 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5720 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5723 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5725 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5726 (type_all_derivations_known_p): New predicate.
5727 (type_all_ctors_visible_p): New predicate.
5728 (type_possibly_instantiated_p): New predicate.
5729 (get_odr_type): Compute all_derivations_known.
5730 (dump_odr_type): Dump the flag.
5731 (maybe_record_type): Cleanup.
5732 (record_target_from_binfo): Add bases_to_consider array;
5733 record bases for types w/o instances and skip CXX destructor.
5734 (possible_polymorphic_call_targets_1): Add bases_to_consider
5735 and consider_construction parameters; check if type may have instance.
5736 (get_polymorphic_call_info): Set maybe_in_construction to true
5737 when we know nothing.
5738 (record_targets_from_bases): Skip CXX destructors; they are
5739 never called for types in construction.
5740 (possible_polymorphic_call_targets): Do not record target when
5741 type may not have instance.
5743 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5746 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5747 external aliases alive, too.
5749 2014-04-16 Andrew Pinski <apinski@cavium.com>
5751 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5754 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5756 * final.c (compute_alignments): Do not apply loop alignment to a block
5757 falling through to the exit.
5759 2014-04-16 Catherine Moore <clm@codesourcery.com>
5761 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5762 Adjust constraints for microMIPS store patterns.
5764 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5766 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5768 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5770 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5771 (append_use): Run at -O0.
5772 (append_vdef): Likewise.
5773 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5774 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5776 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5778 PR tree-optimization/60844
5779 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5780 (propagate_op_to_single_use, remove_visited_stmt_chain,
5781 linearize_expr, repropagate_negates, reassociate_bb): Use it
5782 instead of gsi_remove.
5784 2014-04-16 Martin Jambor <mjambor@suse.cz>
5786 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5787 ipa_transforms_to_apply.
5788 (cgraph_function_versioning): Assert that old_node has empty
5789 ipa_transforms_to_apply.
5790 * trans-mem.c (ipa_tm_create_version): Likewise.
5791 * tree-inline.c (tree_function_versioning): Do not duplicate
5792 ipa_transforms_to_apply.
5794 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5797 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5799 Pass necessary as flags on 64-bit Solaris/x86.
5800 Use lowercase relocs for x86_64-*-*.
5801 * configure: Regenerate.
5803 2014-04-15 Jan Hubicka <jh@suse.cz>
5805 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5806 (maybe_record_node, likely_target_p): Use it.
5808 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5811 Revert following patch
5813 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5816 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5817 software floating point or no floating point registers, do not
5818 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5819 in GPRs that occurs after we tested for GPRs that would never be
5822 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5823 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5824 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5825 specifically allow DDmode, since that does not use the SPE SIMD
5828 2014-03-21 Mark Wielaard <mjw@redhat.com>
5830 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5831 as unsigned or int depending on type and value used.
5833 2014-04-15 Richard Biener <rguenther@suse.de>
5835 PR rtl-optimization/56965
5836 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5837 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5839 * alias.c (true_dependence_1): Do not call
5840 nonoverlapping_component_refs_p.
5841 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5842 nonoverlapping_component_refs_p.
5843 (indirect_refs_may_alias_p): Likewise.
5845 2014-04-15 Teresa Johnson <tejohnson@google.com>
5847 * cfg.c (dump_bb_info): Fix flags check.
5848 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5850 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5852 PR rtl-optimization/60663
5853 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5856 2014-04-15 Richard Biener <rguenther@suse.de>
5858 * lto-streamer.h (LTO_major_version): Bump to 4.
5860 2014-04-15 Richard Biener <rguenther@suse.de>
5862 * common.opt (lto_partition_model): New enum.
5863 (flto-partition=): Merge separate options with a single with argument,
5864 add -flto-partition=one support.
5865 * flag-types.h (enum lto_partition_model): Declare.
5866 * opts.c (finish_options): Remove duplicate -flto-partition=
5868 * lto-wrapper.c (run_gcc): Adjust.
5870 2014-04-15 Richard Biener <rguenther@suse.de>
5872 * alias.c (ncr_compar): New function.
5873 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5875 2014-04-15 Richard Biener <rguenther@suse.de>
5877 * alias.c (record_component_aliases): Do not walk BINFOs.
5879 2014-04-15 Richard Biener <rguenther@suse.de>
5881 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5882 Add struct function argument and adjust.
5883 (find_func_aliases_for_call): Likewise.
5884 (find_func_aliases): Likewise.
5885 (find_func_clobbers): Likewise.
5886 (intra_create_variable_infos): Likewise.
5887 (compute_points_to_sets): Likewise.
5888 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5890 2014-04-15 Richard Biener <rguenther@suse.de>
5892 * tree.c (iterative_hash_expr): Use enum tree_code_class
5893 to store TREE_CODE_CLASS.
5894 (tree_block): Likewise.
5895 (tree_set_block): Likewise.
5896 * tree.h (fold_build_pointer_plus_loc): Use
5897 convert_to_ptrofftype_loc.
5899 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5902 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5905 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5907 * cfgloop.h (struct loop): Move force_vectorize down.
5908 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5909 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5910 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5911 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5912 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5913 * tree-core.h (enum annot_expr_kind): Add new kind values.
5914 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5915 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5917 * tree.def (ANNOTATE_EXPR): Tweak comment.
5919 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5921 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5924 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5926 * tree.h (TYPE_IDENTIFIER): Declare.
5927 * tree.c (subrange_type_for_debug_p): Use it.
5928 * godump.c (go_format_type): Likewise.
5929 * dwarf2out.c (is_cxx_auto, modified_type_die,
5930 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5931 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5933 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5936 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5938 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5940 * config/i386/i386.c (examine_argument): Return bool. Return true if
5941 parameter should be passed in memory.
5942 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5943 (construct_container): Update calls to examine_argument.
5944 (function_arg_advance_64): Ditto.
5945 (return_in_memory_32): Merge with ix86_return_in_memory.
5946 (return_in_memory_64): Ditto.
5947 (return_in_memory_ms_64): Ditto.
5949 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5951 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5952 * coverage.c (coverage_compute_profile_id): Handle externally visible
5955 2014-04-14 Martin Jambor <mjambor@suse.cz>
5957 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5958 DECL_DISREGARD_INLINE_LIMITS functions.
5960 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5963 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5965 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5968 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5969 optimize_insn_for_speed_p instead of
5970 optimize_function_for_speed_p.
5972 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5974 * doc/invoke.texi (free): Document AArch64.
5976 2014-04-14 Richard Biener <rguenther@suse.de>
5978 PR tree-optimization/60042
5979 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5980 (insert_into_preds_of_block): Do not prevent PHI insertion
5981 for REFERENCE exprs here ...
5982 (eliminate_dom_walker::before_dom_children): ... but prevent
5983 their use here under similar conditions when applied to the
5984 IL after PRE optimizations.
5986 2014-04-14 Richard Biener <rguenther@suse.de>
5988 * passes.def: Move early points-to after early SRA.
5990 2014-04-14 Richard Biener <rguenther@suse.de>
5992 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5993 check for which sign-changes we allow when forwarding
5994 a converted value into a switch.
5996 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5998 * stor-layout.c (place_field): Finalize non-constant offset for the
6001 2014-04-14 Richard Biener <rguenther@suse.de>
6003 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6005 (expand_switch_using_bit_tests_p): Likewise.
6006 (process_switch): Compute and pass on speed_p based on the
6008 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6009 optimize_bb_for_speed_p.
6011 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6013 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6014 * function.h (struct function): Rename has_force_vect_loops into
6015 has_force_vectorize_loops.
6016 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6017 (input_struct_function_base): Likewise.
6018 * lto-streamer-out.c (output_cfg): Likewise.
6019 (output_struct_function_base): Likewise.
6020 * omp-low.c (expand_omp_simd): Likewise.
6021 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6022 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6023 (version_loop_for_if_conversion): Likewise.
6024 (tree_if_conversion): Likewise.
6025 (main_tree_if_conversion): Likewise.
6026 (gate_tree_if_conversion): Likewise.
6027 * tree-inline.c (copy_loops): Likewise.
6028 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6029 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6030 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6031 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6032 * tree-vectorizer.c (vectorize_loops): Likewise.
6033 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6035 2014-04-14 Richard Biener <rguenther@suse.de>
6038 * lto-streamer-out.c (wrap_refs): New function.
6039 (lto_output): Wrap symbol references in global initializes in
6040 type-preserving MEM_REFs.
6042 2014-04-14 Christian Bruel <christian.bruel@st.com>
6044 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6046 2014-04-14 Christian Bruel <christian.bruel@st.com>
6048 * config/sh/sh.md (setmemqi): New expand pattern.
6049 * config/sh/sh.h (CLEAR_RATIO): Define.
6050 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6051 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6053 2014-04-14 Richard Biener <rguenther@suse.de>
6056 * fold-const.c (negate_expr_p): Don't negate directional rounding
6058 (fold_negate_expr): Likewise.
6060 2014-04-14 Richard Biener <rguenther@suse.de>
6062 PR tree-optimization/59817
6063 PR tree-optimization/60453
6064 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6065 recursion to catch all CHRECs in the scalar evolution and restrict
6066 the predicate for the remains appropriately.
6068 2014-04-12 Catherine Moore <clm@codesourcery.com>
6070 * config/mips/constraints.md: Add new register constraint "kb".
6071 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6072 (*movhi_internal): Likewise.
6073 (*movqi_internal): Likewise.
6074 * config/mips/mips.h (M16_STORE_REGS): New register class.
6075 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6076 (REG_CLASS_CONTENTS): Likewise.
6077 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6079 2014-04-11 Tobias Burnus <burnus@net-b.de>
6082 * doc/invoke.texi (-Wformat-signedness): Document it.
6083 (Wformat=2): Mention that this enables -Wformat-signedness.
6085 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6087 * common/config/epiphany/epiphany-common.c
6088 (epiphany_option_optimization_table): Enable section anchors by
6089 default at -O1 or higher.
6090 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6091 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6092 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6093 carries no extra cost.
6094 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6095 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6096 * config/epiphany/predicates.md (memclob_operand): New predicate.
6097 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6098 Use memclob_operand predicate and X constraint for operand 3.
6100 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6102 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6103 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6106 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6108 PR rtl-optimization/60651
6109 * mode-switching.c (optimize_mode_switching): Make sure to emit
6110 sets of a lower numbered entity before sets of a higher numbered
6111 entity to a mode of the same or lower priority.
6112 When creating a seginfo for a basic block that starts with a code
6113 label, move the insertion point past the code label.
6114 (new_seginfo): Document and enforce requirement that
6115 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6116 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6117 * doc/tm.texi: Regenerate.
6119 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6122 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6124 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6126 * BASE-VER: Set to 4.10.0.
6128 2014-04-11 Tobias Burnus <burnus@net-b.de>
6131 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6132 * doc/gcc.texi (Service): Update description in the @menu
6133 * doc/invoke.texi (Option Summary): Remove misplaced and
6136 2014-04-11 Steve Ellcey <sellcey@mips.com>
6137 Jakub Jelinek <jakub@redhat.com>
6140 * expr.c (convert_move): Use emit_store_flag_force instead of
6141 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6144 2014-04-11 Richard Biener <rguenther@suse.de>
6147 * varasm.c (assemble_alias): Avoid endless error reporting
6148 recursion by setting TREE_ASM_WRITTEN.
6150 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6152 * config/s390/s390.md: Add a splitter for NOT rtx.
6154 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6156 PR rtl-optimization/60663
6157 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6159 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6160 Jakub Jelinek <jakub@redhat.com>
6163 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6164 flag from decl_node to node.
6166 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6169 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6170 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6171 ameliorating the cases where it can be.
6173 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6176 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6178 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6179 (loadsync_<mode>): Change mode.
6180 (load_quadpti, store_quadpti): New.
6181 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6182 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6183 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6185 2014-04-09 Cong Hou <congh@google.com>
6188 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6191 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6193 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6194 instead of vnor to exploit possible fusion opportunity in the
6196 (altivec_expand_vec_perm_const_le): Likewise.
6198 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6200 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6201 (loadsync_<mode>): Change mode.
6202 (load_quadpti, store_quadpti): New.
6203 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6204 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6206 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6209 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6210 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6211 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6213 2014-04-08 Richard Biener <rguenther@suse.de>
6216 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6217 a 64bit widest int print double-int similar to on HWI64 hosts.
6219 2014-04-08 Richard Biener <rguenther@suse.de>
6221 PR tree-optimization/60785
6222 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6223 default defs properly.
6225 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6227 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6228 (Weffc++): Likewise.
6230 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6232 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6233 set completep to false rather than true.
6235 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6238 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6239 ARM_TARGET2_DWARF_FORMAT.
6241 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6244 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6245 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6248 2014-04-07 Richard Biener <rguenther@suse.de>
6250 PR tree-optimization/60766
6251 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6252 (may_eliminate_iv): Convert cand_value_at result to desired type.
6254 2014-04-07 Jason Merrill <jason@redhat.com>
6257 * common.opt (-fno-gnu-unique): Add.
6258 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6260 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6262 * haifa-sched.c: Fix outdated function reference and minor
6263 grammar errors in introductory comment.
6265 2014-04-07 Richard Biener <rguenther@suse.de>
6268 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6270 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6272 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6275 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6276 size accounting for thunks.
6277 (pa_asm_output_mi_thunk): Use final_start_function() and
6278 final_end_function() to output function start and end directives.
6280 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6282 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6283 device specific ISA/ feature information. Remove short_sp and
6284 errata_skip ds. Add avr_device_specific_features enum to have device
6286 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6287 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6288 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6289 updated device specific info.
6290 * config/avr/avr-mcus.def: Merge device specific details to
6291 dev_attribute field.
6292 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6294 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6295 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6296 assembler if RMW isa supported by current device.
6297 * config/avr/genmultilib.awk: Update as device info structure changed.
6298 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6300 2014-04-04 Cong Hou <congh@google.com>
6302 PR tree-optimization/60656
6303 * tree-vect-stmts.c (supportable_widening_operation):
6304 Fix a bug that elements in a vector with vect_used_by_reduction
6305 property are incorrectly reordered when the operation on it is not
6306 consistant with the one in reduction operation.
6308 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6310 PR rtl-optimization/60155
6311 * gcse.c (record_set_data): New function.
6312 (single_set_gcse): New function.
6313 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6314 (hoist_code): Likewise.
6315 (get_pressure_class_and_nregs): Likewise.
6317 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6319 * explow.c (probe_stack_range): Emit a final optimization blockage.
6321 2014-04-04 Anthony Green <green@moxielogic.com>
6323 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6326 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6329 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6330 flags are set only during streaming.
6331 * ipa.c (process_references, walk_polymorphic_call_targets,
6332 symtab_remove_unreachable_nodes): Drop bodies of always inline
6333 after early inlining.
6334 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6336 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6337 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6340 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6343 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6346 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6348 (cortex_a53_fdivd): Likewise.
6350 2014-04-04 Martin Jambor <mjambor@suse.cz>
6353 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6355 * cgraph.c (clone_of_p): Also return true if thunks match.
6356 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6357 cgraph_function_or_thunk_node and an obsolete comment.
6358 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6360 (build_function_decl_skip_args): Likewise.
6361 (set_new_clone_decl_and_node_flags): New function.
6362 (duplicate_thunk_for_node): Likewise.
6363 (redirect_edge_duplicating_thunks): Likewise.
6364 (cgraph_clone_node): New parameter args_to_skip, pass it to
6365 redirect_edge_duplicating_thunks which is called instead of
6366 cgraph_redirect_edge_callee.
6367 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6368 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6370 2014-04-04 Jeff Law <law@redhat.com>
6373 * config/arm/predicates.md (const_int_I_operand): New predicate.
6374 (const_int_M_operand): Similarly.
6375 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6377 (insv_t2, extv_reg, extzv_t2): Likewise.
6378 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6379 (pop_multiple_with_writeback_and_return): Likewise.
6380 (vfp_pop_multiple_with_writeback): Likewise
6382 2014-04-04 Richard Biener <rguenther@suse.de>
6385 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6386 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6388 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6389 * gimplify.c (gimple_add_tmp_var_fn): New function.
6390 * gimple-expr.h (create_tmp_reg_fn): Declare.
6391 * gimple-expr.c (create_tmp_reg_fn): New function.
6392 * gimple-low.c (record_vars_into): Don't change cfun.
6393 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6394 code generation without cfun.
6396 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6399 * Makefile.in (install-driver): Fix shell scripting.
6401 2014-04-03 Cong Hou <congh@google.com>
6403 PR tree-optimization/60505
6404 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6405 threshold of number of iterations below which no vectorization
6407 * tree-vect-loop.c (new_loop_vec_info):
6408 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6409 * tree-vect-loop.c (vect_analyze_loop_operations):
6410 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6411 * tree-vect-loop.c (vect_transform_loop):
6412 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6413 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6414 of iterations of the loop and see if we should build the epilogue.
6416 2014-04-03 Richard Biener <rguenther@suse.de>
6418 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6419 (streamer_tree_cache_create): Adjust.
6420 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6421 to allow optional nodes array.
6422 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6423 (streamer_tree_cache_append): Likewise.
6424 (streamer_tree_cache_create): Create nodes array optionally
6425 as specified by parameter.
6426 * lto-streamer-out.c (create_output_block): Avoid maintaining
6427 the node array in the writer cache.
6428 (DFS_write_tree): Remove assertion.
6429 (produce_asm_for_decls): Free the out decl state hash table early.
6430 * lto-streamer-in.c (lto_data_in_create): Adjust for
6431 streamer_tree_cache_create prototype change.
6433 2014-04-03 Richard Biener <rguenther@suse.de>
6435 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6436 set TREE_CHAIN to NULL_TREE.
6438 2014-04-03 Richard Biener <rguenther@suse.de>
6440 PR tree-optimization/60740
6441 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6442 over all GIMPLE_COND operands.
6444 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6446 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6447 (Weffc++): Remove Scott's numbering, merge lists and reference
6450 2014-04-03 Nick Clifton <nickc@redhat.com>
6452 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6455 2014-04-03 Martin Jambor <mjambor@suse.cz>
6457 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6458 mention gcc_unreachable before failing.
6459 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6462 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6465 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6466 inconsistent code and instead mark the context inconsistent.
6467 (possible_polymorphic_call_targets): For inconsistent contexts
6468 return empty complete list.
6470 2014-04-02 Anthony Green <green@moxielogic.com>
6472 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6473 (extendqisi2, extendhisi2): Define.
6474 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6475 (WCHAR_TYPE): Change to unsigned int.
6477 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6479 PR tree-optimization/60733
6480 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6481 insertion point for PHI candidates to be the end of the feeding
6482 block for the PHI argument.
6484 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6486 PR rtl-optimization/60650
6487 * lra-constraints.c (process_alt_operands): Decrease reject for
6488 earlyclobber matching.
6490 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6492 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6494 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6496 * config/spu/spu.c (pad_bb): Do not crash when the last
6497 insn is CODE_FOR_blockage.
6499 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6501 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6502 lies outside the target mode.
6504 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6507 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6508 software floating point or no floating point registers, do not
6509 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6510 in GPRs that occurs after we tested for GPRs that would never be
6513 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6514 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6515 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6516 specifically allow DDmode, since that does not use the SPE SIMD
6519 2014-04-02 Richard Biener <rguenther@suse.de>
6522 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6523 MODE_INTs. Properly use negv_optab.
6524 (expand_abs): Likewise.
6526 2014-04-02 Richard Biener <rguenther@suse.de>
6529 * Makefile.in (install-driver): Guard extra installs with special
6532 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6534 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6537 2014-04-01 Richard Henderson <rth@redhat.com>
6540 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6541 alternative enabled before register allocation.
6543 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6545 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6546 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6548 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6549 (nios2_got_address): Update nios2_large_got_address call site.
6550 (nios2_delegitimize_address): New function.
6551 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6552 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6553 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6555 2014-04-01 Martin Husemann <martin@duskware.de>
6557 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6560 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6562 PR rtl-optimization/60604
6563 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6564 check from register_operand.
6565 (register_operand): Redefine in terms of general_operand.
6566 (nonmemory_operand): Use register_operand for the non-constant cases.
6568 2014-04-01 Richard Biener <rguenther@suse.de>
6570 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6572 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6574 * doc/invoke.texi (mapp-regs): Clarify.
6576 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6578 * config/i386/avx512fintrin.h (__v32hi): Define type.
6579 (__v64qi): Likewise.
6580 (_mm512_set1_epi8): Define.
6581 (_mm512_set1_epi16): Define.
6582 (_mm512_set4_epi32): Define.
6583 (_mm512_set4_epi64): Define.
6584 (_mm512_set4_pd): Define.
6585 (_mm512_set4_ps): Define.
6586 (_mm512_setr4_epi64): Define.
6587 (_mm512_setr4_epi32): Define.
6588 (_mm512_setr4_pd): Define.
6589 (_mm512_setr4_ps): Define.
6590 (_mm512_setzero_epi32): Define.
6592 2014-03-31 Martin Jambor <mjambor@suse.cz>
6595 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6596 callsite_arguments_match_p. Updated all callers. Also check types of
6597 corresponding formal parameters and actual arguments.
6598 (not_all_callers_have_enough_arguments_p) Renamed to
6599 some_callers_have_mismatched_arguments_p.
6601 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6603 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6605 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6608 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6611 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6613 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6614 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6616 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6617 Use FMAMODE_NOVF512 mode iterator.
6618 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6619 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6620 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6621 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6623 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6624 Use VF_128_256 mode iterator.
6625 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6628 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6630 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6631 static chain if needed.
6633 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6636 * lra-constraints.c (index_part_to_reg): New.
6637 (process_address): Use it.
6639 2014-03-27 Jeff Law <law@redhat.com>
6640 Jakub Jelinek <jakub@redhat.com>
6643 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6644 gen_rtx_{PLUS,MULT} to build up the address expression.
6646 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6647 creating non-canonical RTL.
6649 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6652 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6653 functions; reorganize to make cheap checks first.
6654 (inline_small_functions): Do not estimate growth when dumping;
6656 * ipa-inline.h (inline_summary): Add min_size.
6657 (growth_likely_positive): New function.
6658 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6659 (set_cond_stmt_execution_predicate): Cleanup.
6660 (estimate_edge_size_and_time): Compute min_size.
6661 (estimate_calls_size_and_time): Likewise.
6662 (estimate_node_size_and_time): Likewise.
6663 (inline_update_overall_summary): Update min_size.
6664 (do_estimate_edge_time): Likewise.
6665 (do_estimate_edge_size): Update.
6666 (do_estimate_edge_hints): Update.
6667 (growth_likely_positive): New function.
6669 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6672 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6673 also if addr has VOIDmode.
6675 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6677 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6678 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6680 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6681 instructions as well as AdvancedSIMD loads.
6683 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6685 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6686 Use crypto_aese type.
6687 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6688 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6689 crypto_aese, crypto_aesmc. Move to types.md.
6690 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6692 * config/arm/iterators.md (crypto_type): Likewise.
6694 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6696 * cgraph.c: Include expr.h and tree-dfa.h.
6697 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6700 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6703 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6704 regs from checking multi-reg pseudos.
6706 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6708 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6710 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6712 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6713 if it would clobber the stack pointer, even temporarily.
6715 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6717 * mode-switching.c: Make small adjustments to the top comment.
6719 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6721 * config/rs6000/constraints.md (wD constraint): New constraint to
6722 match the constant integer to get the top DImode/DFmode out of a
6723 vector in a VSX register.
6725 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6726 match the constant integer to get the top DImode/DFmode out of a
6727 vector in a VSX register.
6729 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6732 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6735 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6736 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6738 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6739 Optimize vec_extract of 64-bit values, where the value being
6740 extracted is in the top word, where we can use scalar
6741 instructions. Add direct move and store support. Combine the big
6742 endian/little endian vector select load support into a single insn.
6743 (vsx_extract_<mode>_internal1): Likewise.
6744 (vsx_extract_<mode>_internal2): Likewise.
6745 (vsx_extract_<mode>_load): Likewise.
6746 (vsx_extract_<mode>_store): Likewise.
6747 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6748 combined into vsx_extract_<mode>_load.
6749 (vsx_extract_<mode>_one_le): Likewise.
6751 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6752 define the top 64-bit vector element.
6754 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6757 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6758 Document vec_vbpermq builtin.
6761 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6762 enable use of xxsldwi and xxpermdi builtin functions.
6763 (vec_xxpermdi): Likewise.
6765 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6766 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6768 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6770 PR rtl-optimization/60650
6771 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6773 (find_spills_for): New.
6774 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6775 Spill all pseudos on the second iteration.
6777 2014-03-27 Marek Polacek <polacek@redhat.com>
6780 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6783 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6785 * config/s390/s390.c (s390_can_use_return_insn): Check for
6786 call-saved FPRs on 31 bit.
6788 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6791 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6792 if they need regimplification, just drop them instead of
6793 calling gimple_regimplify_operands on them.
6795 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6798 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6799 (aarch64_frame_pointer_required): Adjust logic.
6800 (aarch64_can_eliminate): Adjust logic.
6801 (aarch64_override_options_after_change): Adjust logic.
6803 2014-03-27 Dehao Chen <dehao@google.com>
6805 * ipa-inline.c (early_inliner): Update node's inline info.
6807 2014-03-26 Dehao Chen <dehao@google.com>
6809 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6810 compiler inserted conditional jumps for NAN float check.
6812 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6814 * ubsan.h (ubsan_create_data): Change second argument's type
6815 to const location_t *.
6816 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6818 (ubsan_create_data): Change second argument to const location_t *PLOC.
6819 Create Loc field whenever PLOC is non-NULL.
6820 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6821 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6825 * real.c (real_to_integer2): Change type of low to UHWI.
6827 2014-03-26 Tobias Burnus <burnus@net-b.de>
6829 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6830 (CILK_SELF_SPECS): New define.
6831 (driver_self_specs): Use it.
6833 2014-03-26 Richard Biener <rguenther@suse.de>
6835 * tree-pretty-print.c (percent_K_format): Implement special
6836 case for LTO and its stripped down BLOCK tree.
6838 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6841 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6843 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6844 one range is range_int_cst_p, but not both, at least optimize
6845 addition/subtraction of 0 and multiplication by 0 or 1.
6846 * gimple-fold.c (gimple_fold_call): Fold
6847 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6848 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6849 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6851 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6853 PR rtl-optimization/60452
6854 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6855 <case REG>: Return 1 for invalid offsets from the frame pointer.
6857 2014-03-26 Marek Polacek <polacek@redhat.com>
6860 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6863 2014-03-26 Marek Polacek <polacek@redhat.com>
6866 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6869 2014-03-26 Marek Polacek <polacek@redhat.com>
6872 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6873 multiplication in unsigned type.
6875 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6877 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6879 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6881 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6883 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6886 * cif-code.def (UNREACHABLE) New code.
6887 * ipa-inline.c (inline_small_functions): Skip edges to
6888 __builtlin_unreachable.
6889 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6890 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6891 predicate to __bulitin_unreachable.
6892 (set_cond_stmt_execution_predicate): Fix issue when
6893 invert_tree_comparison returns ERROR_MARK.
6894 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6895 propagate to inline clones.
6896 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6898 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6899 * cgraphclones.c (cgraph_clone_node): If call destination is already
6900 ureachable, do not redirect it back.
6901 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6904 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6906 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6907 Do not modify inline clones.
6909 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6911 * config/i386/i386.md (general_sext_operand): New mode attr.
6912 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6913 don't generate (sign_extend (const_int)).
6914 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6915 operands[2]. Use We constraint instead of <i> and
6916 <general_sext_operand> predicate instead of <general_operand>.
6917 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6918 * config/i386/constraints.md (We): New constraint.
6919 * config/i386/predicates.md (x86_64_sext_operand,
6920 sext_operand): New predicates.
6922 2014-03-25 Martin Jambor <mjambor@suse.cz>
6925 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6926 inconsistent devirtualizations to __builtin_unreachable.
6928 2014-03-25 Marek Polacek <polacek@redhat.com>
6931 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6933 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6935 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6936 order of elements for big-endian.
6938 2014-03-25 Richard Biener <rguenther@suse.de>
6941 * gimplify-me.c (gimple_regimplify_operands): Update the
6944 2014-03-25 Martin Jambor <mjambor@suse.cz>
6947 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6948 (lto_output_varpool_node): Likewise.
6949 (input_overwrite_node): Likewise.
6950 (input_varpool_node): Likewise.
6952 2014-03-25 Richard Biener <rguenther@suse.de>
6954 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6955 (run_gcc): Likewise.
6957 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6959 * combine.c (simplify_compare_const): Add MODE argument.
6960 Handle mode_width 0 as very large mode_width.
6961 (try_combine, simplify_comparison): Adjust callers.
6963 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6964 type to avoid signed integer overflow.
6965 * explow.c (plus_constant): Likewise.
6967 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6969 * doc/generic.texi: Correct typos.
6971 2014-03-24 Tobias Burnus <burnus@net-b.de>
6973 * doc/invoke.texi (-flto): Expand section about
6974 using static libraries with LTO.
6976 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6978 PR rtl-optimization/60501
6979 * optabs.def (addptr3_optab): New optab.
6980 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6981 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6982 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6984 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6986 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6988 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6990 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6993 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6994 (_mm256_undefined_ps): Define.
6995 (_mm256_undefined_pd): Define.
6996 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6997 (_mm_undefined_pd): Define.
6998 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6999 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7000 (_mm512_undefined_ps): Define.
7001 (_mm512_undefined_pd): Define.
7002 Use _mm*_undefined_*.
7003 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7005 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7007 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7008 (lshr_simd): DI mode added.
7009 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7010 (aarch64_ushr_simddi): Likewise.
7011 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7012 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7013 (vshrd_n_u64): Likewise.
7015 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7017 * Makefile.in (s-macro_list): Depend on cc1.
7019 2014-03-23 Teresa Johnson <tejohnson@google.com>
7021 * ipa-utils.c (ipa_print_order): Use specified dump file.
7023 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7025 PR rtl-optimization/60601
7026 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7028 * gcc.c (eval_spec_function): Initialize save_growing_value.
7030 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7033 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7034 code == MINUS_EXPR, never swap op0 with op1.
7036 * toplev.c (init_local_tick): Avoid signed integer multiplication
7038 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7039 shift by first operand's bitsize.
7041 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7044 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7046 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7047 TARGET_ISA_64BIT_P(x).
7049 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7051 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7052 pattern for vector nor instead of subtract from splat(-1).
7053 (altivec_expand_vec_perm_const_le): Likewise.
7055 2014-03-21 Richard Henderson <rth@twiddle.net>
7058 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7059 related insns after epilogue_completed.
7061 2014-03-21 Martin Jambor <mjambor@suse.cz>
7064 * cgraph.h (symtab_node): New flag body_removed.
7065 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7066 when removing bodies.
7067 * symtab.c (dump_symtab_base): Dump body_removed flag.
7068 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7069 had their bodies removed.
7071 2014-03-21 Martin Jambor <mjambor@suse.cz>
7074 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7077 2014-03-21 Richard Biener <rguenther@suse.de>
7079 PR tree-optimization/60577
7080 * tree-core.h (struct tree_base): Document nothrow_flag use
7081 in VAR_DECL_NONALIASED.
7082 * tree.h (VAR_DECL_NONALIASED): New.
7083 (may_be_aliased): Adjust.
7084 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7086 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7088 * expr.c (expand_expr_real_1): Remove outdated comment.
7090 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7093 * ira.c (adjust_cleared_regs): Call copy_rtx on
7094 *reg_equiv[REGNO (loc)].src_p before passing it to
7095 simplify_replace_fn_rtx.
7098 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7099 into CONST, put pic register as first operand of PLUS. Use
7100 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7102 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7104 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7106 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7108 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7109 around for store forwarding issue in the FPU on the UT699.
7110 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7111 loads and operations if -mfix-ut699 is specified.
7112 (divtf3_hq): Tweak attribute.
7113 (sqrttf2_hq): Likewise.
7115 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7117 * calls.c (store_one_arg): Remove incorrect const qualification on the
7118 type of the temporary.
7119 * cfgexpand.c (expand_return): Likewise.
7120 * expr.c (expand_constructor): Likewise.
7121 (expand_expr_real_1): Likewise.
7123 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7125 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7128 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7131 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7133 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7135 * config/arm/aarch-common-protos.h
7136 (alu_cost_table): Fix spelling of "extend".
7137 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7139 2014-03-19 Richard Biener <rguenther@suse.de>
7142 * tree-core.h (tree_type_common): Re-order pointer members
7143 to reduce recursion depth during GC walks.
7145 2014-03-19 Marek Polacek <polacek@redhat.com>
7148 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7149 before accessing it.
7151 2014-03-19 Richard Biener <rguenther@suse.de>
7154 * lto-streamer-in.c (input_function): In WPA stage do not drop
7157 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7159 PR tree-optimization/60559
7160 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7161 with build_zero_cst assignment.
7163 2014-03-18 Kai Tietz <ktietz@redhat.com>
7165 PR rtl-optimization/56356
7166 * sdbout.c (sdbout_parms): Verify that parms'
7167 incoming argument is valid.
7168 (sdbout_reg_parms): Likewise.
7170 2014-03-18 Richard Henderson <rth@redhat.com>
7173 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7174 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7175 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7177 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7179 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7180 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7181 Italicize plugin event names in description. Explain that
7182 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7183 Remind that no GCC functions should be called after PLUGIN_FINISH.
7184 Explain what pragmas with expansion are.
7186 2014-03-18 Martin Liska <mliska@suse.cz>
7188 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7189 gimple call statement is update.
7190 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7191 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7193 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7196 * ubsan.c (ubsan_instrument_unreachable): Call
7197 initialize_sanitizer_builtins.
7198 (ubsan_pass): Likewise.
7201 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7202 varpool_finalize_decl instead of rest_of_decl_compilation.
7204 2014-03-18 Richard Biener <rguenther@suse.de>
7206 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7207 by using bitmap_and_compl instead of bitmap_and_compl_into.
7208 (df_rd_transfer_function): Likewise.
7210 2014-03-18 Richard Biener <rguenther@suse.de>
7212 * doc/lto.texi (fresolution): Fix typo.
7214 2014-03-18 Richard Biener <rguenther@suse.de>
7216 * doc/invoke.texi (flto): Update for changes in 4.9.
7218 2014-03-18 Richard Biener <rguenther@suse.de>
7220 * doc/loop.texi: Remove section on the removed lambda framework.
7221 Update loop docs with recent changes in preserving loop structure.
7223 2014-03-18 Richard Biener <rguenther@suse.de>
7225 * doc/lto.texi (-fresolution): Document.
7227 2014-03-18 Richard Biener <rguenther@suse.de>
7229 * doc/contrib.texi: Adjust my name.
7231 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7234 * internal-fn.c: Include diagnostic-core.h.
7235 (expand_BUILTIN_EXPECT): New function.
7236 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7237 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7238 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7239 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7241 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7242 Revert 3 argument __builtin_expect code.
7243 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7244 * gimple-fold.c (gimple_fold_call): Likewise.
7245 * tree.h (fold_builtin_expect): New prototype.
7246 * builtins.c (build_builtin_expect_predicate): Add predictor
7247 argument, if non-NULL, create 3 argument __builtin_expect.
7248 (fold_builtin_expect): No longer static. Add ARG2 argument,
7249 pass it through to build_builtin_expect_predicate.
7250 (fold_builtin_2): Adjust caller.
7251 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7252 * internal-fn.def (BUILTIN_EXPECT): New.
7254 2014-03-18 Tobias Burnus <burnus@net-b.de>
7257 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7258 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7259 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7261 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7264 * predict.c (combine_predictions_for_bb): Fix up formatting.
7265 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7266 fill what it points to if non-NULL.
7267 (tree_predict_by_opcode): Adjust caller, use the predictor.
7268 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7270 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7272 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7273 proper constant for the store mode.
7275 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7277 * symtab.c (change_decl_assembler_name): Fix transparent alias
7280 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7282 * config/aarch64/aarch64.c: Correct the comments about the
7283 aarch64 stack layout.
7285 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7287 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7288 check for GF_OMP_FOR_KIND_FOR.
7290 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7292 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7293 ymm and zmm register names.
7295 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7298 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7299 note creation for the 2010-08-31 changes.
7301 2014-03-17 Marek Polacek <polacek@redhat.com>
7304 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7305 as -fno-tree-loop-vectorize.
7306 (expand_omp_simd): Likewise.
7308 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7310 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7311 (eligible_for_call_delay): New prototype.
7312 * config/sparc/sparc.c (tls_call_delay): Rename into...
7313 (eligible_for_call_delay): ...this. Return false if the instruction
7314 cannot be put in the delay slot of a branch.
7315 (eligible_for_restore_insn): Simplify.
7316 (eligible_for_return_delay): Return false if the instruction cannot be
7317 put in the delay slot of a branch and simplify.
7318 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7319 put in the delay slot of a branch.
7320 * config/sparc/sparc.md (fix_ut699): New attribute.
7321 (tls_call_delay): Delete.
7322 (in_call_delay): Reimplement.
7323 (eligible_for_sibcall_delay): Rename into...
7324 (in_sibcall_delay): ...this.
7325 (eligible_for_return_delay): Rename into...
7326 (in_return_delay): ...this.
7327 (in_branch_delay): Reimplement.
7328 (in_uncond_branch_delay): Delete.
7329 (in_annul_branch_delay): Delete.
7331 2014-03-14 Richard Henderson <rth@redhat.com>
7334 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7335 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7336 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7337 (floathi splitters): Remove.
7338 (float<SWI48x>xf2): New pattern.
7339 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7340 code that tried to handle DImode for 32-bit, but which was excluded
7341 by the pattern's condition. Drop allocation of stack temporary.
7342 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7343 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7344 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7345 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7346 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7347 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7348 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7349 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7350 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7351 (*float<SWI48x><X87MODEF>2_i387): Remove.
7352 (all float _with_temp splitters): Remove.
7353 (*float<SWI48x><MODEF>2_i387): New pattern.
7354 (*float<SWI48><MODEF>2_sse): New pattern.
7355 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7356 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7358 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7359 Marek Polacek <polacek@redhat.com>
7362 * common.opt (dump_base_name_prefixed): New Variable.
7363 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7364 if x_dump_base_name_prefixed is already set, set it at the end.
7366 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7368 PR rtl-optimization/60508
7369 * lra-constraints.c (get_reload_reg): Add new parameter
7371 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7372 Pass the new parameter values.
7374 2014-03-14 Richard Biener <rguenther@suse.de>
7376 * common.opt: Revert unintented changes from r205065.
7379 2014-03-14 Richard Biener <rguenther@suse.de>
7382 * cfghooks.c (split_block): Properly adjust all loops the
7383 block was a latch of.
7385 2014-03-14 Martin Jambor <mjambor@suse.cz>
7388 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7391 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7394 * config/avr/avr.c (avr_set_current_function): Pass function name
7395 through default_strip_name_encoding before sanity checking instead
7396 of skipping the first char of the assembler name.
7398 2014-03-13 Richard Henderson <rth@redhat.com>
7401 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7402 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7403 * config/i386/i386-protos.h: Likewise.
7404 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7405 in the expander instead of a splitter.
7406 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7407 any possibility of requiring a memory.
7408 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7409 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7410 (fp branch splitters): Update for ix86_split_fp_branch.
7411 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7412 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7413 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7414 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7415 (*fop_<MODEF>_3_i387): Likewise.
7416 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7417 (splitters for the fop_* register patterns): Remove.
7418 (fscalexf4_i387): Rename from *fscalexf4_i387.
7419 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7421 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7423 PR tree-optimization/59779
7424 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7425 type for bitsize and maxsize instead of HOST_WIDE_INT.
7427 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7429 PR rtl-optimization/57320
7430 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7431 the CFG after thread_prologue_and_epilogue_insns.
7433 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7435 PR rtl-optimization/57189
7436 * lra-constraints.c (process_alt_operands): Disfavor spilling
7439 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7441 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7443 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7445 PR tree-optimization/59025
7447 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7448 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7450 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7453 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7454 calls of avr_out_plus_1.
7456 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7458 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7459 BB's single pred and update the father loop's latch info later.
7461 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7463 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7467 (VEC_base): Likewise.
7468 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7469 registers, we need to swap double words in little endian mode.
7471 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7472 to be a container mode for 128-bit integer operations added in ISA
7473 2.07. Unlike TImode and PTImode, the preferred register set is
7474 the Altivec/VMX registers for the 128-bit operations.
7476 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7478 (rs6000_split_128bit_ok_p): Likewise.
7480 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7481 macros for creating ISA 2.07 normal and overloaded builtin
7482 functions with 3 arguments.
7483 (BU_P8V_OVERLOAD_3): Likewise.
7484 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7485 for use as overloaded functions.
7486 (VPERM_1TI_UNS): Likewise.
7487 (VSEL_1TI): Likewise.
7488 (VSEL_1TI_UNS): Likewise.
7489 (ST_INTERNAL_1ti): Likewise.
7490 (LD_INTERNAL_1ti): Likewise.
7491 (XXSEL_1TI): Likewise.
7492 (XXSEL_1TI_UNS): Likewise.
7493 (VPERM_1TI): Likewise.
7494 (VPERM_1TI_UNS): Likewise.
7495 (XXPERMDI_1TI): Likewise.
7496 (SET_1TI): Likewise.
7497 (LXVD2X_V1TI): Likewise.
7498 (STXVD2X_V1TI): Likewise.
7499 (VEC_INIT_V1TI): Likewise.
7500 (VEC_SET_V1TI): Likewise.
7501 (VEC_EXT_V1TI): Likewise.
7502 (EQV_V1TI): Likewise.
7503 (NAND_V1TI): Likewise.
7504 (ORC_V1TI): Likewise.
7505 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7506 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7508 (VADDUQM): Likewise.
7509 (VSUBCUQ): Likewise.
7510 (VADDEUQM): Likewise.
7511 (VADDECUQ): Likewise.
7512 (VSUBEUQM): Likewise.
7513 (VSUBECUQ): Likewise.
7515 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7516 __int128_t and __uint128_t types.
7517 (__uint128_type): Likewise.
7518 (altivec_categorize_keyword): Add support for vector __int128_t,
7519 vector __uint128_t, vector __int128, and vector unsigned __int128
7520 as a container type for TImode operations that need to be done in
7521 VSX/Altivec registers.
7522 (rs6000_macro_to_expand): Likewise.
7523 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7524 to support 128-bit integer instructions vaddcuq, vadduqm,
7525 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7526 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7528 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7529 for V1TImode, and set up preferences to use VSX/Altivec registers.
7530 Setup VSX reload handlers.
7531 (rs6000_debug_reg_global): Likewise.
7532 (rs6000_init_hard_regno_mode_ok): Likewise.
7533 (rs6000_preferred_simd_mode): Likewise.
7534 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7535 (easy_altivec_constant): Likewise.
7536 (output_vec_const_move): Likewise.
7537 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7539 (rs6000_expand_vector_extract): Likewise.
7540 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7542 (rs6000_const_vec): Add support for V1TImode.
7543 (rs6000_emit_le_vsx_load): Swap double words when loading or
7544 storing TImode/V1TImode.
7545 (rs6000_emit_le_vsx_store): Likewise.
7546 (rs6000_emit_le_vsx_move): Likewise.
7547 (rs6000_emit_move): Add support for V1TImode.
7548 (altivec_expand_ld_builtin): Likewise.
7549 (altivec_expand_st_builtin): Likewise.
7550 (altivec_expand_vec_init_builtin): Likewise.
7551 (altivec_expand_builtin): Likewise.
7552 (rs6000_init_builtins): Add support for V1TImode type. Add
7553 support for ISA 2.07 128-bit integer builtins. Define type names
7554 for the VSX/Altivec vector types.
7555 (altivec_init_builtins): Add support for overloaded vector
7556 functions with V1TImode type.
7557 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7558 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7560 (rs6000_split_128bit_ok_p): Likewise.
7561 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7562 __int128_t and vector __uint128_t.
7564 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7565 and mode attributes.
7572 (VS_scalar): Likewise.
7573 (VS_double): Likewise.
7574 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7576 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7577 we support the ISA 2.07 128-bit integer arithmetic instructions.
7578 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7579 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7580 and TImode types for use with the builtin functions.
7581 (V1TI_type_node): Likewise.
7582 (unsigned_V1TI_type_node): Likewise.
7583 (intTI_type_internal_node): Likewise.
7584 (uintTI_type_internal_node): Likewise.
7586 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7587 128-bit builtin functions.
7588 (UNSPEC_VADDEUQM): Likewise.
7589 (UNSPEC_VADDECUQ): Likewise.
7590 (UNSPEC_VSUBCUQ): Likewise.
7591 (UNSPEC_VSUBEUQM): Likewise.
7592 (UNSPEC_VSUBECUQ): Likewise.
7593 (VM): Add V1TImode to vector mode iterators.
7595 (VI_unit): Likewise.
7596 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7597 (altivec_vaddcuq): Likewise.
7598 (altivec_vsubuqm): Likewise.
7599 (altivec_vsubcuq): Likewise.
7600 (altivec_vaddeuqm): Likewise.
7601 (altivec_vaddecuq): Likewise.
7602 (altivec_vsubeuqm): Likewise.
7603 (altivec_vsubecuq): Likewise.
7605 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7607 (BOOL_128): Likewise.
7608 (BOOL_REGS_OUTPUT): Likewise.
7609 (BOOL_REGS_OP1): Likewise.
7610 (BOOL_REGS_OP2): Likewise.
7611 (BOOL_REGS_UNARY): Likewise.
7612 (BOOL_REGS_AND_CR0): Likewise.
7614 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7615 128-bit integer builtin support.
7616 (vec_vadduqm): Likewise.
7617 (vec_vaddecuq): Likewise.
7618 (vec_vaddeuqm): Likewise.
7619 (vec_vsubecuq): Likewise.
7620 (vec_vsubeuqm): Likewise.
7621 (vec_vsubcuq): Likewise.
7622 (vec_vsubuqm): Likewise.
7624 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7625 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7626 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7627 128-bit integer add/subtract to ISA 2.07.
7629 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7631 * config/arc/arc.c (arc_predicate_delay_insns):
7632 Fix third argument passed to conditionalize_nonjump.
7634 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7636 * config/aarch64/aarch64-builtins.c
7637 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7638 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7639 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7640 instead of __builtin_lfloor.
7641 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7643 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7645 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7646 (tree_ssa_ifcombine_bb_1): New function.
7647 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7648 is an empty forwarder block to then_bb or vice versa and then_bb
7649 and else_bb are effectively swapped.
7651 2014-03-12 Christian Bruel <christian.bruel@st.com>
7654 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7655 REG_CFA_DEF_CFA note.
7656 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7657 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7659 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7661 PR tree-optimization/60454
7662 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7664 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7666 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7667 Do not define target_cpu_default2 to generic.
7668 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7669 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7670 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7672 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7673 Marc Glisse <marc.glisse@inria.fr>
7675 PR tree-optimization/60502
7676 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7677 instead of build_low_bits_mask.
7679 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7682 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7683 if there are multiple uses, but op doesn't live on E edge.
7684 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7685 clobber stmts before __builtin_unreachable.
7687 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7689 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7690 hard_frame_pointer_rtx.
7691 * cse.c (cse_insn): Remove volatile check.
7692 * cselib.c (cselib_process_insn): Likewise.
7693 * dse.c (scan_insn): Likewise.
7695 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7697 * config/arc/arc.c (conditionalize_nonjump): New function,
7699 (arc_ifcvt): ... this.
7700 (arc_predicate_delay_insns): Use it.
7702 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7704 * config/arc/predicates.md (extend_operand): During/after reload,
7705 allow const_int_operand.
7706 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7707 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7708 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7710 (umulsi3_highpart_i): Likewise.
7712 2014-03-11 Richard Biener <rguenther@suse.de>
7714 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7715 Add asserts to guard possible wrong-code bugs.
7717 2014-03-11 Richard Biener <rguenther@suse.de>
7719 PR tree-optimization/60429
7720 PR tree-optimization/60485
7721 * tree-ssa-structalias.c (set_union_with_increment): Properly
7722 take into account all fields that overlap the shifted vars.
7723 (do_sd_constraint): Likewise.
7724 (do_ds_constraint): Likewise.
7725 (get_constraint_for_ptr_offset): Likewise.
7727 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7729 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7730 (nios2_compute_frame_layout):
7731 Add calculation of cfun->machine->fp_save_offset.
7732 (nios2_expand_prologue): Correct setting of frame pointer register
7734 (nios2_expand_epilogue): Update recovery of stack pointer from
7735 frame pointer accordingly.
7736 (nios2_initial_elimination_offset): Update calculation of offset
7737 for eliminating to HARD_FRAME_POINTER_REGNUM.
7739 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7742 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7743 cgraph_get_create_node on VAR_DECLs.
7745 2014-03-10 Richard Biener <rguenther@suse.de>
7748 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7750 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7752 * config/vms/vms.opt (vms_float_format): New variable.
7754 2014-03-08 Tobias Burnus <burnus@net-b.de>
7756 * doc/invoke.texi (-fcilkplus): Update implementation status.
7758 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7759 Richard Biener <rguenther@suse.de>
7761 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7762 consistently accross all TUs.
7763 (run_gcc): Enable -fshort-double automatically at link at link-time
7764 and disallow override.
7766 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7769 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7770 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7771 if they can't be used.
7773 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7775 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7776 for Solaris 11/x86 ld.
7777 * configure: Regenerate.
7779 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7781 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7782 (LIB_TLS_SPEC): Save as ld_tls_libs.
7783 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7784 (HAVE_AS_IX86_TLSLDM): New test.
7785 * configure, config.in: Regenerate.
7786 * config/i386/i386.c (legitimize_tls_address): Fall back to
7787 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7788 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7789 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7790 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7792 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7794 * common.opt (fira-loop-pressure): Mark as optimization.
7796 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7798 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7799 an OpenMP mappable type.
7801 2014-03-06 Matthias Klose <doko@ubuntu.com>
7803 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7804 MULTILIB_OSDIRNAMES is not defined.
7806 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7807 Meador Inge <meadori@codesourcery.com>
7810 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7811 (arm_legitimize_address): Call legitimize_tls_address for any
7812 arm_tls_referenced_p expression, handle constant addend. Call it
7813 before testing for !TARGET_ARM.
7814 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7816 2014-03-06 Richard Biener <rguenther@suse.de>
7822 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7824 * tree-streamer.c (record_common_node): Assert we don't record
7825 nodes with type double.
7826 (preload_common_node): Skip type double, complex double and double
7827 pointer since it is now frontend dependent due to fshort-double option.
7829 2014-03-06 Richard Biener <rguenther@suse.de>
7831 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7832 or -fno-lto is specified and the linker has full plugin support.
7833 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7834 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7835 * lto-wrapper.c (merge_and_complain): Merge compile-time
7836 optimization levels.
7837 (run_gcc): And pass it through to the link options.
7839 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7843 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7845 * cselib.c (remove_useless_values): Skip to avoid quadratic
7846 behavior if the condition moved from...
7847 (cselib_process_insn): ... here holds.
7849 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7852 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7853 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7856 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7857 (TM_H): Add x86-tune.def.
7859 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7861 * config/aarch64/aarch64.c (generic_tunings):
7862 Use cortexa57_extra_costs.
7864 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7867 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7868 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7869 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7872 2014-03-04 Heiher <r@hev.cc>
7874 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7875 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7877 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7879 * config/i386/predicates.md (const2356_operand): Change to ...
7880 (const2367_operand): ... this.
7881 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7883 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7884 (*avx512pf_scatterpf<mode>sf): Ditto.
7885 (avx512pf_scatterpf<mode>df): Ditto.
7886 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7887 (*avx512pf_scatterpf<mode>df): Ditto.
7888 * config/i386/i386.c (ix86_expand_builtin): Update
7889 incorrect hint operand error message.
7891 2014-03-04 Richard Biener <rguenther@suse.de>
7893 * lto-section-in.c (lto_get_section_data): Fix const cast.
7895 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7897 * tree-streamer.c (record_common_node): Assert we don't record
7898 nodes with type double.
7899 (preload_common_node): Skip type double, complex double and double
7900 pointer since it is now frontend dependent due to fshort-double option.
7902 2014-03-04 Richard Biener <rguenther@suse.de>
7905 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7906 (lto_input_toplevel_asms): Likewise.
7907 * lto-section-in.c (lto_get_section_data): Instead do it here
7910 2014-03-04 Richard Biener <rguenther@suse.de>
7912 PR tree-optimization/60382
7913 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7914 dead PHIs a reduction.
7916 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7918 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7920 (_mm_prefetch): Move out of GCC target("sse") pragma.
7921 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7922 GCC target("prfchw") pragma.
7923 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7925 * config/i386/i386.c (ix86_option_override_internal): Enable
7926 -mprfchw with -mprefetchwt1.
7928 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7930 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7933 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7935 * opts.h (CL_PCH_IGNORE): Define.
7936 * targhooks.c (option_affects_pch_p):
7937 Return false for options that have CL_PCH_IGNORE set.
7938 * opt-functions.awk: Process PchIgnore.
7939 * doc/options.texi: Document PchIgnore.
7941 * config/arc/arc.opt (misize): Add PchIgnore property.
7943 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7945 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7946 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7947 constraint on constants to permit them being loaded into
7948 GENERAL_REGS or BASE_REGS.
7950 2014-03-03 Nick Clifton <nickc@redhat.com>
7952 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7953 anti-cacnonical alternatives.
7954 (negandhi3_real): New pattern.
7955 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7957 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7959 * config/avr/avr-mcus.def: Remove atxmega16x1.
7960 * config/avr/avr-tables.opt: Regenerate.
7961 * config/avr/t-multilib: Regenerate.
7962 * doc/avr-mmcu.texi: Regenerate.
7964 2014-03-03 Tobias Grosser <tobias@grosser.es>
7965 Mircea Namolaru <mircea.namolaru@inria.fr>
7967 PR tree-optimization/58028
7968 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7971 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7973 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7974 not handled by recognizers.
7976 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7979 * function.c (expand_function_end): Don't emit
7980 clobber_return_register sequence if clobber_after is a BARRIER.
7981 * cfgexpand.c (construct_exit_block): Append instructions before
7982 return_label to prev_bb.
7984 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7986 * config/rs6000/constraints.md: Document reserved use of "wc".
7988 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7991 * ipa.c (function_and_variable_visibility): When dissolving comdat
7992 group, also set all symbols to local.
7994 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7999 2013-12-14 Jan Hubicka <jh@suse.cz>
8001 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8003 2014-03-02 Jon Beniston <jon@beniston.com>
8009 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8010 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8011 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8012 (simple_return, *simple_return): New patterns
8013 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8014 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8016 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8018 * dwarf2out.c (gen_subprogram_die): Tidy.
8020 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8023 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8024 (*mov_t_msb_neg_negc): ... this new insn.
8026 2014-02-28 Jason Merrill <jason@redhat.com>
8029 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8032 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8035 * dwarf2out.c (decltype_auto_die): New static.
8036 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8037 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8038 (is_cxx_auto): Likewise.
8040 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8042 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8043 we are not using general regs only.
8045 2014-02-28 Richard Biener <rguenther@suse.de>
8048 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8049 previous fix and only allow to remove trivial pre-headers
8050 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8051 (remove_forwarder_block): Properly update the latch of a loop.
8053 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8056 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8057 (cselib_preserved_hash_table): New.
8058 (preserve_constants_and_equivs): Move preserved vals to it.
8059 (cselib_find_slot): Look it up first.
8060 (cselib_init): Initialize it.
8061 (cselib_finish): Release it.
8062 (dump_cselib_table): Dump it.
8064 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8067 * cselib.c (remove_useless_values): Skip to avoid quadratic
8068 behavior if the condition moved from...
8069 (cselib_process_insn): ... here holds.
8071 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8074 * var-tracking.c (vt_initialize): Apply the same condition to
8075 preserve the CFA base value.
8077 2014-02-28 Joey Ye <joey.ye@arm.com>
8080 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8081 if reload in progress or completed.
8083 2014-02-28 Tobias Burnus <burnus@net-b.de>
8086 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8089 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8091 * doc/tm.texi.in (Condition Code Status): Update documention for
8092 relative locations of cc0-setter and cc0-user.
8094 2014-02-27 Jeff Law <law@redhat.com>
8096 PR rtl-optimization/52714
8097 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8098 into two independent simple sets, if I3 is a jump, ensure the
8099 pattern we place into I3 is a (set (pc) ...).
8101 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8102 Jeff Law <law@redhat.com>
8104 PR rtl-optimization/49847
8105 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8106 are in different blocks.
8107 * doc/tm.texi (Condition Code Status): Update documention for
8108 relative locations of cc0-setter and cc0-user.
8110 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8113 * lra.c (lra_emit_add): Check SUBREG too.
8115 2014-02-27 Andreas Schwab <schwab@suse.de>
8117 * config/m68k/m68k.c (m68k_option_override): Disable
8118 -flive-range-shrinkage for classic m68k.
8119 (m68k_override_options_after_change): Likewise.
8121 2014-02-27 Marek Polacek <polacek@redhat.com>
8124 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8125 -Wmaybe-uninitialized.
8127 2014-02-27 Alan Modra <amodra@gmail.com>
8130 * reload1.c (emit_input_reload_insns): When reload_override_in,
8131 set old to rl->in_reg when rl->in_reg is a subreg.
8133 2014-02-26 Richard Biener <rguenther@suse.de>
8136 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8138 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8140 * common/config/i386/predicates.md (const1256_operand): Remove.
8141 (const2356_operand): New.
8142 (const_1_to_2_operand): Remove.
8143 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8144 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8145 (*avx512pf_gatherpf<mode>sf): Ditto.
8146 (avx512pf_gatherpf<mode>df): Ditto.
8147 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8148 (*avx512pf_gatherpf<mode>df): Ditto.
8149 (avx512pf_scatterpf<mode>sf): Ditto.
8150 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8151 (*avx512pf_scatterpf<mode>sf): Ditto.
8152 (avx512pf_scatterpf<mode>df): Ditto.
8153 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8154 (*avx512pf_scatterpf<mode>df): Ditto.
8155 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8157 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8159 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8160 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8161 (_mm512_mask_testn_epi64_mask): Move to ...
8162 * config/i386/avx512cdintrin.h: Here.
8163 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8164 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8165 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8166 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8167 TARGET_AVX512F from TARGET_AVX512CD.
8169 2014-02-26 Richard Biener <rguenther@suse.de>
8172 * ipa.c (walk_polymorphic_call_targets): Properly guard
8173 call to inline_update_overall_summary.
8175 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8178 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8179 and latches only if requested. Fix latch if it is removed.
8180 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8181 LOOPS_HAVE_PREHEADERS.
8183 2014-02-25 Andrew Pinski <apinski@cavium.com>
8185 * builtins.c (expand_builtin_thread_pointer): Create a new target
8186 when the target is NULL.
8188 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8190 PR rtl-optimization/60317
8191 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8192 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8193 * lra-assigns.c: Include params.h.
8194 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8195 other reload pseudos considerations.
8197 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8199 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8200 to use canonical form for nor<mode>3.
8202 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8205 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8208 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8210 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8211 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8212 (ix86_handle_option): Handle OPT_mprefetchwt1.
8213 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8214 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8216 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8217 OPTION_MASK_ISA_PREFETCHWT1.
8218 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8219 (PTA_PREFETCHWT1): New.
8220 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8221 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8222 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8223 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8224 (*prefetch_avx512pf_<mode>_: Change into ...
8225 (*prefetch_prefetchwt1_<mode>: This.
8226 * config/i386/i386.opt (mprefetchwt1): New.
8227 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8228 (_mm_prefetch): Handle intent to write.
8229 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8231 2014-02-25 Richard Biener <rguenther@suse.de>
8234 * emit-rtl.c (mem_attrs_htab): Remove.
8235 (mem_attrs_htab_hash): Likewise.
8236 (mem_attrs_htab_eq): Likewise.
8237 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8238 (init_emit_once): Do not allocate mem_attrs_htab.
8240 2014-02-25 Richard Biener <rguenther@suse.de>
8243 * lto-opts.c (lto_write_options): Output non-explicit conservative
8244 -fwrapv, -fno-trapv and -fno-strict-overflow.
8245 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8246 (run_gcc): And pass them through.
8248 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8250 * sel-sched.c (calculate_new_fences): New parameter ptime.
8251 Calculate it as a maximum over all fence cycles.
8252 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8253 Print the final schedule timing when sched_verbose.
8255 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8257 PR rtl-optimization/60292
8258 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8259 bit fot the fence instruction.
8261 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8263 * calls.h: Fix typo in comment.
8265 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8267 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8268 adjusting offsetable addresses.
8270 2014-02-24 Guozhi Wei <carrot@google.com>
8272 * sparseset.h (sparseset_pop): Fix the wrong index.
8274 2014-02-24 Walter Lee <walt@tilera.com>
8276 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8277 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8279 * common/config/tilegx/tilegx-common.c
8280 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8281 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8283 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8284 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8285 (tilegx_gimplify_va_arg_expr): Handle big endian.
8286 (tilegx_expand_unaligned_load): Ditto.
8287 (tilegx_expand_unaligned_store): Ditto.
8288 (TARGET_RETURN_IN_MSB): New.
8289 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8290 (TARGET_ENDIAN_DEFAULT): New.
8291 (TARGET_BIG_ENDIAN): Handle big endian.
8292 (BYTES_BIG_ENDIAN): Ditto.
8293 (WORDS_BIG_ENDIAN): Ditto.
8294 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8297 * config/tilegx/tilegx.md (extv): Handle big endian.
8299 (insn_st<n>): Ditto.
8300 (insn_st<n>_add<bitsuffix>): Ditto.
8301 (insn_stnt<n>): Ditto.
8302 (insn_stnt<n>_add<bitsuffix>):Ditto.
8303 (vec_interleave_highv8qi): Handle big endian.
8304 (vec_interleave_highv8qi_be): New.
8305 (vec_interleave_highv8qi_le): New.
8306 (insn_v1int_h): Handle big endian.
8307 (vec_interleave_lowv8qi): Handle big endian.
8308 (vec_interleave_lowv8qi_be): New.
8309 (vec_interleave_lowv8qi_le): New.
8310 (insn_v1int_l): Handle big endian.
8311 (vec_interleave_highv4hi): Handle big endian.
8312 (vec_interleave_highv4hi_be): New.
8313 (vec_interleave_highv4hi_le): New.
8314 (insn_v2int_h): Handle big endian.
8315 (vec_interleave_lowv4hi): Handle big endian.
8316 (vec_interleave_lowv4hi_be): New.
8317 (vec_interleave_lowv4hi_le): New.
8318 (insn_v2int_l): Handle big endian.
8319 (vec_interleave_highv2si): Handle big endian.
8320 (vec_interleave_highv2si_be): New.
8321 (vec_interleave_highv2si_le): New.
8322 (insn_v4int_h): Handle big endian.
8323 (vec_interleave_lowv2si): Handle big endian.
8324 (vec_interleave_lowv2si_be): New.
8325 (vec_interleave_lowv2si_le): New.
8326 (insn_v4int_l): Handle big endian.
8327 * config/tilegx/tilegx.opt (mbig-endian): New option.
8328 (mlittle-endian): New option.
8329 * doc/install.texi: Document tilegxbe-linux.
8330 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8332 2014-02-24 Martin Jambor <mjambor@suse.cz>
8335 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8336 there are no parameter descriptors.
8338 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8340 PR rtl-optimization/60268
8341 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8342 initialization to ...
8343 (sched_rgn_init): ... here.
8344 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8346 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8348 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8351 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8353 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8356 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8358 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8359 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8361 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8363 * config/microblaze/predicates.md: Add cmp_op predicate.
8364 * config/microblaze/microblaze.md: Add branch_compare instruction
8365 which uses cmp_op predicate and emits cmp insn before branch.
8366 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8367 to microblaze_expand_conditional_branch and consolidate logic.
8368 (microblaze_expand_conditional_branch): emit branch_compare
8369 insn instead of handling cmp op separate from branch insn.
8371 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8373 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8376 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8378 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8379 define_insn with define_expand and new define_insn
8380 *altivec_lve<VI_char>x_internal.
8381 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8382 and new define_insn *altivec_stve<VI_char>x_internal.
8383 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8385 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8387 (altivec_expand_stvex_be): New function.
8389 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8391 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8392 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8393 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8394 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8396 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8399 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8400 instead of emit_move_insn.
8402 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8404 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8406 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8407 gen_altivec_vsumsws.
8409 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8411 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8412 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8413 (altivec_lvxl_<mode>): New define_expand incorporating
8414 -maltivec=be semantics where needed.
8415 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8416 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8417 semantics where needed.
8418 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8419 (altivec_stvx_<mode>): New define_expand incorporating
8420 -maltivec=be semantics where needed.
8421 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8422 VM2 iterator instead of V4SI.
8423 (altivec_stvxl_<mode>): New define_expand incorporating
8424 -maltivec=be semantics where needed.
8425 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8426 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8427 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8428 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8429 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8430 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8431 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8432 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8433 ALTIVEC_BUILTIN_STVXL.
8434 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8435 (altivec_expand_stvx_be): Likewise.
8436 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8437 (altivec_expand_lvx_be): Likewise.
8438 (altivec_expand_stvx_be): Likewise.
8439 (altivec_expand_builtin): Add cases for
8440 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8441 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8442 (altivec_init_builtins): Add definitions for
8443 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8444 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8446 2014-02-21 Catherine Moore <clm@codesourcery.com>
8448 * doc/invoke.texi (mvirt, mno-virt): Document.
8449 * config/mips/mips.opt (mvirt): New option.
8450 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8452 2014-02-21 Richard Biener <rguenther@suse.de>
8454 PR tree-optimization/60276
8455 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8456 (STMT_VINFO_MIN_NEG_DIST): New macro.
8457 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8458 STMT_VINFO_MIN_NEG_DIST.
8459 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8460 made for negative dependence distances still hold.
8462 2014-02-21 Richard Biener <rguenther@suse.de>
8465 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8466 DECL_INITIAL for globals not in the current function context.
8468 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8470 PR tree-optimization/56490
8471 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8472 * tree-ssa-uninit.c: Include params.h.
8473 (compute_control_dep_chain): Add num_calls argument, return false
8474 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8475 num_calls to recursive call.
8476 (find_predicates): Change dep_chain into normal array,
8477 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8478 variable and adjust compute_control_dep_chain caller.
8479 (find_def_preds): Likewise.
8481 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8483 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8484 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8486 2014-02-21 Nick Clifton <nickc@redhat.com>
8488 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8489 (pushhi1): Likewise.
8490 (popqi1): Add mode to pre_dec.
8493 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8495 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8496 mode for mask of V8SFmode permutation.
8498 2014-02-20 Richard Henderson <rth@redhat.com>
8501 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8502 a new pseudo for OLDVAL.
8504 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8507 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8508 gen_reg_rtx if d->testing_p.
8509 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8510 if d->testing_p and we will certainly return true.
8511 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8514 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8516 * emit-rtl.c (gen_reg_rtx): Assert that
8517 crtl->emit.regno_pointer_align_length is non-zero.
8519 2014-02-20 Richard Henderson <rth@redhat.com>
8522 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8523 on failure the store back into EXPECT.
8525 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8526 Sandra Loosemore <sandra@codesourcery.com>
8528 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8529 * config/nios2/nios2.c (nios2_function_profiler): Add
8530 -fPIC (flag_pic == 2) support.
8531 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8532 (nios2_large_offset_p): New function.
8533 (nios2_unspec_reloc_p): Move up position, update to use
8534 nios2_large_offset_p.
8535 (nios2_unspec_address): Remove function.
8536 (nios2_unspec_offset): New function.
8537 (nios2_large_got_address): New function.
8538 (nios2_got_address): Add large offset support.
8539 (nios2_legitimize_tls_address): Update usage of removed and new
8541 (nios2_symbol_binds_local_p): New function.
8542 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8543 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8544 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8545 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8546 processing for (const (unspec ...)).
8547 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8549 2014-02-20 Richard Biener <rguenther@suse.de>
8551 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8552 doing the substitution.
8553 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8555 2014-02-20 Martin Jambor <mjambor@suse.cz>
8558 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8559 info when checking whether lattices are bottom.
8561 2014-02-20 Richard Biener <rguenther@suse.de>
8564 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8567 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8570 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8571 parameter specifying the scaling.
8572 (inline_call): Update.
8573 (want_inline_recursively): Guard division by zero.
8574 (recursive_inlining): Update.
8575 * ipa-inline.h (clone_inlined_nodes): Update.
8577 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8580 * config/i386/i386.c (classify_argument): Pass structures of size
8581 64 bytes or less in register.
8583 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8584 Kirill Yukhin <kirill.yukhin@intel.com>
8586 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8587 (_mm_rcp28_round_ss): Ditto.
8588 (_mm_rsqrt28_round_sd): Ditto.
8589 (_mm_rsqrt28_round_ss): Ditto.
8590 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8591 (_mm_rcp14_round_ss): Ditto.
8592 (_mm_rsqrt14_round_sd): Ditto.
8593 (_mm_rsqrt14_round_ss): Ditto.
8594 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8595 the first input operand, get rid of match_dup.
8596 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8598 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8600 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8601 operand as the first input operand, set type attribute.
8602 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8604 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8605 operand as the first input operand, set type attribute.
8607 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8609 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8612 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8615 * config/i386/i386.c (construct_container): Remove TFmode check
8616 for X86_64_INTEGER_CLASS.
8618 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8621 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8622 only when -Wpsabi is enabled.
8624 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8627 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8628 passing arrays in registers are the same as for structs, so remove the
8629 special case for them.
8631 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8633 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8634 destination type, extract only the valid bits if the source type is not
8635 integral and has a different mode.
8637 2014-02-19 Richard Biener <rguenther@suse.de>
8640 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8643 2014-02-19 Richard Biener <rguenther@suse.de>
8646 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8647 (ipa_modify_call_arguments): Emit an argument load explicitely and
8648 preserve virtual SSA form there and for the replacement call.
8649 Do not update SSA form nor free dominance info.
8651 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8653 * ipa.c (function_and_variable_visibility): Also clear WEAK
8654 flag when disolving COMDAT_GROUP.
8656 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8658 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8659 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8661 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8662 do more sanity checks.
8663 (detect_type_change): Return true when giving up early.
8664 (compute_complex_assign_jump_func): Fix type parameter of
8665 ipa_set_ancestor_jf.
8666 (compute_complex_ancestor_jump_func): Likewise.
8667 (update_jump_functions_after_inlining): Fix updating of
8669 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8671 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8673 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8674 inline clones when edge disappears.
8676 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8679 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8680 Split 64-bit moves into 2 patterns. Do not allow the use of
8681 direct move for TDmode in little endian, since the decimal value
8682 has little endian bytes within a word, but the 64-bit pieces are
8683 ordered in a big endian fashion, and normal subreg's of TDmode are
8685 (mov<mode>_64bit_dm): Likewise.
8686 (movtd_64bit_nodm): Likewise.
8688 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8690 PR tree-optimization/60174
8691 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8692 statement of an SSA_NAME that occurs in an abnormal PHI node.
8694 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8697 * final.c (SEEN_BB): Remove.
8698 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8699 (final_scan_insn): Don't force_source_line on second
8700 NOTE_INSN_BASIC_BLOCK.
8702 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8705 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8706 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8707 (type_natural_mode): Warn ABI change when %zmm register is not
8708 available for AVX512F vector value passing.
8710 2014-02-18 Kai Tietz <ktietz@redhat.com>
8713 * config/i386/i386.c (ix86_expand_prologue): Use value in
8714 rax register as displacement when restoring %r10 or %rax.
8715 Fix wrong offset when restoring both registers.
8717 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8719 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8720 assertion with conditional return.
8722 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8723 Uros Bizjak <ubizjak@gmail.com>
8726 * config/i386/driver-i386.c (host_detect_local_cpu): If
8727 YMM state is not saved by the OS, also clear has_f16c. Move
8728 CPUID 0x80000001 handling before YMM state saving checking.
8730 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8732 PR rtl-optimization/58960
8733 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8734 factored out from ...
8735 (sched_init): ... here.
8736 (free_global_sched_pressure_data): New, factored out from ...
8737 (sched_finish): ... here.
8738 * sched-int.h (free_global_sched_pressure_data): Declare.
8739 * sched-rgn.c (nr_regions_initial): New static global.
8740 (haifa_find_rgns): Initialize it.
8741 (schedule_region): Disable sched-pressure for the newly
8744 2014-02-17 Richard Biener <rguenther@suse.de>
8746 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8747 release SSA defs of pattern stmts.
8749 2014-02-17 Richard Biener <rguenther@suse.de>
8751 * tree-inline.c (expand_call_inline): Release the virtual
8752 operand defined by the call we are about to inline.
8754 2014-02-17 Richard Biener <rguenther@suse.de>
8756 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8758 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8759 Ilya Tocar <ilya.tocar@intel.com>
8761 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8762 arguments order in builtin.
8763 (_mm512_permutexvar_epi64): Ditto.
8764 (_mm512_mask_permutexvar_epi64): Ditto
8765 (_mm512_maskz_permutexvar_epi32): Ditto
8766 (_mm512_permutexvar_epi32): Ditto
8767 (_mm512_mask_permutexvar_epi32): Ditto
8769 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8771 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8772 (p8_vmrgow): Likewise.
8774 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8776 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8779 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8782 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8783 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8784 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8785 using direct move instructions on ISA 2.07. Also adjust
8786 instruction length for 64-bit.
8787 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8788 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8790 2014-02-15 Alan Modra <amodra@gmail.com>
8794 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8795 find_replacement on parts of insn rtl that might be reloaded.
8797 2014-02-15 Richard Biener <rguenther@suse.de>
8799 PR tree-optimization/60183
8800 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8801 (tree_ssa_phiprop): Calculate and free post-dominators.
8803 2014-02-14 Jeff Law <law@redhat.com>
8805 PR rtl-optimization/60131
8806 * ree.c (get_extended_src_reg): New function.
8807 (combine_reaching_defs): Use it rather than assuming location of REG.
8808 (find_and_remove_re): Verify first operand of extension is
8809 a REG before adding the insns to the copy list.
8811 2014-02-14 Roland McGrath <mcgrathr@google.com>
8813 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8814 * configure: Regenerated.
8815 * config.in: Regenerated.
8816 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8817 instead of ASM_SHORT.
8819 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8820 Richard Earnshaw <rearnsha@arm.com>
8822 PR rtl-optimization/59535
8823 * lra-constraints.c (process_alt_operands): Encourage alternative
8824 when unassigned pseudo class is superset of the alternative class.
8825 (inherit_reload_reg): Don't inherit when optimizing for code size.
8826 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8827 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8828 modes not less than 4 for Thumb1.
8830 2014-02-14 Kyle McMartin <kyle@redhat.com>
8833 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8835 2014-02-14 Richard Biener <rguenther@suse.de>
8837 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8838 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8839 Do not strip INDIRECT_REFs.
8841 2014-02-14 Richard Biener <rguenther@suse.de>
8844 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8845 DECL_FUNCTION_SPECIFIC_TARGET.
8846 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8847 * tree-streamer-out.c (pack_ts_target_option): Remove.
8848 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8849 (write_ts_function_decl_tree_pointers): Do not stream
8850 DECL_FUNCTION_SPECIFIC_TARGET.
8851 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8852 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8853 (lto_input_ts_function_decl_tree_pointers): Do not stream
8854 DECL_FUNCTION_SPECIFIC_TARGET.
8856 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8858 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8859 (get_initial_def_for_induction, vectorizable_induction): Ignore
8860 debug stmts when looking for exit_phi.
8861 (vectorizable_live_operation): Fix up condition.
8863 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8865 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8866 nreverse() because it changes the content of original tree list.
8868 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8870 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8871 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8873 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8875 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8876 GNU coding standards.
8878 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8881 * dwarf2out.c (gen_subprogram_die): Don't call
8882 add_calling_convention_attribute if subr_die is old_die.
8884 2014-02-13 Sharad Singhai <singhai@google.com>
8886 * doc/optinfo.texi: Fix order of nodes.
8888 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8890 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8891 operands[2], not operands[3].
8893 2014-02-13 Richard Biener <rguenther@suse.de>
8896 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8897 mention the possibility of an in-tree build.
8898 (CLooG): Update recommended version to 0.18.1, mention the
8899 possibility of an in-tree build and clarify that the ISL
8900 bundled with CLooG does not work.
8902 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8905 * expr.c (compress_float_constant): If x is a hard register,
8906 extend into a pseudo and then move to x.
8908 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8910 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8911 caused by bad second argument to warning_at() with -mhotpatch and
8912 nested functions (e.g. with gfortran).
8914 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8916 * opts.c (option_name): Remove "enabled by default" rider.
8918 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8920 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8922 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8923 Uros Bizjak <ubizjak@gmail.com>
8926 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8927 * configure: Regenerated.
8929 2014-02-12 Richard Biener <rguenther@suse.de>
8931 * vec.c (vec_prefix::calculate_allocation): Move as
8932 inline variant to vec.h.
8933 (vec_prefix::calculate_allocation_1): New out-of-line version.
8934 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8935 (vec_prefix::m_has_auto_buf): Rename to ...
8936 (vec_prefix::m_using_auto_storage): ... this.
8937 (vec_prefix::calculate_allocation): Inline the easy cases
8938 and dispatch to calculate_allocation_1 which doesn't need the
8940 (va_heap::reserve): Use gcc_checking_assert.
8941 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8942 m_using_auto_storage.
8943 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8945 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8946 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8947 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8949 2014-02-12 Richard Biener <rguenther@suse.de>
8951 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8952 when we found a dependence.
8954 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8956 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8958 (maybe_fold_stmt): ... into this new function.
8959 * omp-low.c (lower_omp): Update comment.
8961 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8964 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8967 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8969 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8970 identifiers in comments.
8971 (cortexa53_extra_costs): Likewise.
8972 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8973 (cortexa7_extra_costs): Likewise.
8974 (cortexa12_extra_costs): Likewise.
8975 (cortexa15_extra_costs): Likewise.
8976 (v7m_extra_costs): Likewise.
8978 2014-02-12 Richard Biener <rguenther@suse.de>
8981 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8982 of posix_memalign being successful.
8983 (lower_stmt): Restrict lowering of posix_memalign to when
8984 -ftree-bit-ccp is enabled.
8986 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8988 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8990 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8992 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8994 PR rtl-optimization/60116
8995 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8996 other_insn once the combination has been validated.
8998 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9001 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9003 * ipa-devirt.c: Include demangle.h
9004 (odr_violation_reported): New static variable.
9005 (add_type_duplicate): Update odr_violations.
9006 (maybe_record_node): Add completep parameter; update it.
9007 (record_target_from_binfo): Add COMPLETEP parameter;
9008 update it as needed.
9009 (possible_polymorphic_call_targets_1): Likewise.
9010 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9011 rename FINAL to COMPLETE.
9012 (record_targets_from_bases): Sanity check we found the binfo;
9013 fix COMPLETEP updating.
9014 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9015 parameter, fix computing of COMPLETEP.
9016 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9017 at LTO time do demangling.
9018 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9019 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9021 (gimple_get_virt_method_for_binfo): Likewise.
9022 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9023 gimple_get_virt_method_for_vtable): Update prototypes.
9025 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9028 * genautomata.c (add_presence_absence): Fix typo with
9029 {final_}presence_list.
9031 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9034 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9035 for VSX/Altivec vectors that land in GPR registers.
9037 2014-02-11 Richard Henderson <rth@redhat.com>
9038 Jakub Jelinek <jakub@redhat.com>
9041 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9042 around drhs if type conversion to lacc->type is not useless.
9044 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9046 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9048 (cortex-a57.cortex-a53): Likewise.
9049 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9051 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9053 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9056 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9058 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9059 add_options_for_arm_vfp3.
9061 2014-02-11 Jeff Law <law@redhat.com>
9064 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9065 object with an undesirable mode.
9067 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9070 * config/i386/sol2-9.h: New file.
9071 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9072 *-*-solaris2.9*): Use it.
9074 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9076 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9077 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9079 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9081 * config/microblaze/microblaze.c: Extend mcpu version format
9083 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9085 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9087 2014-02-10 Richard Henderson <rth@redhat.com>
9090 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9091 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9092 ms-abi vs -mno-accumulate-outgoing-args.
9093 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9094 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9097 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9100 * cfgexpand.c (expand_asm_operands): Attach source location to
9101 ASM_INPUT rtx objects.
9102 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9104 2014-02-10 Nick Clifton <nickc@redhat.com>
9106 * config/mn10300/mn10300.c (popcount): New function.
9107 (mn10300_expand_prologue): Include saved registers in stack usage
9110 2014-02-10 Jeff Law <law@redhat.com>
9113 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9114 when changing the SET_DEST of a prior insn to avoid an input reload.
9116 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9118 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9119 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9120 -mcall-openbsd, or -mcall-linux.
9121 (CC1_ENDIAN_BIG_SPEC): Remove.
9122 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9123 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9124 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9125 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9126 and %cc1_endian_default.
9127 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9129 2014-02-10 Richard Biener <rguenther@suse.de>
9131 PR tree-optimization/60115
9132 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9133 MEM_REF handling. Properly verify that the accesses are not
9134 out of the objects bound.
9136 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9138 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9141 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9143 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9144 proper constants and fix formatting.
9145 (possible_polymorphic_call_targets): Fix formatting.
9147 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9148 Ilya Tocar <ilya.tocar@intel.com>
9150 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9151 (_mm512_loadu_epi32): Renamed into...
9152 (_mm512_loadu_si512): This.
9153 (_mm512_storeu_epi32): Renamed into...
9154 (_mm512_storeu_si512): This.
9155 (_mm512_maskz_ceil_ps): Removed.
9156 (_mm512_maskz_ceil_pd): Ditto.
9157 (_mm512_maskz_floor_ps): Ditto.
9158 (_mm512_maskz_floor_pd): Ditto.
9159 (_mm512_floor_round_ps): Ditto.
9160 (_mm512_floor_round_pd): Ditto.
9161 (_mm512_ceil_round_ps): Ditto.
9162 (_mm512_ceil_round_pd): Ditto.
9163 (_mm512_mask_floor_round_ps): Ditto.
9164 (_mm512_mask_floor_round_pd): Ditto.
9165 (_mm512_mask_ceil_round_ps): Ditto.
9166 (_mm512_mask_ceil_round_pd): Ditto.
9167 (_mm512_maskz_floor_round_ps): Ditto.
9168 (_mm512_maskz_floor_round_pd): Ditto.
9169 (_mm512_maskz_ceil_round_ps): Ditto.
9170 (_mm512_maskz_ceil_round_pd): Ditto.
9171 (_mm512_expand_pd): Ditto.
9172 (_mm512_expand_ps): Ditto.
9173 * config/i386/i386.c (ix86_builtins): Remove
9174 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9175 (bdesc_args): Ditto.
9176 * config/i386/predicates.md (const1256_operand): New.
9177 (const_1_to_2_operand): Ditto.
9178 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9179 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9180 (*avx512pf_gatherpf<mode>sf): Ditto.
9181 (avx512pf_gatherpf<mode>df): Ditto.
9182 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9183 (*avx512pf_gatherpf<mode>df): Ditto.
9184 (avx512pf_scatterpf<mode>sf): Ditto.
9185 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9186 (*avx512pf_scatterpf<mode>sf): Ditto.
9187 (avx512pf_scatterpf<mode>df): Ditto.
9188 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9189 (*avx512pf_scatterpf<mode>df): Ditto.
9190 (avx512f_expand<mode>): Removed.
9191 (<shift_insn><mode>3<mask_name>): Change predicate type.
9193 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9195 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9196 not at the end of datarefs vector use ordered_remove to avoid
9197 reordering datarefs vector.
9200 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9201 mark local addressable non-static vars as GOVD_PRIVATE
9202 instead of GOVD_LOCAL.
9203 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9204 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9208 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9209 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9210 assume_aligned or alloc_align attributes.
9211 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9212 arguments. Handle also assume_aligned and alloc_align attributes.
9213 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9214 calls to functions with assume_aligned or alloc_align attributes.
9215 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9217 2014-02-08 Terry Guo <terry.guo@arm.com>
9219 * doc/invoke.texi: Document ARM -march=armv7e-m.
9221 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9223 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9224 flag on __cilkrts_rethrow builtin.
9227 * ipa-cp.c (determine_versionability): Fail at -O0
9228 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9229 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9232 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9235 * tree-inline.c (copy_forbidden): Fail for
9236 __attribute__((optimize (0))) functions.
9238 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9240 * varpool.c: Include pointer-set.h.
9241 (varpool_remove_unreferenced_decls): Variables in other partitions
9242 will not be output; be however careful to not lose information
9245 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9247 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9248 lookup in the vtable constructor.
9250 2014-02-07 Jeff Law <law@redhat.com>
9253 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9254 define_insn_and_split.
9256 * ipa-inline.c (inline_small_functions): Fix typos.
9258 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9260 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9261 (s390_can_use_return_insn): Declare.
9262 * config/s390/s390.h (EPILOGUE_USES): Define.
9263 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9265 (s390_chunkify_start): Handle return JUMP_LABELs.
9266 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9267 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9268 (s390_can_use_return_insn): New functions.
9269 (s390_fix_long_loop_prediction): Handle conditional returns.
9270 (TARGET_SET_UP_BY_PROLOGUE): Define.
9271 * config/s390/s390.md (ANY_RETURN): New code iterator.
9272 (*creturn, *csimple_return, return, simple_return): New patterns.
9274 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9276 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9277 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9278 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9279 REG_CFA_RESTORE list when deciding not to restore a register.
9281 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9283 * config/s390/s390.c: Include tree-pass.h and context.h.
9284 (s390_early_mach): New function, split out from...
9285 (s390_emit_prologue): ...here.
9286 (pass_data_s390_early_mach): New pass structure.
9287 (pass_s390_early_mach): New class.
9288 (s390_option_override): Create and register early_mach pass.
9289 Move to end of file.
9291 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9293 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9294 to match for the exit block.
9296 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9298 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9299 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9300 Reject misaligned operands.
9302 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9304 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9306 2014-02-07 Richard Biener <rguenther@suse.de>
9309 * gimple-low.c (lower_builtin_posix_memalign): New function.
9310 (lower_stmt): Call it to lower posix_memalign in a way
9311 to make alignment info accessible.
9313 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9316 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9317 __builtin_setjmp_receiver.
9319 2014-02-07 Richard Biener <rguenther@suse.de>
9322 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9323 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9324 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9325 Handle BUILT_IN_POSIX_MEMALIGN.
9326 (find_func_clobbers): Likewise.
9327 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9328 (call_may_clobber_ref_p_1): Likewise.
9330 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9333 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9336 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9339 * lto-cgraph.c (lto_output_node): Use
9340 symtab_get_symbol_partitioning_class.
9341 (lto_output_varpool_node): likewise.
9342 (symtab_get_symbol_partitioning_class): Move here from
9344 * cgraph.h (symbol_partitioning_class): Likewise.
9345 (symtab_get_symbol_partitioning_class): Declare.
9347 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9349 * ggc.h (ggc_internal_cleared_alloc): New macro.
9350 * vec.h (vec_safe_copy): Handle memory stats.
9351 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9352 * target-globals.c (save_target_globals): Likewise.
9354 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9357 * expr.c (emit_move_resolve_push): Export; be bit more selective
9358 on when to clear alias set.
9359 * expr.h (emit_move_resolve_push): Declare.
9360 * function.h (struct function): Add tail_call_marked.
9361 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9362 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9363 * config/i386/i386.md (TImode move expander): De not call
9365 (FP push expanders): Preserve memory attributes.
9366 * config/i386/sse.md (push<mode>1): Remove.
9367 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9368 (ix86_expand_push): Remove.
9369 * config/i386/mmx.md (push<mode>1): Remove.
9371 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9373 PR rtl-optimization/60030
9374 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9375 lopart with paradoxical subreg before shifting it up by hprec.
9377 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9379 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9380 Remove extra newline at end of file.
9381 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9382 (arm_issue_rate): Handle cortexa57.
9383 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9384 (cortex-a57.cortex-a53): Likewise.
9386 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9389 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9390 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9392 (arm_expand_prologue): Adjust all callers.
9393 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9394 info, registers also at the lowest numbered registers side. Use
9395 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9399 * var-tracking.c (adjust_mems): Before adding a SET to
9400 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9402 2014-02-06 Alan Modra <amodra@gmail.com>
9405 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9406 change SDmode to DDmode when lra_in_progress.
9408 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9411 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9412 free_data_ref on the dr first, and before goto again also set dr
9413 to the next dr. For simd_lane_access, free old datarefs[i] before
9414 overwriting it. For get_vectype_for_scalar_type failure, don't
9415 free_data_ref if simd_lane_access.
9417 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9420 * tree.h (opts_for_fn): New inline function.
9421 (opt_for_fn): Define.
9422 * config/i386/i386.c (ix86_function_regparm): Use
9423 opt_for_fn (decl, optimize) instead of optimize.
9425 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9427 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9428 for SYMBOL_REF in large memory model.
9430 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9432 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9434 (cortex-a57): Likewise.
9435 (cortex-a57.cortex-a53): Likewise.
9437 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9438 Kugan Vivekanandarajah <kuganv@linaro.org>
9440 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9442 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9444 2014-02-06 Richard Biener <rguenther@suse.de>
9446 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9447 set_loop_copy and initialize_original_copy_tables.
9449 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9451 * config/aarch64/aarch64-simd.md
9452 (aarch64_ashr_simddi): Change QI to SI.
9454 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9455 Jakub Jelinek <jakub@redhat.com>
9458 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9461 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9463 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9464 CODE_FOR_altivec_vpku[hw]um to
9465 CODE_FOR_altivec_vpku[hw]um_direct.
9466 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9467 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9468 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9469 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9471 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9473 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9474 generation for -maltivec=be.
9475 (altivec_vsumsws): Simplify redundant test.
9477 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9479 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9480 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9481 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9482 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9483 gen_altivec_vpkuwum.
9484 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9486 (altivec_vpks<VI_char>ss): Likewise.
9487 (altivec_vpks<VI_char>us): Likewise.
9488 (altivec_vpku<VI_char>us): Likewise.
9489 (altivec_vpku<VI_char>um): Likewise.
9490 (altivec_vpku<VI_char>um_direct): New (copy of
9491 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9493 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9494 target is little endian and -maltivec=be is not specified.
9495 (*altivec_vupkhs<VU_char>_direct): New (copy of
9496 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9497 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9498 target is little endian and -maltivec=be is not specified.
9499 (*altivec_vupkls<VU_char>_direct): New (copy of
9500 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9501 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9502 little endian and -maltivec=be is not specified.
9503 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9504 little endian and -maltivec=be is not specified.
9506 2014-02-05 Richard Henderson <rth@redhat.com>
9509 * combine-stack-adj.c: Revert r206943.
9510 * sched-int.h (struct deps_desc): Add last_args_size.
9511 * sched-deps.c (init_deps): Initialize it.
9512 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9513 contain REG_ARGS_SIZE notes.
9515 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9517 * lto-cgraph.c (asm_nodes_output): Make global.
9518 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9519 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9520 (driver_handle_option): Handle OPT_fwpa.
9522 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9525 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9526 a comment typo and formatting issue. If odr_hash hasn't been
9527 created, return vNULL and set *completep to false.
9530 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9531 bb with no successors.
9533 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9536 * doc/invoke.texi (-march): Clarify documentation for ARM.
9540 2014-02-05 Richard Biener <rguenther@suse.de>
9542 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9543 when not vectorizing because of too many alias checks.
9544 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9545 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9547 2014-02-05 Nick Clifton <nickc@redhat.com>
9549 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9550 accept extended registers in any mode when compiling for the MN10300.
9552 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9554 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9555 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9556 sanitization attributes.
9557 (can_inline_edge_p): Likewise.
9558 (sanitize_attrs_match_for_inline_p): New function.
9560 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9562 * ipa-prop.c (detect_type_change): Shor circuit testing of
9563 type changes on THIS pointer.
9565 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9568 * config/pa/pa.c (legitimize_tls_address): Return original address
9569 if not passed a SYMBOL_REF rtx.
9570 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9572 (pa_emit_move_sequence): Simplify TLS source operands.
9573 (pa_legitimate_constant_p): Reject all TLS constants.
9574 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9575 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9577 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9579 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9580 groups when we know they are controlled by LTO.
9581 * varasm.c (default_binds_local_p_1): If object is in other partition,
9582 it will be resolved locally.
9584 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9586 * config/host-linux.c (linux_gt_pch_use_address): Don't
9587 use SSIZE_MAX because it is not always defined.
9589 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9592 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9593 threshold for pseudo splitting.
9594 (update_ebb_live_info): Process call argument hard registers and
9595 hard registers from insn definition too.
9596 (max_small_class_regs_num): New constant.
9597 (inherit_in_ebb): Update live hard regs through EBBs. Update
9598 reloads_num only for small register classes. Don't split for
9601 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9604 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9607 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9609 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9612 2014-02-04 Marek Polacek <polacek@redhat.com>
9614 * gdbinit.in (pel): Define.
9616 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9618 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9621 2014-02-04 Richard Biener <rguenther@suse.de>
9624 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9625 in function context local.
9626 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9627 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9628 similar to LTO_imported_decl_ref.
9630 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9632 PR tree-optimization/60002
9633 * cgraphclones.c (build_function_decl_skip_args): Clear
9636 PR tree-optimization/60023
9637 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9638 false to gsi_replace.
9639 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9640 has been in some EH region and vec_stmt could throw, add
9641 vec_stmt into the same EH region.
9642 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9643 has no lhs, ignore it.
9644 * internal-fn.c (expand_MASK_LOAD): Likewise.
9647 * tree-inline.c (copy_forbidden): Fail for
9648 __attribute__((optimize (0))) functions.
9651 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9652 is set, copy one less argument.
9653 (expand_simd_clones): Don't subtract clone_info->inbranch
9654 from simd_clone_struct_alloc argument.
9656 PR rtl-optimization/57915
9657 * recog.c (simplify_while_replacing): If all unary/binary/relational
9658 operation arguments are constant, attempt to simplify those.
9661 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9662 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9664 2014-02-04 Richard Biener <rguenther@suse.de>
9666 PR tree-optimization/60012
9667 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9668 TBAA disambiguation to all DDRs.
9670 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9673 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9674 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9676 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9679 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9681 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9683 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9684 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9686 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9689 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9690 to figure out targets of polymorphic calls with known decl.
9691 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9692 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9693 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9694 (get_polymorphic_call_info): ... here.
9695 (get_polymorphic_call_info_from_invariant): New function.
9697 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9699 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9700 lookup via vtable pointer; check for type consistency
9701 and turn inconsitent facts into UNREACHABLE.
9702 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9703 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9704 type inconsistent querries; return UNREACHABLE instead.
9706 2014-02-03 Richard Henderson <rth@twiddle.net>
9709 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9710 already processed this node.
9711 (normalize_one_pred_1): Pass along mark_set.
9712 (normalize_one_pred): Create and destroy a pointer_set_t.
9713 (normalize_one_pred_chain): Likewise.
9715 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9717 PR gcov-profile/58602
9718 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9720 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9723 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9724 -fno-devirtualize; try to devirtualize by the knowledge of
9725 virtual table pointer given by aggregate propagation.
9726 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9727 (ipa_print_node_jump_functions): Dump also offset that
9728 is relevant for polymorphic calls.
9729 (determine_known_aggregate_parts): Add arg_type parameter; use it
9730 instead of determining the type from pointer type.
9731 (ipa_compute_jump_functions_for_edge): Update call of
9732 determine_known_aggregate_parts.
9733 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9734 (gimple_get_virt_method_for_binfo): ... here; simplify using
9735 vtable_pointer_value_to_vtable.
9736 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9737 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9738 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9739 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9741 (vtable_pointer_value_to_binfo): ... here.
9742 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9744 2014-02-03 Teresa Johnson <tejohnson@google.com>
9746 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9747 redef of outer loop index variable.
9749 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9753 * doc/extend.texi (Function Attributes): Typo.
9755 2014-02-03 Cong Hou <congh@google.com>
9757 PR tree-optimization/60000
9758 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9759 if the vectorized statement is a store. A store statement can only
9760 appear at the end of pattern statements.
9762 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9764 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9765 (ix86_option_override_internal): Default long double to 64-bit for
9766 32-bit Bionic and to 128-bit for 64-bit Bionic.
9768 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9769 TARGET_LONG_DOUBLE_128 is true.
9770 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9772 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9773 (mlong-double-64): Negate -mlong-double-128.
9774 (mlong-double-128): New option.
9776 * config/i386/i386-c.c (ix86_target_macros): Define
9777 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9779 * doc/invoke.texi: Document -mlong-double-128.
9781 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9783 PR rtl-optimization/60024
9784 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9786 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9788 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9790 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9792 PR rtl-optimization/57662
9793 * sel-sched.c (code_motion_path_driver): Do not mark already not
9794 existing blocks in the visiting bitmap.
9796 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9798 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9799 on the insn being emitted.
9801 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9802 Will Deacon <will.deacon@arm.com>
9804 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9806 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9808 * config/arm/arm-tables.opt: Regenerate.
9810 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9812 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9813 for vector types other than V16QImode.
9814 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9815 define_expand, and call altivec_expand_vec_perm_le when producing
9816 code with little endian element order.
9817 (*altivec_vperm_<mode>_internal): New insn having previous
9818 behavior of altivec_vperm_<mode>.
9819 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9820 altivec_expand_vec_perm_le when producing code with little endian
9822 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9823 behavior of altivec_vperm_<mode>_uns.
9825 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9827 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9828 (altivec_vsumsws): Add handling for -maltivec=be with a little
9830 (altivec_vsumsws_direct): New.
9831 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9832 gen_altivec_vsumsws.
9834 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9836 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9837 vtable_pointer_value_to_binfo): New functions.
9838 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9839 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9841 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9843 * config/nios2/nios2.md (load_got_register): Initialize GOT
9844 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9845 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9847 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9849 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9850 preserverd by passthrough, do not propagate the type.
9852 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9854 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9855 (mips_atomic_assign_expand_fenv): New function.
9856 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9858 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9860 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9861 (__builtin_mips_set_fcsr): Likewise.
9862 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9863 MIPS_USI_FTYPE_VOID.
9864 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9865 (mips16_expand_set_fcsr): Likewise.
9866 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9867 (mips16_set_fcsr_stub): Likewise.
9868 (mips16_get_fcsr_one_only_stub): New class.
9869 (mips16_set_fcsr_one_only_stub): Likewise.
9870 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9871 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9872 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9873 (hard_float): New availability predicate.
9874 (mips_builtins): Add get_fcsr and set_fcsr.
9875 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9876 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9877 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9878 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9879 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9882 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9884 * config/mips/mips.c (mips_one_only_stub): New class.
9885 (mips_need_mips16_rdhwr_p): Replace with...
9886 (mips16_rdhwr_stub): ...this new variable.
9887 (mips16_stub_call_address): New function.
9888 (mips16_rdhwr_one_only_stub): New class.
9889 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9890 (mips_output_mips16_rdhwr): Delete.
9891 (mips_finish_stub): New function.
9892 (mips_code_end): Use it to handle rdhwr stubs.
9894 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9897 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9898 when calculating size of integer atomic types.
9900 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9902 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9904 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9906 PR tree-optimization/60003
9907 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9908 * profile.c (branch_prob): Use gimple_call_builtin_p
9909 to check for BUILT_IN_SETJMP_RECEIVER.
9910 * tree-inline.c (copy_bb): Call notice_special_calls.
9912 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9915 * lra-constraints.c (process_alt_operands): Update reload_sum only
9918 2014-01-31 Richard Henderson <rth@redhat.com>
9921 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9922 until after else_eh is processed.
9924 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9926 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9927 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9928 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9929 in smmintrin.h, remove them.
9930 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9931 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9932 * config/i386/i386.md (ROUND_SAE): Fix value.
9933 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9934 (const48_operand): New.
9935 * config/i386/subst.md (round), (round_expand): Use
9936 const_4_or_8_to_11_operand.
9937 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9939 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9941 * config/i386/constraints.md (Yk): Swap meaning with k.
9942 * config/i386/i386.md (movhi_internal): Change Yk to k.
9943 (movqi_internal): Ditto.
9944 (*k<logic><mode>): Ditto.
9947 (kandn<mode>): Ditto.
9948 (*<code>hi_1): Ditto.
9949 (*<code>qi_1): Ditto.
9950 (kxnor<mode>): Ditto.
9951 (kortestzhi): Ditto.
9952 (kortestchi): Ditto.
9954 (*one_cmplhi2_1): Ditto.
9955 (*one_cmplqi2_1): Ditto.
9956 * config/i386/sse.md (): Change k to Yk.
9957 (avx512f_load<mode>_mask): Ditto.
9958 (avx512f_blendm<mode>): Ditto.
9959 (avx512f_store<mode>_mask): Ditto.
9960 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9961 (avx512f_storedqu<mode>_mask): Ditto.
9962 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9964 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9965 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9966 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9967 (avx512f_maskcmp<mode>3): Ditto.
9968 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9969 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9970 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9971 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9972 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9973 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9974 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9975 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9976 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9977 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9978 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9979 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9980 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9981 (vec_extract_lo_<mode>_maskm): Ditto.
9982 (vec_extract_hi_<mode>_maskm): Ditto.
9983 (avx512f_vternlog<mode>_mask): Ditto.
9984 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9985 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9986 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9987 (avx512f_<code>v8div16qi2_mask): Ditto.
9988 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9989 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9990 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9991 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9992 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9993 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9994 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9995 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9996 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9997 (avx512cd_maskb_vec_dupv8di): Ditto.
9998 (avx512cd_maskw_vec_dupv16si): Ditto.
9999 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10000 (avx512f_vpermi2var<mode>3_mask): Ditto.
10001 (avx512f_vpermi2var<mode>3_mask): Ditto.
10002 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10003 (*avx512f_gathersi<mode>): Ditto.
10004 (*avx512f_gathersi<mode>_2): Ditto.
10005 (*avx512f_gatherdi<mode>): Ditto.
10006 (*avx512f_gatherdi<mode>_2): Ditto.
10007 (*avx512f_scattersi<mode>): Ditto.
10008 (*avx512f_scatterdi<mode>): Ditto.
10009 (avx512f_compress<mode>_mask): Ditto.
10010 (avx512f_compressstore<mode>_mask): Ditto.
10011 (avx512f_expand<mode>_mask): Ditto.
10012 * config/i386/subst.md (mask): Change k to Yk.
10013 (mask_scalar_merge): Ditto.
10016 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10018 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10020 2014-01-31 Richard Biener <rguenther@suse.de>
10022 PR middle-end/59990
10023 * builtins.c (fold_builtin_memory_op): Make sure to not
10024 use a floating-point mode or a boolean or enumeral type for
10025 the copy operation.
10027 2014-01-30 DJ Delorie <dj@redhat.com>
10029 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10030 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10031 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10032 whenever main() has an epilogue.
10034 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10036 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10037 unused variable "field".
10038 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10039 (vsx_mergeh_<mode>): Likewise.
10040 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10041 (altivec_vmrghh): Likewise.
10042 (altivec_vmrghw): Likewise.
10043 (altivec_vmrglb): Likewise.
10044 (altivec_vmrglh): Likewise.
10045 (altivec_vmrglw): Likewise.
10046 (altivec_vspltb): Add missing uses.
10047 (altivec_vsplth): Likewise.
10048 (altivec_vspltw): Likewise.
10049 (altivec_vspltsf): Likewise.
10051 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10054 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10055 frame related instructions.
10057 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10059 PR rtl-optimization/59959
10060 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10061 any reload of register whose subreg is invalid.
10063 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10065 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10066 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10067 Add missing return type - void.
10069 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10071 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10072 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10073 remove element index adjustment for endian (now handled in vsx.md
10075 (altivec_expand_vec_perm_const): Use
10076 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10077 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10078 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10079 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10080 define_expand and a new define_insn *altivec_vspltb_internal;
10081 adjust for -maltivec=be on a little endian target.
10082 (altivec_vspltb_direct): New.
10083 (altivec_vsplth): Divide into a define_expand and a new
10084 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10085 little endian target.
10086 (altivec_vsplth_direct): New.
10087 (altivec_vspltw): Divide into a define_expand and a new
10088 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10089 little endian target.
10090 (altivec_vspltw_direct): New.
10091 (altivec_vspltsf): Divide into a define_expand and a new
10092 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10093 a little endian target.
10095 2014-01-30 Richard Biener <rguenther@suse.de>
10097 PR tree-optimization/59993
10098 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10099 can propagate form the earlier stmt and avoid the transform
10100 when the intermediate result is needed.
10102 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10104 * README.Portability: Fix typo.
10106 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10108 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10109 comparison_operator with ordered_comparison_operator.
10111 2014-01-30 Nick Clifton <nickc@redhat.com>
10113 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10114 Rename to mn10300_store_multiple_regs.
10115 * config/mn10300/mn10300.c: Likewise.
10116 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10117 store_multiple_regs.
10118 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10119 Call mn10300_store_multiple_regs.
10121 2014-01-30 Nick Clifton <nickc@redhat.com>
10122 DJ Delorie <dj@redhat.com>
10124 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10125 %fp 2 to keep registers after it properly word-aligned.
10126 (rl78_alloc_physical_registers_umul): Handle the case where both
10127 input operands are the same.
10129 2014-01-30 Richard Biener <rguenther@suse.de>
10131 PR tree-optimization/59903
10132 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10135 2014-01-30 Jason Merrill <jason@redhat.com>
10138 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10141 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10143 2014-01-30 Richard Biener <rguenther@suse.de>
10145 PR tree-optimization/59951
10146 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10148 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10151 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10152 SFmode to DFmode case.
10154 2014-01-29 DJ Delorie <dj@redhat.com>
10156 * config/msp430/msp430.opt (-minrt): New.
10157 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10159 (ENDFILE_SPEC): Likewise.
10161 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10163 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10164 (estimate_function_body_sizes): Use it.
10166 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10169 * dwarf2out.c (is_cxx_auto): New.
10170 (is_base_type): Use it.
10171 (gen_type_die_with_usage): Likewise.
10173 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10175 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10176 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10177 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10178 -maltivec=be with LE targets.
10179 (vsx_mergeh_<mode>): Likewise.
10180 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10181 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10182 (altivec_vmrghb): Replace with define_expand and new
10183 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10184 (altivec_vmrghb_direct): New define_insn.
10185 (altivec_vmrghh): Replace with define_expand and new
10186 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10187 (altivec_vmrghh_direct): New define_insn.
10188 (altivec_vmrghw): Replace with define_expand and new
10189 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10190 (altivec_vmrghw_direct): New define_insn.
10191 (*altivec_vmrghsf): Adjust for endianness.
10192 (altivec_vmrglb): Replace with define_expand and new
10193 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10194 (altivec_vmrglb_direct): New define_insn.
10195 (altivec_vmrglh): Replace with define_expand and new
10196 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10197 (altivec_vmrglh_direct): New define_insn.
10198 (altivec_vmrglw): Replace with define_expand and new
10199 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10200 (altivec_vmrglw_direct): New define_insn.
10201 (*altivec_vmrglsf): Adjust for endianness.
10202 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10203 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10204 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10205 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10206 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10207 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10208 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10209 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10211 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10213 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10214 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10217 2014-01-29 Richard Biener <rguenther@suse.de>
10219 PR tree-optimization/58742
10220 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10221 associate_pointerplus_align.
10222 (associate_pointerplus_diff): New function.
10223 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10224 and associate_pointerplus_diff.
10226 2014-01-29 Richard Biener <rguenther@suse.de>
10228 * lto-streamer.h (LTO_major_version): Bump to 3.
10229 (LTO_minor_version): Reset to 0.
10231 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10233 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10234 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10235 (arm_file_start): Generate correct asm header for armv7ve.
10236 * config/arm/bpabi.h: Add multilib support for armv7ve.
10237 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10238 and cortex-a15 to armv7ve.
10239 * config/arm/t-aprofile: Add multilib support for armv7ve.
10240 * doc/invoke.texi: Document -march=armv7ve.
10242 2014-01-29 Richard Biener <rguenther@suse.de>
10244 PR tree-optimization/58742
10245 * tree-ssa-forwprop.c (associate_plusminus): Return true
10246 if we changed sth, defer EH cleanup to ...
10247 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10248 (simplify_mult): New function.
10250 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10252 PR middle-end/59917
10253 PR tree-optimization/59920
10254 * tree.c (build_common_builtin_nodes): Remove
10255 __builtin_setjmp_dispatcher initialization.
10256 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10257 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10258 instead of gsi_after_labels + manually skipping debug stmts.
10259 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10260 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10261 * tree-inline.c (copy_edges_for_bb): Remove
10262 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10263 argument. Ignore computed_goto_p stmts. Don't call
10264 make_abnormal_goto_edges. If a call might need abnormal edges
10265 for non-local gotos, see if it already has an edge to
10266 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10267 with true argument, don't do anything then, otherwise add
10268 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10269 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10271 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10272 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10273 (lower_stmt): Don't set data->calls_builtin_setjmp.
10274 (lower_builtin_setjmp): Adjust comment.
10275 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10276 * tree-cfg.c (found_computed_goto): Remove.
10277 (factor_computed_gotos): Remove.
10278 (make_goto_expr_edges): Return bool, true for computed gotos.
10279 Don't call make_abnormal_goto_edges.
10280 (build_gimple_cfg): Don't set found_computed_goto, don't call
10281 factor_computed_gotos.
10282 (computed_goto_p): No longer static.
10283 (make_blocks): Don't set found_computed_goto.
10284 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10285 (make_edges): If make_goto_expr_edges returns true, push bb
10286 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10287 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10288 vector. Record mapping between bbs and OpenMP regions if there
10289 are any, adjust make_gimple_omp_edges caller. Call
10290 handle_abnormal_edges.
10291 (make_abnormal_goto_edges): Remove.
10292 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10293 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10294 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10295 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10296 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10297 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10298 filling *region also set *region_idx to (*region)->entry->index.
10301 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10302 For REGs set ORIGINAL_REGNO.
10304 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10306 * doc/md.texi: Mention that a target shouldn't implement
10307 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10310 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10312 PR tree-optimization/59594
10313 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10314 a copy of the datarefs vector rather than the vector itself.
10316 2014-01-28 Jason Merrill <jason@redhat.com>
10319 * dwarf2out.c (auto_die): New static.
10320 (gen_type_die_with_usage): Handle C++1y 'auto'.
10321 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10324 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10327 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10328 (SPEC_X32): Likewise.
10329 (SPEC_64): Likewise.
10330 * config/i386/i386.c (ix86_option_override_internal): Turn off
10331 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10333 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10334 * config/i386/i386.h (TARGET_16BIT): New macro.
10335 (TARGET_16BIT_P): Likewise.
10336 * config/i386/i386.opt: Add m16.
10337 * doc/invoke.texi: Document -m16.
10339 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10341 PR preprocessor/59935
10342 * input.c (location_get_source_line): Bail out on when line number
10343 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10345 2014-01-28 Richard Biener <rguenther@suse.de>
10347 PR tree-optimization/58742
10348 * tree-ssa-forwprop.c (associate_plusminus): Handle
10349 pointer subtraction of the form (T)(P + A) - (T)P.
10351 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10353 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10356 2014-01-28 Richard Biener <rguenther@suse.de>
10359 2014-01-28 Richard Biener <rguenther@suse.de>
10361 PR rtl-optimization/45364
10362 PR rtl-optimization/59890
10363 * var-tracking.c (local_get_addr_clear_given_value): Handle
10364 already cleared slot.
10365 (val_reset): Handle not allocated local_get_addr_cache.
10366 (vt_find_locations): Use post-order on the inverted CFG.
10368 2014-01-28 Richard Biener <rguenther@suse.de>
10370 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10372 2014-01-28 Richard Biener <rguenther@suse.de>
10374 PR rtl-optimization/45364
10375 PR rtl-optimization/59890
10376 * var-tracking.c (local_get_addr_clear_given_value): Handle
10377 already cleared slot.
10378 (val_reset): Handle not allocated local_get_addr_cache.
10379 (vt_find_locations): Use post-order on the inverted CFG.
10381 2014-01-28 Alan Modra <amodra@gmail.com>
10383 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10384 * configure.ac <recursive call for build != host>: Define
10385 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10386 and LD_FOR_BUILD too.
10387 * configure: Regenerate.
10389 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10391 * config/i386/i386.c (get_builtin_code_for_version): Separate
10392 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10393 Broadwell from Haswell.
10395 2014-01-27 Steve Ellcey <sellcey@mips.com>
10397 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10398 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10399 * config/mips/mips.c (mips_option_override): Change setting
10401 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10402 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10403 Change from Mask to Var.
10405 2014-01-27 Jeff Law <law@redhat.com>
10407 * ipa-inline.c (inline_small_functions): Fix typo.
10409 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10411 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10412 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10413 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10414 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10415 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10416 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10417 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10418 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10419 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10420 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10421 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10422 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10423 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10424 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10425 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10426 (_mm512_storeu_epi64): Ditto.
10427 (_mm512_cmpge_epi32_mask): Ditto.
10428 (_mm512_cmpge_epu32_mask): Ditto.
10429 (_mm512_cmpge_epi64_mask): Ditto.
10430 (_mm512_cmpge_epu64_mask): Ditto.
10431 (_mm512_cmple_epi32_mask): Ditto.
10432 (_mm512_cmple_epu32_mask): Ditto.
10433 (_mm512_cmple_epi64_mask): Ditto.
10434 (_mm512_cmple_epu64_mask): Ditto.
10435 (_mm512_cmplt_epi32_mask): Ditto.
10436 (_mm512_cmplt_epu32_mask): Ditto.
10437 (_mm512_cmplt_epi64_mask): Ditto.
10438 (_mm512_cmplt_epu64_mask): Ditto.
10439 (_mm512_cmpneq_epi32_mask): Ditto.
10440 (_mm512_cmpneq_epu32_mask): Ditto.
10441 (_mm512_cmpneq_epi64_mask): Ditto.
10442 (_mm512_cmpneq_epu64_mask): Ditto.
10443 (_mm512_expand_pd): Ditto.
10444 (_mm512_expand_ps): Ditto.
10445 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10446 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10447 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10448 * config/i386/i386.c (ix86_builtins): Add
10449 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10450 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10451 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10452 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10453 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10454 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10455 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10456 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10457 IX86_BUILTIN_PMOVUSQW512_MEM.
10458 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10459 __builtin_ia32_pmovsqd512mem_mask,
10460 __builtin_ia32_pmovqd512mem_mask,
10461 __builtin_ia32_pmovusqw512mem_mask,
10462 __builtin_ia32_pmovsqw512mem_mask,
10463 __builtin_ia32_pmovqw512mem_mask,
10464 __builtin_ia32_pmovusdw512mem_mask,
10465 __builtin_ia32_pmovsdw512mem_mask,
10466 __builtin_ia32_pmovdw512mem_mask,
10467 __builtin_ia32_pmovqb512mem_mask,
10468 __builtin_ia32_pmovusqb512mem_mask,
10469 __builtin_ia32_pmovsqb512mem_mask,
10470 __builtin_ia32_pmovusdb512mem_mask,
10471 __builtin_ia32_pmovsdb512mem_mask,
10472 __builtin_ia32_pmovdb512mem_mask.
10473 (bdesc_args): Add __builtin_ia32_expanddf512,
10474 __builtin_ia32_expandsf512.
10475 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10476 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10477 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10478 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10479 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10480 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10481 (avx512f_<code>v8div16qi2_mask_store): This.
10482 (avx512f_expand<mode>): New.
10484 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10486 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10488 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10489 (_mm512_prefetch_i32scatter_pd): Ditto.
10490 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10491 (_mm512_prefetch_i64scatter_pd): Ditto.
10492 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10493 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10494 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10495 (_mm512_prefetch_i32scatter_ps): Ditto.
10496 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10497 (_mm512_prefetch_i64scatter_ps): Ditto.
10498 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10499 * config/i386/i386-builtin-types.def: Define
10500 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10501 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10502 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10503 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10504 IX86_BUILTIN_SCATTERPFQPD.
10505 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10506 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10507 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10508 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10509 __builtin_ia32_scatterpfqps.
10510 (ix86_expand_builtin): Expand new built-ins.
10511 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10512 fix memory access data type.
10513 (*avx512pf_gatherpf<mode>_mask): Ditto.
10514 (*avx512pf_gatherpf<mode>): Ditto.
10515 (avx512pf_scatterpf<mode>): Ditto.
10516 (*avx512pf_scatterpf<mode>_mask): Ditto.
10517 (*avx512pf_scatterpf<mode>): Ditto.
10518 (GATHER_SCATTER_SF_MEM_MODE): New.
10519 (avx512pf_gatherpf<mode>df): Ditto.
10520 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10521 (*avx512pf_scatterpf<mode>df): Ditto.
10523 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10526 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10527 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10530 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10532 * common/config/arm/arm-common.c
10533 (arm_rewrite_mcpu): Handle multiple names.
10535 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10537 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10539 * gimple-builder.h (create_gimple_tmp): Delete.
10541 2014-01-27 Christian Bruel <christian.bruel@st.com>
10543 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10546 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10548 * config/pa/pa.md (call): Generate indirect long calls to non-local
10549 functions when outputing 32-bit code.
10550 (call_value): Likewise except for special call to buggy powf function.
10552 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10553 portable runtime and PIC indirect calls.
10554 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10555 and PIC call sequences. Use ldo instead of blr to set return register
10556 in PIC call sequence.
10558 2014-01-25 Walter Lee <walt@tilera.com>
10560 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10561 avoid clobbering a live register.
10563 2014-01-25 Walter Lee <walt@tilera.com>
10565 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10566 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10567 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10568 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10570 2014-01-25 Walter Lee <walt@tilera.com>
10572 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10573 arguments on even registers.
10574 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10576 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10577 (BIGGEST_ALIGNMENT): Ditto.
10578 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10580 2014-01-25 Walter Lee <walt@tilera.com>
10582 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10583 insns before bundling.
10584 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10586 2014-01-25 Walter Lee <walt@tilera.com>
10588 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10589 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10590 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10592 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10594 * config/mips/constraints.md (kl): Delete.
10595 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10596 define expands, using...
10597 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10598 instructions for MIPS16.
10599 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10600 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10602 2014-01-25 Walter Lee <walt@tilera.com>
10604 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10608 2014-01-25 Walter Lee <walt@tilera.com>
10610 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10611 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10613 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10615 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10618 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10620 * print-rtl.c (in_call_function_usage): New var.
10621 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10622 EXPR_LIST mode as mode and not as reg note name.
10624 PR middle-end/59561
10625 * cfgloopmanip.c (copy_loop_info): If
10626 loop->warned_aggressive_loop_optimizations, make sure
10627 the flag is set in target loop too.
10629 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10631 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10633 * builtins.def: Likewise.
10634 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10635 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10636 * ira.c (ira_setup_eliminable_regset): Likewise.
10637 * omp-low.c (gate_expand_omp): Likewise.
10638 (execute_lower_omp): Likewise.
10639 (diagnose_sb_0): Likewise.
10640 (gate_diagnose_omp_blocks): Likewise.
10641 (simd_clone_clauses_extract): Likewise.
10644 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10646 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10647 correction for little endian...
10648 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10651 2014-01-24 Jeff Law <law@redhat.com>
10653 PR tree-optimization/59919
10654 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10655 for non-returning calls.
10657 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10659 * common/config/aarch64/aarch64-common.c
10660 (aarch64_rewrite_mcpu): Handle multiple names.
10661 * config/aarch64/aarch64.h
10662 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10664 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10666 * input.c (add_file_to_cache_tab): Handle the case where fopen
10669 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10672 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10673 from push operand if code of push isn't PRE_DEC.
10675 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10678 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10679 -mquad-memory-atomic. Update -mquad-memory documentation to say
10680 it is only used for non-atomic loads/stores.
10682 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10683 -mquad-memory or -mquad-memory-atomic switches.
10685 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10686 -mquad-memory-atomic to ISA 2.07 support.
10688 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10689 to separate support of normal quad word memory operations (ldq, stq)
10690 from the atomic quad word memory operations.
10692 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10693 support to separate non-atomic quad word operations from atomic
10694 quad word operations. Disable non-atomic quad word operations in
10695 little endian mode so that we don't have to swap words after the
10696 load and before the store.
10697 (quad_load_store_p): Add comment about atomic quad word support.
10698 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10699 options printed with -mdebug=reg.
10701 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10702 -mquad-memory-atomic as the test for whether we have quad word
10703 atomic instructions.
10704 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10705 or -mp8-vector are used, allow byte/half-word atomic operations.
10707 * config/rs6000/sync.md (load_lockedti): Insure that the address
10708 is a proper indexed or indirect address for the lqarx instruction.
10709 On little endian systems, swap the hi/lo registers after the lqarx
10711 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10712 insure the address is valid for the lqarx instruction.
10713 (store_conditionalti): Insure that the address is a proper indexed
10714 or indirect address for the stqcrx. instruction. On little endian
10715 systems, swap the hi/lo registers before doing the stqcrx.
10717 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10718 insure the address is valid for the stqcrx. instruction.
10720 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10721 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10722 type of quad memory support is available.
10724 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10726 PR regression/59915
10727 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10728 there is a danger of looping.
10730 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10732 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10733 force flag_ira_loop_pressure if set via command line.
10735 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10737 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10738 (ashr_simd): New builtin handling DI mode.
10739 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10740 (aarch64_sshr_simddi): New match pattern.
10741 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10742 (vshrd_n_s64): Likewise.
10743 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10745 2014-01-23 Nick Clifton <nickc@redhat.com>
10747 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10748 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10749 favour of mcu specific scripts.
10750 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10753 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10754 Alex Velenko <Alex.Velenko@arm.com>
10756 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10757 (vaddv_s16): Likewise.
10758 (vaddv_s32): Likewise.
10759 (vaddv_u8): Likewise.
10760 (vaddv_u16): Likewise.
10761 (vaddv_u32): Likewise.
10762 (vaddvq_s8): Likewise.
10763 (vaddvq_s16): Likewise.
10764 (vaddvq_s32): Likewise.
10765 (vaddvq_s64): Likewise.
10766 (vaddvq_u8): Likewise.
10767 (vaddvq_u16): Likewise.
10768 (vaddvq_u32): Likewise.
10769 (vaddvq_u64): Likewise.
10770 (vaddv_f32): Likewise.
10771 (vaddvq_f32): Likewise.
10772 (vaddvq_f64): Likewise.
10773 (vmaxv_f32): Likewise.
10774 (vmaxv_s8): Likewise.
10775 (vmaxv_s16): Likewise.
10776 (vmaxv_s32): Likewise.
10777 (vmaxv_u8): Likewise.
10778 (vmaxv_u16): Likewise.
10779 (vmaxv_u32): Likewise.
10780 (vmaxvq_f32): Likewise.
10781 (vmaxvq_f64): Likewise.
10782 (vmaxvq_s8): Likewise.
10783 (vmaxvq_s16): Likewise.
10784 (vmaxvq_s32): Likewise.
10785 (vmaxvq_u8): Likewise.
10786 (vmaxvq_u16): Likewise.
10787 (vmaxvq_u32): Likewise.
10788 (vmaxnmv_f32): Likewise.
10789 (vmaxnmvq_f32): Likewise.
10790 (vmaxnmvq_f64): Likewise.
10791 (vminv_f32): Likewise.
10792 (vminv_s8): Likewise.
10793 (vminv_s16): Likewise.
10794 (vminv_s32): Likewise.
10795 (vminv_u8): Likewise.
10796 (vminv_u16): Likewise.
10797 (vminv_u32): Likewise.
10798 (vminvq_f32): Likewise.
10799 (vminvq_f64): Likewise.
10800 (vminvq_s8): Likewise.
10801 (vminvq_s16): Likewise.
10802 (vminvq_s32): Likewise.
10803 (vminvq_u8): Likewise.
10804 (vminvq_u16): Likewise.
10805 (vminvq_u32): Likewise.
10806 (vminnmv_f32): Likewise.
10807 (vminnmvq_f32): Likewise.
10808 (vminnmvq_f64): Likewise.
10810 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10812 * config/aarch64/aarch64-simd.md
10813 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10814 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10815 (*aarch64_mul3_elt<mode>): Likewise.
10816 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10817 (*aarch64_mul3_elt_to_64v2df): Likewise.
10818 (*aarch64_mla_elt<mode>): Likewise.
10819 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10820 (*aarch64_mls_elt<mode>): Likewise.
10821 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10822 (*aarch64_fma4_elt<mode>): Likewise.
10823 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10824 (*aarch64_fma4_elt_to_64v2df): Likewise.
10825 (*aarch64_fnma4_elt<mode>): Likewise.
10826 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10827 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10828 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10829 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10830 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10831 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10832 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10833 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10834 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10836 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10838 * config/aarch64/aarch64-simd.md
10839 (aarch64_be_checked_get_lane<mode>): New define_expand.
10840 * config/aarch64/aarch64-simd-builtins.def
10841 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10842 New builtin definition.
10843 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10844 Use new safe be builtin.
10846 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10848 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10850 (aarch64_be_st1<mode>): Likewise.
10851 (aarch_ld1<VALL:mode>): Define_expand modified.
10852 (aarch_st1<VALL:mode>): Likewise.
10853 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10854 (UNSPEC_ST1): Likewise.
10856 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10858 * config/microblaze/microblaze.md: Add trap insn and attribute
10860 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10862 PR preprocessor/58580
10863 * input.h (location_get_source_line): Take an additional line_size
10865 (void diagnostics_file_cache_fini): Declare new function.
10866 * input.c (struct fcache): New type.
10867 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10868 New static constants.
10869 (diagnostic_file_cache_init, total_lines_num)
10870 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10871 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10872 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10873 (get_next_line, read_next_line, goto_next_line, read_line_num):
10874 New static function definitions.
10875 (diagnostic_file_cache_fini): New function.
10876 (location_get_source_line): Take an additional output line_len
10877 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10879 * diagnostic.c (diagnostic_finish): Call
10880 diagnostic_file_cache_fini.
10881 (adjust_line): Take an additional input parameter for the length
10882 of the line, rather than calculating it with strlen.
10883 (diagnostic_show_locus): Adjust the use of
10884 location_get_source_line and adjust_line with respect to their new
10885 signature. While displaying a line now, do not stop at the first
10886 null byte. Rather, display the zero byte as a space and keep
10887 going until we reach the size of the line.
10888 * Makefile.in: Add vec.o to OBJS-libcommon
10890 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10891 Ilya Tocar <ilya.tocar@intel.com>
10893 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10894 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10895 (__builtin_ia32_kmov16): Ditto.
10896 * config/i386/i386.md (UNSPEC_KMOV): New.
10899 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10901 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10902 (_mm512_storeu_si512): Ditto.
10904 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10907 * rtl.h (get_referenced_operands): Declare.
10908 * recog.c (get_referenced_operands): New function.
10909 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10910 operands have been referenced when recording LO_SUM references.
10912 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10914 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10916 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10918 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10919 Enable for generic and recent AMD targets.
10921 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10923 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10924 ARG_SIZE note when adjustment was eliminated.
10926 2014-01-22 Jeff Law <law@redhat.com>
10928 PR tree-optimization/59597
10929 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10930 in file. Accept new argument REGISTERING and use it to modify
10931 dump output appropriately.
10932 (register_jump_thread): Corresponding changes.
10933 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10934 thread paths involving joiner blocks. Add code to dump cancelled
10935 jump threading paths.
10937 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10939 PR rtl-optimization/59477
10940 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10941 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10943 2014-01-22 Tom Tromey <tromey@redhat.com>
10945 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10947 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10949 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10951 PR rtl-optimization/59896
10952 * lra-constraints.c (process_alt_operands): Check unused note for
10953 matched operands of insn with no output reloads.
10955 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10957 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10958 (mips_move_from_gpr_cost): Likewise.
10960 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10962 PR rtl-optimization/59858
10963 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10964 ira_class_hard_regs_num.
10965 (process_alt_operands): Increase reject for dying matched operand.
10967 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10970 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10971 smaller than size, perform several stores or loads and stores
10972 at dst + count - size to store or copy all of size bytes, rather
10973 than just last modesize bytes.
10975 2014-01-20 DJ Delorie <dj@redhat.com>
10977 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10978 that CLOBBERs are REGs before propogating their values.
10980 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10982 PR middle-end/59789
10983 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10984 (cgraph_inline_failed_type): New function.
10985 * cgraph.h (DEFCIFCODE): Add type.
10986 (cgraph_inline_failed_type_t): New enum.
10987 (cgraph_inline_failed_type): New prototype.
10988 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10989 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10990 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10991 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10992 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10993 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10994 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10995 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10996 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10997 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10998 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10999 OPTIMIZATION_MISMATCH.
11000 * tree-inline.c (expand_call_inline): Emit errors during
11001 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11003 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11006 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11007 mode attribute in insn output.
11009 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11011 * output.h (output_constant): Delete.
11012 * varasm.c (output_constant): Make private.
11014 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11016 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11018 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11020 PR middle-end/59860
11021 * tree.h (fold_builtin_strcat): New prototype.
11022 * builtins.c (fold_builtin_strcat): No longer static. Add len
11023 argument, if non-NULL, don't call c_strlen. Optimize
11024 directly into __builtin_memcpy instead of __builtin_strcpy.
11025 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11026 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11028 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11030 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11031 for SImode_address_operand operands, having only a REG argument.
11033 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11035 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11036 loader name using mbig-endian.
11037 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11039 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11041 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11042 (-mtune): Likewise.
11045 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11047 * config/aarch64/aarch64-protos.h
11048 (aarch64_cannot_change_mode_class_ptr): Declare.
11049 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11050 aarch64_cannot_change_mode_class_ptr): New.
11051 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11052 backend hook aarch64_cannot_change_mode_class.
11054 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11056 * common/config/aarch64/aarch64-common.c
11057 (aarch64_handle_option): Don't handle any option order logic here.
11058 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11059 selected_cpu, warn on architecture version mismatch.
11060 (aarch64_override_options): Fix parsing order for option strings.
11062 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11063 Iain Sandoe <iain@codesourcery.com>
11066 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11067 warning. Amend comment to reflect current functionality.
11069 2014-01-20 Richard Biener <rguenther@suse.de>
11071 PR middle-end/59860
11072 * builtins.c (fold_builtin_strcat): Remove case better handled
11073 by tree-ssa-strlen.c.
11075 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11077 * config/aarch64/aarch64.opt
11078 (mcpu, march, mtune): Make case-insensitive.
11080 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11083 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11084 if operands[1] is a REG or ZERO_EXTEND of a REG.
11086 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11088 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11090 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11092 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11093 long non-pic millicode calls.
11095 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11097 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11099 2014-01-19 Kito Cheng <kito@0xlab.org>
11101 * builtins.c (expand_movstr): Check movstr expand done or fail.
11103 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11104 H.J. Lu <hongjiu.lu@intel.com>
11107 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11108 to DImode for zero-extended addresses.
11110 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11112 PR rtl-optimization/57763
11113 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11114 on the new indirect jump_insn and increment LABEL_NUSES (label).
11116 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11120 * config.gcc (x86_archs): New variable.
11121 (x86_64_archs): Likewise.
11122 (x86_cpus): Likewise.
11123 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11124 --with-arch/--with-cpu= options.
11125 Support --with-arch=/--with-cpu={nehalem,westmere,
11126 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11128 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11130 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11131 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11133 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11135 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11137 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11140 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11141 clear cpp_get_options (parse_in)->warn_unused_macros for
11142 ix86_target_macros_internal with cpp_define.
11144 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11146 * jump.c (delete_related_insns): Keep (use (insn))s.
11147 * reorg.c (redundant_insn): Check for barriers too.
11149 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11151 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11153 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11155 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11156 call to $$dyncall when TARGET_LONG_CALLS is true.
11158 2014-01-17 Jeff Law <law@redhat.com>
11160 * ree.c (combine_set_extension): Temporarily disable test for
11161 changing number of hard registers.
11163 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11165 PR middle-end/58125
11166 * ipa-inline-analysis.c (inline_free_summary):
11167 Do not free summary of aliases.
11169 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11171 PR middle-end/59706
11172 * gimplify.c (gimplify_expr): Use create_tmp_var
11173 instead of create_tmp_var_raw. If cond doesn't have
11174 integral type, don't add the IFN_ANNOTATE builtin at all.
11176 2014-01-17 Martin Jambor <mjambor@suse.cz>
11179 * ipa-cp.c (prev_edge_clone): New variable.
11180 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11181 Also resize prev_edge_clone vector.
11182 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11183 (ipcp_edge_removal_hook): New function.
11184 (ipcp_driver): Register ipcp_edge_removal_hook.
11186 2014-01-17 Andrew Pinski <apinski@cavium.com>
11187 Steve Ellcey <sellcey@mips.com>
11190 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11193 2014-01-17 Jeff Law <law@redhat.com>
11195 PR middle-end/57904
11196 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11197 so that pass_ccp runs first.
11199 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11201 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11202 (ix86_adjust_cost): Use !TARGET_XXX.
11203 (do_reorder_for_imul): Likewise.
11204 (swap_top_of_ready_list): Likewise.
11205 (ix86_sched_reorder): Likewise.
11207 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11209 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11210 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11211 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11212 (intel_memset): New. Duplicate slm_memset.
11213 (intel_cost): New. Duplicate slm_cost.
11214 (m_INTEL): New macro.
11215 (processor_target_table): Add "intel".
11216 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11217 with PROCESSOR_INTEL for "intel".
11218 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11219 PROCESSOR_SILVERMONT.
11220 (ix86_issue_rate): Likewise.
11221 (ix86_adjust_cost): Likewise.
11222 (ia32_multipass_dfa_lookahead): Likewise.
11223 (swap_top_of_ready_list): Likewise.
11224 (ix86_sched_reorder): Likewise.
11225 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11226 instead of TARGET_OPT_AGU.
11227 * config/i386/i386.h (TARGET_INTEL): New.
11228 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11229 (processor_type): Add PROCESSOR_INTEL.
11230 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11231 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11233 2014-01-17 Marek Polacek <polacek@redhat.com>
11236 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11239 2014-01-17 Richard Biener <rguenther@suse.de>
11241 PR tree-optimization/46590
11242 * opts.c (default_options_table): Add entries for
11243 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11244 all enabled at -O1 but not for -Og.
11245 * common.opt (fbranch-count-reg): Remove Init(1).
11246 (fmove-loop-invariants): Likewise.
11247 (ftree-pta): Likewise.
11249 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11251 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11252 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11253 decls to at least the GCC 4.8 used alignments.
11256 * tree-nested.c (convert_nonlocal_reference_stmt,
11257 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11258 of GIMPLE_BIND stmts, adjust associated decls.
11260 2014-01-17 Richard Biener <rguenther@suse.de>
11262 PR tree-optimization/46590
11263 * vec.h (vec<>::bseach): New member function implementing
11264 binary search according to C89 bsearch.
11265 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11266 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11267 bitmap pointer again. Make accesses_in_loop a flat array.
11268 (mem_ref_obstack): New global.
11269 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11270 (mark_ref_stored): Likewise.
11271 (ref_indep_loop_p_2): Likewise.
11272 (set_ref_stored_in_loop): New helper function.
11273 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11274 (memref_free): Adjust.
11275 (record_mem_ref_loc): Simplify.
11276 (gather_mem_refs_stmt): Adjust.
11277 (sort_locs_in_loop_postorder_cmp): New function.
11278 (analyze_memory_references): Sort accesses_in_loop after
11279 loop postorder number.
11280 (find_ref_loc_in_loop_cmp): New function.
11281 (for_all_locs_in_loop): Find relevant cluster of locs in
11282 accesses_in_loop and iterate without recursion.
11283 (execute_sm): Avoid uninit warning.
11284 (struct ref_always_accessed): Simplify.
11285 (ref_always_accessed::operator ()): Likewise.
11286 (ref_always_accessed_p): Likewise.
11287 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11288 loop postorder numbers here.
11289 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11292 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11295 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11296 on decls for which assemble_alias has been called.
11298 2014-01-17 Nick Clifton <nickc@redhat.com>
11300 * config/msp430/msp430.opt: (mcpu): New option.
11301 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11302 (msp430_option_override): Parse target_cpu. If the MCU name
11303 matches a generic string, clear target_mcu.
11304 (msp430_attr): Allow numeric interrupt values up to 63.
11305 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11306 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11308 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11310 * config/msp430/msp430.md (popm): Use %J rather than %I.
11311 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11312 (addhi_cy_i): Use immediate_operand for operand 2.
11313 * doc/invoke.texi: Document -mcpu option.
11315 2014-01-17 Richard Biener <rguenther@suse.de>
11317 PR rtl-optimization/38518
11318 * df.h (df_analyze_loop): Declare.
11319 * df-core.c: Include cfgloop.h.
11320 (df_analyze_1): Split out main part of df_analyze.
11321 (df_analyze): Adjust.
11322 (loop_inverted_post_order_compute): New function.
11323 (loop_post_order_compute): Likewise.
11324 (df_analyze_loop): New function avoiding whole-function
11325 postorder computes.
11326 * loop-invariant.c (find_defs): Use df_analyze_loop.
11327 (find_invariants): Adjust.
11328 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11330 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11332 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11333 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11335 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11337 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11338 traversal when removing references.
11340 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11343 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11345 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11347 PR middle-end/56791
11348 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11349 pushing a reload for an autoinc when we had previously reloaded an
11350 inner part of the address.
11352 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11354 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11356 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11357 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11358 when not giving up or versioning for alias only because of
11360 (vect_analyze_data_ref_dependences): Set to true.
11361 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11363 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11364 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11367 PR middle-end/58344
11368 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11371 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11372 operand 0 predicate for gathers, use a new pseudo as subtarget.
11374 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11376 PR middle-end/59609
11377 * lra-constraints.c (process_alt_operands): Add printing debug info.
11378 Check absence of input/output reloads for matched operands too.
11380 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11382 PR rtl-optimization/59835
11383 * ira.c (ira_init_register_move_cost): Increase cost for
11386 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11388 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11390 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11393 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11394 non-register objects. Use gen_(high/low)part more consistently.
11397 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11400 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11401 endian support, remove tests for WORDS_BIG_ENDIAN.
11402 (p8_mfvsrd_3_<mode>): Likewise.
11403 (reload_gpr_from_vsx<mode>): Likewise.
11404 (reload_gpr_from_vsxsf): Likewise.
11405 (p8_mfvsrd_4_disf): Likewise.
11407 2014-01-16 Richard Biener <rguenther@suse.de>
11409 PR rtl-optimization/46590
11410 * lcm.c (compute_antinout_edge): Use postorder iteration.
11411 (compute_laterin): Use inverted postorder iteration.
11413 2014-01-16 Nick Clifton <nickc@redhat.com>
11415 PR middle-end/28865
11416 * varasm.c (output_constant): Return the number of bytes actually
11418 (output_constructor_array_range): Update the field size with the
11419 number of bytes emitted by output_constant.
11420 (output_constructor_regular_field): Likewise. Also do not
11421 complain if the total number of bytes emitted is now greater
11422 than the expected fieldpos.
11423 * output.h (output_constant): Update prototype and descriptive comment.
11425 2014-01-16 Marek Polacek <polacek@redhat.com>
11427 PR middle-end/59827
11428 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11429 it is error_mark_node.
11431 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11433 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11434 VALID_AVX256_REG_OR_OI_MODE.
11436 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11438 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11439 current procedure should be profiled.
11441 2014-01-15 Andrew Pinski <apinski@cavium.com>
11443 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11444 of moving from/to the STACK_REG register class.
11446 2014-01-15 Richard Henderson <rth@redhat.com>
11449 * reginfo.c (global_regs_decl): Globalize.
11450 * rtl.h (global_regs_decl): Declare.
11451 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11452 reserved via global_regs.
11454 2014-01-15 Teresa Johnson <tejohnson@google.com>
11456 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11458 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11460 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11461 and vmulosh rather than call gen_vec_widen_smult_*.
11462 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11463 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11464 (vec_widen_smult_even_v16qi): Likewise.
11465 (vec_widen_umult_even_v8hi): Likewise.
11466 (vec_widen_smult_even_v8hi): Likewise.
11467 (vec_widen_umult_odd_v16qi): Likewise.
11468 (vec_widen_smult_odd_v16qi): Likewise.
11469 (vec_widen_umult_odd_v8hi): Likewise.
11470 (vec_widen_smult_odd_v8hi): Likewise.
11471 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11472 vmuloub rather than call gen_vec_widen_umult_*.
11473 (vec_widen_umult_lo_v16qi): Likewise.
11474 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11475 vmulosb rather than call gen_vec_widen_smult_*.
11476 (vec_widen_smult_lo_v16qi): Likewise.
11477 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11478 rather than call gen_vec_widen_umult_*.
11479 (vec_widen_umult_lo_v8hi): Likewise.
11480 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11481 rather than call gen_vec_widen_smult_*.
11482 (vec_widen_smult_lo_v8hi): Likewise.
11484 2014-01-15 Jeff Law <law@redhat.com>
11486 PR tree-optimization/59747
11487 * ree.c (find_and_remove_re): Properly handle case where a second
11488 eliminated extension requires widening a copy created for elimination
11489 of a prior extension.
11490 (combine_set_extension): Ensure that the number of hard regs needed
11491 for a destination register does not change when we widen it.
11493 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11495 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11496 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11497 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11498 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11499 (avr-*-rtems*): Likewise.
11500 (bfin*-rtems*): Likewise.
11501 (moxie-*-rtems*): Likewise.
11502 (h8300-*-rtems*): Likewise.
11503 (i[34567]86-*-rtems*): Likewise.
11504 (lm32-*-rtems*): Likewise.
11505 (m32r-*-rtems*): Likewise.
11506 (m68k-*-rtems*): Likewise.
11507 (microblaze*-*-rtems*): Likewise.
11508 (mips*-*-rtems*): Likewise.
11509 (powerpc-*-rtems*): Likewise.
11510 (sh-*-rtems*): Likewise.
11511 (sparc-*-rtems*): Likewise.
11512 (sparc64-*-rtems*): Likewise.
11513 (v850-*-rtems*): Likewise.
11514 (m32c-*-rtems*): Likewise.
11516 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11518 PR rtl-optimization/59511
11519 * ira.c (ira_init_register_move_cost): Use memory costs for some
11520 cases of register move cost calculations.
11521 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11522 instead of BB frequency.
11523 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11524 * lra-assigns.c (find_hard_regno_for): Ditto.
11526 2014-01-15 Richard Biener <rguenther@suse.de>
11528 PR tree-optimization/59822
11529 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11530 (vectorizable_load): Use it to hoist defs of uses of invariant
11531 loads out of the loop.
11533 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11534 Kugan Vivekanandarajah <kuganv@linaro.org>
11537 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11540 2014-01-15 Richard Biener <rguenther@suse.de>
11542 PR rtl-optimization/59802
11543 * lcm.c (compute_available): Use inverted postorder to seed
11544 the initial worklist.
11546 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11549 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11550 ADDR_REGS for invalid symrefs in non-PIC code.
11552 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11555 * builtins.c (determine_block_size): Initialize *probable_max_size
11556 even if len_rtx is CONST_INT.
11558 2014-01-14 Andrew Pinski <apinski@cavium.com>
11560 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11561 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11562 (cortexa53_tunings): Likewise.
11563 (aarch64_sched_issue_rate): New function.
11564 (TARGET_SCHED_ISSUE_RATE): Define.
11566 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11568 * ira-costs.c (find_costs_and_classes): Add missed
11569 ira_init_register_move_cost_if_necessary.
11571 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11574 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11576 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11579 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11580 to indicate if type is used for function return value. Warn ABI
11581 change if the vector mode isn't available for function return value.
11582 (ix86_function_arg_advance): Pass false to type_natural_mode.
11583 (ix86_function_arg): Likewise.
11584 (ix86_gimplify_va_arg): Likewise.
11585 (function_arg_32): Don't warn ABI change.
11586 (ix86_function_value): Pass true to type_natural_mode.
11587 (ix86_return_in_memory): Likewise.
11588 (ix86_struct_value_rtx): Removed.
11589 (TARGET_STRUCT_VALUE_RTX): Likewise.
11591 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11593 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11594 converting a conditional jump into a conditional return.
11596 2014-01-14 Richard Biener <rguenther@suse.de>
11598 PR tree-optimization/58921
11599 PR tree-optimization/59006
11600 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11601 hoisting invariant stmts.
11602 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11603 invariant loads on the preheader edge if possible.
11605 2014-01-14 Joey Ye <joey.ye@arm.com>
11607 * doc/plugin.texi (Building GCC plugins): Update to C++.
11609 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11611 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11612 (_mm_rcp28_round_ss): Ditto.
11613 (_mm_rsqrt28_round_sd): Ditto.
11614 (_mm_rsqrt28_round_ss): Ditto.
11615 (_mm_rcp28_sd): Ditto.
11616 (_mm_rcp28_ss): Ditto.
11617 (_mm_rsqrt28_sd): Ditto.
11618 (_mm_rsqrt28_ss): Ditto.
11619 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11620 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11621 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11622 (IX86_BUILTIN_RCP28SD): Ditto.
11623 (IX86_BUILTIN_RCP28SS): Ditto.
11624 (IX86_BUILTIN_RSQRT28SD): Ditto.
11625 (IX86_BUILTIN_RSQRT28SS): Ditto.
11626 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11627 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11628 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11629 (ix86_expand_special_args_builtin): Expand new FTYPE.
11630 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11631 (srcp14<mode>): Make insn unary.
11632 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11633 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11634 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11635 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11636 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11637 Fix rounding: make it SAE only.
11638 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11640 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11642 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11643 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11644 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11645 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11646 (round_saeonly_mask_scalar_operand4): Ditto.
11647 (round_saeonly_mask_scalar_op3): Ditto.
11648 (round_saeonly_mask_scalar_op4): Ditto.
11650 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11652 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11653 Implement -maltivec=be for vec_insert and vec_extract.
11655 2014-01-10 DJ Delorie <dj@redhat.com>
11657 * config/msp430/msp430.md (call_internal): Don't allow memory
11658 references with SP as the base register.
11659 (call_value_internal): Likewise.
11660 * config/msp430/constraints.md (Yc): New. For memory references
11661 that don't use SP as a base register.
11663 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11664 "an integer without a # prefix"
11665 * config/msp430/msp430.md (epilogue_helper): Use it.
11667 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11670 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11671 AVX512F gather builtins.
11672 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11673 on gather decls with INTEGER_TYPE masktype.
11674 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11675 directly into the builtin rather than hoisting it before loop.
11677 PR tree-optimization/59387
11678 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11679 (scev_const_prop): If folded_casts and type has undefined overflow,
11680 use force_gimple_operand instead of force_gimple_operand_gsi and
11681 for each added stmt if it is assign with
11682 arith_code_with_undefined_signed_overflow, call
11683 rewrite_to_defined_overflow.
11684 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11685 gimple-fold.h instead.
11686 (arith_code_with_undefined_signed_overflow,
11687 rewrite_to_defined_overflow): Moved to ...
11688 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11689 rewrite_to_defined_overflow): ... here. No longer static.
11690 Include gimplify-me.h.
11691 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11692 rewrite_to_defined_overflow): New prototypes.
11694 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11696 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11698 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11700 * builtins.c (get_object_alignment_2): Minor tweak.
11701 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11703 2014-01-13 Christian Bruel <christian.bruel@st.com>
11705 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11706 optimized non constant lengths.
11708 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11711 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11712 load as __atomic_load_N if possible.
11714 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11716 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11718 (rs6000_expand_builtin): Adjust call.
11720 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11723 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11724 * config/rs6000/rs6000.c: Include target-globals.h.
11725 (rs6000_set_current_function): Instead of doing target_reinit
11726 unconditionally, use save_target_globals_default_opts and
11727 restore_target_globals.
11729 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11731 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11732 (rs6000_expand_builtin): Handle mffs and mtfsf.
11733 (rs6000_init_builtins): Define mffs and mtfsf.
11734 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11735 (rs6000_mffs): New pattern.
11736 (rs6000_mtfsf): New pattern.
11738 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11740 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11741 Start narrowing with START. Apply candidate-use pair
11742 and check overall cost in narrowing.
11743 (iv_ca_prune): Pass new argument.
11745 2014-01-10 Jeff Law <law@redhat.com>
11747 PR middle-end/59743
11748 * ree.c (combine_reaching_defs): Ensure the defining statement
11749 occurs before the extension when optimizing extensions with
11750 different source and destination hard registers.
11752 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11755 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11756 vtables into the type inheritance graph.
11758 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11760 PR rtl-optimization/59754
11761 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11762 modes in the REGNO != REGNO case.
11764 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11766 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11768 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11770 PR tree-optimization/59745
11771 * tree-predcom.c (tree_predictive_commoning_loop): Call
11772 free_affine_expand_cache if giving up because components is NULL.
11774 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11775 GC in payload of target_globals struct instead of allocating them on
11776 the heap and the larger structs separately using GC.
11777 * target-globals.h (struct target_globals): Make regs, hard_regs,
11778 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11779 of GTY((skip)) and change type to void *.
11780 (reset_target_globals): Cast loads from those fields to corresponding
11783 2014-01-10 Steve Ellcey <sellcey@mips.com>
11786 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11787 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11788 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11790 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11793 * aarch64-modes.def (CC_Zmode): New flags mode.
11794 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11795 represents an equality.
11796 (aarch64_get_condition_code): Handle CC_Zmode.
11797 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11799 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11801 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11802 extraction in good case.
11804 2014-01-10 Richard Biener <rguenther@suse.de>
11806 PR tree-optimization/59374
11807 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11808 checking after SLP discovery. Mark stmts not participating
11809 in any SLP instance properly.
11811 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11813 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11814 when handling a SET rtx.
11816 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11818 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11819 (cortex-a57): Likewise.
11820 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11822 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11824 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11825 non-iwmmxt builtins.
11827 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11831 * ipa-devirt.c record_target_from_binfo): Take as argument
11832 stack of binfos and lookup matching one for virtual inheritance.
11833 (possible_polymorphic_call_targets_1): Update.
11835 2014-01-10 Huacai Chen <chenhc@lemote.com>
11837 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11838 kernel strings for Loongson-2E/2F/3A.
11840 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11842 PR middle-end/59670
11843 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11844 is_gimple_call before calling gimple_call_internal_p.
11846 2014-01-09 Steve Ellcey <sellcey@mips.com>
11848 * Makefile.in (TREE_FLOW_H): Remove.
11849 (TREE_SSA_H): Add file names from tree-flow.h.
11850 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11851 * tree.h: Remove tree-flow.h reference.
11852 * hash-table.h: Remove tree-flow.h reference.
11853 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11854 reference with tree-ssa-loop.h.
11856 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11858 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11859 default element-order behavior for -maltivec.
11860 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11861 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11862 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11863 when targeting big endian, at least for now.
11864 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11866 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11868 PR middle-end/47735
11869 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11870 var satisfies use_register_for_decl, just take into account type
11871 alignment, rather than decl alignment.
11873 PR tree-optimization/59622
11874 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11875 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11876 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11877 Don't devirtualize for inplace at all. For targets.length () == 1,
11878 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11880 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11882 * config/i386/i386.md (cpu): Remove the unused btver1.
11884 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11886 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11888 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11891 * tree-core.h (struct target_globals): New forward declaration.
11892 (struct tree_target_option): Add globals field.
11893 * tree.h (TREE_TARGET_GLOBALS): Define.
11894 (prepare_target_option_nodes_for_pch): New prototype.
11895 * target-globals.h (struct target_globals): Define even if
11896 !SWITCHABLE_TARGET.
11897 * tree.c (prepare_target_option_node_for_pch,
11898 prepare_target_option_nodes_for_pch): New functions.
11899 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11900 * config/i386/i386.c: Include target-globals.h.
11901 (ix86_set_current_function): Instead of doing target_reinit
11902 unconditionally, use save_target_globals_default_opts and
11903 restore_target_globals.
11905 2014-01-09 Richard Biener <rguenther@suse.de>
11907 PR tree-optimization/59715
11908 * tree-cfg.h (split_critical_edges): Declare.
11909 * tree-cfg.c (split_critical_edges): Export.
11910 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11912 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11914 * cfgexpand.c (expand_stack_vars): Optionally disable
11915 asan stack protection.
11916 (expand_used_vars): Likewise.
11917 (partition_stack_vars): Likewise.
11918 * asan.c (asan_emit_stack_protection): Optionally disable
11919 after return stack usage.
11920 (instrument_derefs): Optionally disable memory access instrumentation.
11921 (instrument_builtin_call): Likewise.
11922 (instrument_strlen_call): Likewise.
11923 (asan_protect_global): Optionally disable global variables protection.
11924 * doc/invoke.texi: Added doc for new options.
11925 * params.def: Added new options.
11926 * params.h: Likewise.
11928 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11930 PR rtl-optimization/59724
11931 * ifcvt.c (cond_exec_process_if_block): Don't call
11932 flow_find_head_matching_sequence with 0 longest_match.
11933 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11934 non-active insns if !stop_after.
11935 (try_head_merge_bb): Revert 2014-01-07 changes.
11937 2014-01-08 Jeff Law <law@redhat.com>
11939 * ree.c (get_sub_rtx): New function, extracted from...
11940 (merge_def_and_ext): Here.
11941 (combine_reaching_defs): Use get_sub_rtx.
11943 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11945 * cgraph.h (varpool_variable_node): Do not choke on null node.
11947 2014-01-08 Catherine Moore <clm@codesourcery.com>
11949 * config/mips/mips.md (simple_return): Attempt to use JRC
11951 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11953 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11955 PR rtl-optimization/59137
11956 * reorg.c (steal_delay_list_from_target): Call update_block for
11958 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11960 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11962 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11963 two duplicate entries.
11965 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11968 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11970 * config/mips/mips.c (mips_truncated_op_cost): New function.
11971 (mips_rtx_costs): Adjust test for BADDU.
11972 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11974 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11976 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11977 (*baddu_si): ...this new pattern.
11979 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11982 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11984 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11986 PR middle-end/57748
11987 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11989 (expand_expr, expand_normal): Adjust.
11990 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11991 inner_reference_p. Use inner_reference_p to expand inner references.
11992 (store_expr): Adjust.
11993 * cfgexpand.c (expand_call_stmt): Adjust.
11995 2014-01-08 Rong Xu <xur@google.com>
11997 * gcov-io.c (gcov_var): Move from gcov-io.h.
11998 (gcov_position): Ditto.
11999 (gcov_is_error): Ditto.
12000 (gcov_rewrite): Ditto.
12001 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12002 only part to libgcc/libgcov.h.
12004 2014-01-08 Marek Polacek <polacek@redhat.com>
12006 PR middle-end/59669
12007 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12009 2014-01-08 Marek Polacek <polacek@redhat.com>
12012 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12014 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12016 PR rtl-optimization/59649
12017 * stor-layout.c (get_mode_bounds): For BImode return
12018 0 and STORE_FLAG_VALUE.
12020 2014-01-08 Richard Biener <rguenther@suse.de>
12022 PR middle-end/59630
12023 * gimple.h (is_gimple_builtin_call): Remove.
12024 (gimple_builtin_call_types_compatible_p): New.
12025 (gimple_call_builtin_p): New overload.
12026 * gimple.c (is_gimple_builtin_call): Remove.
12027 (validate_call): Rename to ...
12028 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12029 check return types.
12030 (validate_type): New static function.
12031 (gimple_call_builtin_p): New overload and adjust.
12032 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12033 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12034 (gimple_fold_stmt_to_constant_1): Likewise.
12035 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12037 2014-01-08 Richard Biener <rguenther@suse.de>
12039 PR middle-end/59471
12040 * gimplify.c (gimplify_expr): Gimplify register-register type
12041 VIEW_CONVERT_EXPRs to separate stmts.
12043 2014-01-07 Jeff Law <law@redhat.com>
12045 PR middle-end/53623
12046 * ree.c (combine_set_extension): Handle case where source
12047 and destination registers in an extension insn are different.
12048 (combine_reaching_defs): Allow source and destination registers
12049 in extension to be different under limited circumstances.
12050 (add_removable_extension): Remove restriction that the
12051 source and destination registers in the extension are the same.
12052 (find_and_remove_re): Emit a copy from the extension's
12053 destination to its source after the defining insn if
12054 the source and destination registers are different.
12056 PR middle-end/59285
12057 * ifcvt.c (merge_if_block): If we are merging a block with more than
12058 one successor with a block with no successors, remove any BARRIER
12059 after the second block.
12061 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12063 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12065 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12068 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12069 for 14-bit register offsets when INT14_OK_STRICT is false.
12071 2014-01-07 Roland Stigge <stigge@antcom.de>
12072 Michael Meissner <meissner@linux.vnet.ibm.com>
12075 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12076 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12078 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12080 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12083 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12085 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12086 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12087 rtx is const0_rtx or not.
12089 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12092 * target-globals.c (save_target_globals): Remove this_fn_optab
12094 * toplev.c: Include optabs.h.
12095 (target_reinit): Temporarily restore the global options if another
12096 set of options are in force.
12098 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12100 PR rtl-optimization/58668
12101 * cfgcleanup.c (flow_find_cross_jump): Don't count
12102 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12103 to determine what is counted.
12104 (flow_find_head_matching_sequence): Use active_insn_p to determine
12106 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12108 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12109 determine what is counted.
12111 PR tree-optimization/59643
12112 * tree-predcom.c (split_data_refs_to_components): If one dr is
12113 read and one write, determine_offset fails and the write isn't
12114 in the bad component, just put the read into the bad component.
12116 2014-01-07 Mike Stump <mikestump@comcast.net>
12117 Jakub Jelinek <jakub@redhat.com>
12120 * tree-core.h (struct tree_optimization_option): Change optabs
12121 type from unsigned char * to void *.
12122 * optabs.c (init_tree_optimization_optabs): Adjust
12123 TREE_OPTIMIZATION_OPTABS initialization.
12125 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12128 * config/i386/i386.h (struct machine_function): Add
12129 no_drap_save_restore field.
12130 * config/i386/i386.c (ix86_save_reg): Use
12131 !cfun->machine->no_drap_save_restore instead of
12132 crtl->stack_realign_needed.
12133 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12134 this function clears frame_pointer_needed. Set
12135 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12136 and DRAP reg is needed.
12138 2014-01-06 Marek Polacek <polacek@redhat.com>
12141 * doc/implement-c.texi: Mention that other integer types are
12142 permitted as bit-field types in strictly conforming mode.
12144 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12146 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12147 is newly allocated.
12149 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12151 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12153 2014-01-06 Martin Jambor <mjambor@suse.cz>
12156 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12158 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12160 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12164 * var-tracking.c (add_stores): Preserve the value of the source even if
12165 we don't record the store.
12167 2014-01-06 Terry Guo <terry.guo@arm.com>
12169 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12171 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12174 * config/darwin.c (darwin_function_section): Adjust return values to
12175 correspond to optimisation changes made in r206070.
12177 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12179 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12180 from prefetch_block tune setting.
12181 (nocona_cost): Correct size of prefetch block to 64.
12183 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12185 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12186 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12187 used to save the static chain register in the computation of the offset
12188 from which the FP registers need to be restored.
12190 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12192 PR tree-optimization/59519
12193 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12194 ICE if get_current_def (current_new_name) is already non-NULL, as long
12195 as it is a phi result of some other phi in *new_exit_bb that has
12198 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12199 or vmovdqu* for misaligned_operand.
12200 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12201 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12202 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12203 aligned_mem for AVX512F masked aligned load and store builtins and for
12204 non-temporal moves.
12206 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12208 PR tree-optimization/59651
12209 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12210 Address range for negative step should be added by TYPE_SIZE_UNIT.
12212 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12214 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12215 overlapping registers also for registers other than the stack pointer.
12217 2014-01-03 Marek Polacek <polacek@redhat.com>
12220 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12223 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12226 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12229 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12230 (push splitter): Use <P:MODE_SIZE> instead of
12231 GET_MODE_SIZE (<P:MODE>mode).
12232 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12233 (mov -1, reg peephole2): Likewise.
12234 * config/i386/sse.md (*mov<mode>_internal,
12235 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12236 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12237 *<code><mode>3, *andnot<mode>3<mask_name>,
12238 <mask_codefor><code><mode>3<mask_name>): Likewise.
12239 * config/i386/subst.md (mask_mode512bit_condition,
12240 sd_mask_mode512bit_condition): Likewise.
12242 2014-01-02 Xinliang David Li <davidxl@google.com>
12244 PR tree-optimization/59303
12245 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12246 (dump_predicates): Better output format.
12247 (pred_equal_p): New function.
12248 (is_neq_relop_p): Ditto.
12249 (is_neq_zero_form_p): Ditto.
12250 (pred_expr_equal_p): Ditto.
12251 (pred_neg_p): Ditto.
12252 (simplify_pred): Ditto.
12253 (simplify_preds_2): Ditto.
12254 (simplify_preds_3): Ditto.
12255 (simplify_preds_4): Ditto.
12256 (simplify_preds): Ditto.
12257 (push_pred): Ditto.
12258 (push_to_worklist): Ditto.
12259 (get_pred_info_from_cmp): Ditto.
12260 (is_degenerated_phi): Ditto.
12261 (normalize_one_pred_1): Ditto.
12262 (normalize_one_pred): Ditto.
12263 (normalize_one_pred_chain): Ditto.
12264 (normalize_preds): Ditto.
12265 (normalize_cond_1): Remove function.
12266 (normalize_cond): Ditto.
12267 (is_gcond_subset_of): Ditto.
12268 (is_subset_of_any): Ditto.
12269 (is_or_set_subset_of): Ditto.
12270 (is_and_set_subset_of): Ditto.
12271 (is_norm_cond_subset_of): Ditto.
12272 (pred_chain_length_cmp): Ditto.
12273 (convert_control_dep_chain_into_preds): Type change.
12274 (find_predicates): Ditto.
12275 (find_def_preds): Ditto.
12276 (destroy_predicates_vecs): Ditto.
12277 (find_matching_predicates_in_rest_chains): Ditto.
12278 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12279 (is_pred_expr_subset): Ditto.
12280 (is_pred_chain_subset_of): Ditto.
12281 (is_included_in): Ditto.
12282 (is_superset_of): Ditto.
12284 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12286 Update copyright years.
12288 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12290 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12291 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12292 config/arc/arc.md, config/arc/arc.opt,
12293 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12294 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12295 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12296 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12297 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12298 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12299 vtable-verify.c, vtable-verify.h: Use the standard form for the
12302 2014-01-02 Tobias Burnus <burnus@net-b.de>
12304 * gcc.c (process_command): Update copyright notice dates.
12305 * gcov-dump.c: Ditto.
12307 * doc/cpp.texi: Bump @copying's copyright year.
12308 * doc/cppinternals.texi: Ditto.
12309 * doc/gcc.texi: Ditto.
12310 * doc/gccint.texi: Ditto.
12311 * doc/gcov.texi: Ditto.
12312 * doc/install.texi: Ditto.
12313 * doc/invoke.texi: Ditto.
12315 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12317 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12319 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12321 * config/i386/sse.md (*mov<mode>_internal): Guard
12322 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12324 PR rtl-optimization/59647
12325 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12326 new_rtx into UNSIGNED_FLOAT rtxes.
12328 Copyright (C) 2014 Free Software Foundation, Inc.
12330 Copying and distribution of this file, with or without modification,
12331 are permitted in any medium without royalty provided the copyright
12332 notice and this notice are preserved.