1 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
3 * config/rs6000/440.md (ppc440-integer): Include shift without
5 (ppc440-compare): Include shift with dot.
6 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without
8 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
10 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
12 (e6500_sfx2): Include it.
13 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
14 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
15 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
16 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
17 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
18 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
19 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
20 *lshiftrt_internal1le, *lshiftrt_internal1be,
21 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
22 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
23 *rotldi3_internal10le, *rotldi3_internal10be,
24 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
25 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
26 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
27 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
28 define_insns): Use type "shift" in the appropriate alternatives.
30 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
32 * config/rs6000/rs6000.md (type): Add "logical". Delete
34 (dot): Adjust comment.
35 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
36 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
37 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
38 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
39 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
40 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
41 *mov<mode>_internal2, and 10 anonymous define_insns): Use
43 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
45 * config/rs6000/40x.md: (ppc403-integer, ppc403-compare): Adjust.
46 * config/rs6000/440.md: (ppc440-integer, ppc440-compare): Adjust.
47 * config/rs6000/476.md: (ppc476-simple-integer, ppc476-compare):
49 * config/rs6000/603.md: (ppc603-integer, ppc603-compare): Adjust.
50 * config/rs6000/6xx.md: (ppc604-integer, ppc604-compare): Adjust.
51 * config/rs6000/7450.md: (ppc7450-integer, ppc7450-compare):
53 * config/rs6000/7xx.md: (ppc750-integer, ppc750-compare): Adjust.
54 * config/rs6000/8540.md: (ppc8540_su): Adjust.
55 * config/rs6000/cell.md: (cell-integer, cell-fast-cmp,
56 cell-cmp-microcoded): Adjust.
57 * config/rs6000/e300c2c3.md: (ppce300c3_cmp, ppce300c3_iu):
59 * config/rs6000/e500mc.md: (e500mc_su): Adjust.
60 * config/rs6000/e500mc64.md: (e500mc64_su, e500mc64_su2): Adjust.
61 * config/rs6000/e5500.md: (e5500_sfx, e5500_sfx2): Adjust.
62 * config/rs6000/e6500.md: (e6500_sfx, e6500_sfx2): Adjust.
63 * config/rs6000/mpc.md: (mpccore-integer, mpccore-compare):
65 * config/rs6000/power4.md: (power4-integer, power4-cmp): Adjust.
66 * config/rs6000/power5.md: (power5-integer, power5-cmp): Adjust.
67 * config/rs6000/power6.md: (power6-integer, power6-fast-compare):
69 * config/rs6000/power7.md: (power7-integer, power7-cmp): Adjust.
70 * config/rs6000/power8.md: (power8-1cyc, power8-fast-compare):
71 Adjust. Adjust comment.
72 * config/rs6000/rs64.md: (rs64a-integer, rs64a-compare): Adjust.
73 * config/rs6000/titan.md: (titan_fxu_adder, titan_fxu_alu):
76 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
78 * config/rs6000/rs6000.md (type): Add "add".
79 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
80 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
81 define_insns): Use it.
82 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
84 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
85 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
86 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare):
88 * config/rs6000/601.md (ppc601-integer): Adjust.
89 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
90 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
91 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
92 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
93 * config/rs6000/8540.md (ppc8540_su): Adjust.
94 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
95 cell-cmp-microcoded): Adjust.
96 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
97 * config/rs6000/e500mc.md (e500mc_su): Adjust.
98 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
99 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
100 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
101 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
102 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
103 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
104 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
106 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
107 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
109 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
110 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
112 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
114 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
115 "delayed_compare", "var_delayed_compare".
116 (var_shift): New attribute.
117 (cell_micro): Adjust.
118 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
119 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
120 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
121 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
122 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
123 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
124 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
125 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
126 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
127 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
128 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
129 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
130 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
131 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
132 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
133 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
134 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
135 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
136 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
137 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
138 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
139 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
140 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
141 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
142 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
144 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
145 * config/rs6000/440.md (ppc440-integer): Adjust.
146 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare):
148 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
149 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
150 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
151 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare):
153 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
154 * config/rs6000/8540.md (ppc8540_su): Adjust.
155 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
156 cell-cmp-microcoded): Adjust.
157 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
158 * config/rs6000/e500mc.md (e500mc_su): Adjust.
159 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
160 e500mc64_delayed): Adjust.
161 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
162 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
163 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
164 * config/rs6000/power4.md (power4-integer, power4-compare):
166 * config/rs6000/power5.md (power5-integer, power5-compare):
168 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
169 power6-delayed-compare, power6-var-delayed-compare): Adjust.
170 * config/rs6000/power7.md (power7-integer, power7-compare):
172 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
174 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
175 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
177 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
179 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add
181 (bits): New mode_attr.
182 (idiv_ldiv): Delete mode_attr.
183 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
184 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
185 rs6000_adjust_priority, is_nonpipeline_insn,
186 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
188 * config/rs6000/40x.md (ppc403-idiv): Adjust.
189 * config/rs6000/440.md (ppc440-idiv): Adjust.
190 * config/rs6000/476.md (ppc476-idiv): Adjust.
191 * config/rs6000/601.md (ppc601-idiv): Adjust.
192 * config/rs6000/603.md (ppc603-idiv): Adjust.
193 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
194 ppc620-ldiv): Adjust.
195 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
196 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
197 * config/rs6000/8540.md (ppc8540_divide): Adjust.
198 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
199 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
200 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
201 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
202 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
203 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
204 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
205 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
206 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
207 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
208 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
209 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
210 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
211 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
212 * config/rs6000/titan.md (titan_fxu_div): Adjust.
214 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
216 * config/rs6000/rs6000.md (type): Delete "insert_word",
217 "insert_dword". Add "insert".
218 (size): Update comment.
219 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
220 insn_must_be_first_in_group): Adjust.
221 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
222 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
223 *insvsi_internal6, insvdi_internal): Adjust.
225 * config/rs6000/40x.md (ppc403-integer): Adjust.
226 * config/rs6000/440.md (ppc440-integer): Adjust.
227 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
228 * config/rs6000/601.md (ppc601-integer): Adjust.
229 * config/rs6000/603.md (ppc603-integer): Adjust.
230 * config/rs6000/6xx.md (ppc604-integer): Adjust.
231 * config/rs6000/7450.md (ppc7450-integer): Adjust.
232 * config/rs6000/7xx.md (ppc750-integer): Adjust.
233 * config/rs6000/8540.md (ppc8540_su): Adjust.
234 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
235 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
236 * config/rs6000/e500mc.md (e500mc_su): Adjust.
237 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
238 * config/rs6000/e5500.md (e5500_sfx): Adjust.
239 * config/rs6000/e6500.md (e6500_sfx): Adjust.
240 * config/rs6000/mpc.md (mpccore-integer): Adjust.
241 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
242 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
243 * config/rs6000/power6.md (power6-insert, power6-insert-dword):
245 * config/rs6000/power7.md (power7-integer): Adjust.
246 * config/rs6000/power8.md (power8-1cyc): Adjust.
247 * config/rs6000/rs64.md (rs64a-integer): Adjust.
248 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
250 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
252 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
253 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
254 (size): New attribute.
255 (dot): New attribute.
256 (cell_micro): Adjust.
257 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
258 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
259 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
260 umuldi3_highpart): Adjust.
261 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
262 rs6000_adjust_priority, is_nonpipeline_insn,
263 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
265 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
266 ppc405-imul3): Adjust.
267 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
268 * config/rs6000/476.md (ppc476-imul): Adjust.
269 * config/rs6000/601.md (ppc601-imul): Adjust.
270 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
271 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
272 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
273 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
274 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
276 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
277 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
278 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
280 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
281 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
282 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
283 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
284 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
285 * config/rs6000/mpc.md (mpccore-imul): Adjust.
286 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
287 power4-lmul, power4-imul, power4-imul3): Adjust.
288 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
289 power5-lmul, power5-imul, power5-imul3): Adjust.
290 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
291 power6-lmul, power6-imul, power6-imul3): Adjust.
292 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
293 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
295 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
297 * config/rs6000/titan.md (titan_imul): Adjust.
299 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
301 * config/rs6000/rs6000.md (type): Add new value "halfmul".
302 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
303 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
304 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
305 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
306 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu):
308 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
309 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
310 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
311 * config/rs6000/titan.md: Delete nonsensical comment.
312 (titan_imul): Add type imul3.
313 (titan_mulhw): Remove type imul3; add type halfmul.
315 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
317 * config/rs6000/rs6000.md (type): Reorder, reformat.
319 2014-05-23 Martin Jambor <mjambor@suse.cz>
321 PR tree-optimization/53787
322 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
323 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
324 analysis_done, update all uses.
325 * ipa-prop.c: Include domwalk.h
326 (param_analysis_info): Removed.
327 (param_aa_status): New type.
328 (ipa_bb_info): Likewise.
329 (func_body_info): Likewise.
330 (ipa_get_bb_info): New function.
331 (aa_overwalked): Likewise.
332 (find_dominating_aa_status): Likewise.
333 (parm_bb_aa_status_for_bb): Likewise.
334 (parm_preserved_before_stmt_p): Changed to use new param AA info.
335 (load_from_unmodified_param): Accept func_body_info as a parameter
336 instead of parms_ainfo.
337 (parm_ref_data_preserved_p): Changed to use new param AA info.
338 (parm_ref_data_pass_through_p): Likewise.
339 (ipa_load_from_parm_agg_1): Likewise. Update callers.
340 (compute_complex_assign_jump_func): Changed to use new param AA info.
341 (compute_complex_ancestor_jump_func): Likewise.
342 (ipa_compute_jump_functions_for_edge): Likewise.
343 (ipa_compute_jump_functions): Removed.
344 (ipa_compute_jump_functions_for_bb): New function.
345 (ipa_analyze_indirect_call_uses): Likewise, moved variable
347 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
348 and info, moved variable declarations down.
349 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
351 (ipa_analyze_stmt_uses): Likewise.
352 (ipa_analyze_params_uses): Removed.
353 (ipa_analyze_params_uses_in_bb): New function.
354 (ipa_analyze_controlled_uses): Likewise.
355 (free_ipa_bb_info): Likewise.
356 (analysis_dom_walker): New class.
357 (ipa_analyze_node): Handle node-specific forbidden analysis,
358 initialize and free func_body_info, use dominator walker.
359 (ipcp_modif_dom_walker): New class.
360 (ipcp_transform_function): Create and free func_body_info, use
361 ipcp_modif_dom_walker, moved a lot of functionality there.
363 2014-05-23 Marek Polacek <polacek@redhat.com>
364 Jakub Jelinek <jakub@redhat.com>
366 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
367 * gcc.c (sanitize_spec_function): Likewise.
368 * convert.c (convert_to_integer): Include "ubsan.h". Add
369 floating-point to integer instrumentation.
370 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
371 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
373 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
374 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
375 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
376 * ubsan.c: Include "realmpfr.h" and "dfp.h".
377 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
378 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
379 float/double/long double.
380 (ubsan_instrument_float_cast): New function.
381 * ubsan.h (ubsan_instrument_float_cast): Declare.
383 2014-05-23 Jiong Wang <jiong.wang@arm.com>
385 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
387 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
388 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
389 Adjust for tailcalling through registers.
390 * config/aarch64/aarch64.h (enum reg_class): New caller save
392 (REG_CLASS_NAMES): Likewise.
393 (REG_CLASS_CONTENTS): Likewise.
394 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
395 Allow tailcalling without decls.
397 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
399 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
400 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
402 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
403 gsi, and variables v_* to v*.
405 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
407 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
409 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
411 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
412 * omp-low.c: Update accordingly.
414 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
415 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
416 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
417 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
418 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
419 GF_OMP_TARGET_KIND_UPDATE.
421 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
422 Explicitly enumerate the expected region types.
424 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
427 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
428 documentation; the old documentation didn't clearly state the
429 constraints on the contents of the pointed-to storage.
431 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
433 Fix bootstrap error on ia64
434 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
435 Return default value.
437 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
439 PR tree-optimization/54733
440 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
442 (find_bswap_or_nop_load): New.
443 (find_bswap_1): Renamed to ...
444 (find_bswap_or_nop_1): This. Also add support for memory source.
445 (find_bswap): Renamed to ...
446 (find_bswap_or_nop): This. Also add support for memory source and
447 detection of bitwise operations equivalent to load in host endianness.
448 (execute_optimize_bswap): Likewise. Also move its leading comment back
449 in place and split statement transformation into ...
450 (bswap_replace): This.
452 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
454 PR rtl-optimization/61215
455 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
456 simplify_gen_subreg until final substitution.
458 2014-05-23 Alan Modra <amodra@gmail.com>
461 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
462 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
463 Use "Y" constraint rather than "m".
465 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
467 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
469 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
470 New function declaration.
471 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
472 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
473 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
474 (aarch64_init_builtins) : Initialize builtins
475 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
476 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
477 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
478 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
479 and __builtins_aarch64_set_fpsr.
480 (aarch64_atomic_assign_expand_fenv): New function.
481 * config/aarch64/aarch64.md (set_fpcr): New pattern.
482 (get_fpcr) : Likewise.
483 (set_fpsr) : Likewise.
484 (get_fpsr) : Likewise.
485 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
486 and UNSPECV_SET_FPSR.
487 * doc/extend.texi (AARCH64 Built-in Functions) : Document
488 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
489 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
491 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
493 PR rtl-optimization/60969
494 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
495 constraints. Set up mem cost for NO_REGS case.
497 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
499 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
501 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
503 * config/darwin.c: Include "lto-section-names.h".
504 (LTO_SEGMENT_NAME): Don't define.
505 * config/i386/winnt.c: Include "lto-section-names.h".
506 * lto-streamer.c: Include "lto-section-names.h".
507 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
508 * lto-wrapper.c: Include "lto-section-names.h".
509 (LTO_SECTION_NAME_PREFIX): Don't define.
510 * lto-section-names.h: New file.
511 * cgraphunit.c: Include "lto-section-names.h".
513 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
515 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
517 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
520 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
522 2014-05-22 Nick Clifton <nickc@redhat.com>
524 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
527 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
529 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
530 -> (T)A transformation to integer types.
532 2014-05-22 Teresa Johnson <tejohnson@google.com>
534 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
535 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
536 (gcov_rewrite): Use gcov_nonruntime_assert.
538 (gcov_write_words): Ditto.
539 (gcov_write_length): Ditto.
540 (gcov_read_words): Use gcov_nonruntime_assert, and remove
541 gcc_assert from IN_LIBGCOV code.
542 (gcov_read_summary): Use gcov_error to flag profile corruption.
543 (gcov_sync): Use gcov_nonruntime_assert.
544 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
545 (gcov_histo_index): Use gcov_nonruntime_assert.
546 (static void gcov_histogram_merge): Ditto.
547 (compute_working_sets): Ditto.
548 * gcov-io.h (gcov_nonruntime_assert): Define.
549 (gcov_error): Define for !IN_LIBGCOV
551 2014-05-22 Richard Biener <rguenther@suse.de>
553 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
554 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
555 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
556 and deallocation site.
557 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
558 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
559 passing through the incoming points-to set.
560 (handle_lhs_call): Use flags argument instead of recomputing it.
561 (find_func_aliases_for_call): Call handle_lhs_call with proper
564 2014-05-22 Jakub Jelinek <jakub@redhat.com>
566 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
567 all padding bits in REAL_VALUE_TYPE are cleared.
569 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
571 Cleanup and improve multipass_dfa_lookahead_guard
572 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
573 (core2i7_first_cycle_multipass_begin,)
574 (core2i7_first_cycle_multipass_issue,)
575 (core2i7_first_cycle_multipass_backtrack): Update signature.
577 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
578 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
579 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
581 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
582 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
584 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
586 * doc/tm.texi: Regenerate.
588 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
589 * haifa-sched.c (ready_try): Make signed to allow negative values.
590 (rebug_ready_list_1): Update.
591 (choose_ready): Simplify.
592 (sched_extend_ready_list): Update.
594 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
596 Remove IA64 speculation tweaking flags
597 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
598 speculation tuning flags.
599 (msched-prefer-non-data-spec-insns,)
600 (msched-prefer-non-control-spec-insns): Obsolete options.
601 * haifa-sched.c (choose_ready): Remove handling of
602 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
603 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
604 and PREFER_NON_DATA_SPEC.
605 * sel-sched.c (process_spec_exprs): Remove handling of
606 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
608 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
610 Improve scheduling debug output
611 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
612 (advance_one_cycle): Update.
613 (schedule_insn, queue_to_ready): Add debug printouts.
614 (debug_ready_list_1): New static function.
615 (debug_ready_list): Update.
616 (max_issue): Add debug printouts.
617 (dump_insn_stream): New static function.
618 (schedule_block): Use it. Also better indent printouts.
620 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
622 Fix sched_insn debug counter
623 * haifa-sched.c (schedule_insn): Update.
624 (struct haifa_saved_data): Add nonscheduled_insns_begin.
625 (save_backtrack_point, restore_backtrack_point): Update.
626 (first_nonscheduled_insn): New static function.
627 (queue_to_ready, choose_ready): Use it.
628 (schedule_block): Init nonscheduled_insns_begin.
629 (sched_emit_insn): Update.
632 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
634 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
636 (aarch64_secondary_reload) : LikeWise.
637 (aarch64_class_max_nregs) : Remove CORE_REGS.
638 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
639 (REG_CLASS_NAMES) : Likewise.
640 (REG_CLASS_CONTENTS) : LikeWise.
641 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
643 2014-05-21 Guozhi Wei <carrot@google.com>
646 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
648 (vqdmulhq_n_s16): Likewise.
650 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
652 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
654 2014-05-21 Marek Polacek <polacek@redhat.com>
657 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
659 2014-05-21 Martin Jambor <mjambor@suse.cz>
661 * doc/invoke.texi (Optimize Options): Document parameters
662 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
663 ipa-cp-array-index-hint-bonus.
665 2014-05-21 Mark Wielaard <mjw@redhat.com>
668 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
669 version >= 3 or not strict DWARF.
670 * langhooks.h (struct lang_hooks_for_types): Add
671 enum_underlying_base_type.
672 * langhooks.c (lhd_enum_underlying_base_type): New function.
673 * gcc/langhooks.h (struct lang_hooks_for_types): Add
674 enum_underlying_base_type.
675 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
676 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
677 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
679 2014-05-21 Richard Biener <rguenther@suse.de>
681 * doc/invoke.texi (-flto-partition=): Document one and
684 2014-05-21 John Marino <gnugcc@marino.st>
686 * config.gcc (*-*-dragonfly*): New target.
687 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
688 * configure: Regenerate.
689 * config/dragonfly-stdint.h: New.
690 * config/dragonfly.h: New.
691 * config/dragonfly.opt: New.
692 * config/i386/dragonfly.h: New.
693 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
695 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
697 * tree.def (VOID_CST): New.
698 * tree-core.h (TI_VOID): New.
699 * tree.h (void_node): New.
700 * tree.c (tree_node_structure_for_code, tree_code_size)
701 (iterative_hash_expr): Handle VOID_CST.
702 (build_common_tree_nodes): Initialize void_node.
704 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
706 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
708 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
710 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
713 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
714 flag_reorder_blocks_and_partition.
715 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
717 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
720 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
722 (*addc_r_t): Add new insn_and_split.
724 2014-05-21 Jakub Jelinek <jakub@redhat.com>
727 * omp-low.c (handle_simd_reference): New function.
728 (lower_rec_input_clauses): Use it. Defer adding reference
729 initialization even for reduction without placeholder if in simd,
730 handle it properly later on.
732 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
734 PR tree-optimization/60899
735 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
736 assume all static symbols will have definition wile parsing and
737 check the do have definition later in compilation; check that
738 variable referring symbol will be output before concluding that
739 reference is safe; be conservative for referring local statics;
740 be more precise about when comdat is output in other partition.
742 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
745 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
746 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
747 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
749 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
751 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
752 set writeonly flag only for vars actually written to.
754 2014-05-20 Dehao Chen <dehao@google.com>
756 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
757 and callee count to get clone count.
758 * tree-inline.c (expand_call_inline): Use callee count instead of bb
761 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
763 PR rtl-optimization/61243
764 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
766 2014-05-20 Xinliang David Li <davidxl@google.com>
768 * cgraphunit.c (walk_polymorphic_call_targets): Add
769 dbgcnt and fopt-info support.
770 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
771 * ipa-devirt.c (ipa_devirt): Ditto.
772 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
774 * ipa.c (walk_polymorphic_call_targets): Ditto.
775 * gimple-fold.c (fold_gimple_assign): Ditto.
776 (gimple_fold_call): Ditto.
777 * dbgcnt.def: New counter.
779 2014-05-20 DJ Delorie <dj@redhat.com>
781 * config/msp430/msp430.md (split): Don't allow subregs when
782 splitting SImode adds.
783 (andneghi): Fix subtraction logic.
784 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
786 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
788 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
790 * except.c (switch_to_exception_section, resolve_unique_section,
791 get_named_text_section, default_function_rodata_section,
792 align_variable, get_block_for_decl, default_section_type_flags):
793 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
794 * symtab.c (symtab_add_to_same_comdat_group,
795 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
796 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
798 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
799 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
800 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
801 (c6x_function_in_section_p): Likewise.
802 * config/darwin.c (machopic_select_section): Likewise.
803 * config/arm/arm.c (arm_function_in_section_p): Likewise.
804 * config/mips/mips.c (mips_function_rodata_section): Likewise.
805 * config/mep/mep.c (mep_select_section): LIkewise.
806 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
808 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
810 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
811 EH region of calls to pure functions that can throw an exception.
812 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
813 (copy_reference_ops_from_call): Also copy the EH region of the call if
814 it can throw an exception.
816 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
818 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
819 nested VEC_SELECTs that are inverses of each other.
821 2014-05-20 Richard Biener <rguenther@suse.de>
823 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
825 (extract_and_process_scc_for_name): not here.
826 (cond_dom_walker::before_dom_children): Only process
827 stmts that end the BB in interesting ways.
828 (run_scc_vn): Mark param uses as visited.
830 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
832 * config/arm/arm.md (arith_shiftsi): Do not predicate for
835 2014-05-20 Nick Clifton <nickc@redhat.com>
837 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
838 (msp430_gimplify_va_arg_expr): New function.
839 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
841 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
842 operand 0 in order to prevent confusion about the number of
845 2014-05-20 Richard Biener <rguenther@suse.de>
847 PR tree-optimization/61221
848 * tree-ssa-pre.c (el_to_update): Remove.
849 (eliminate_dom_walker::before_dom_children): Handle released
850 VDEFs by value-numbering them to the associated VUSE. Update
851 stmt immediately for substituted call address.
852 (eliminate): Remove delayed stmt updating code.
853 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
854 possibly late re-numbered vuses.
855 (vn_reference_lookup_2): Adjust.
856 (vn_reference_lookup_pieces): Likewise.
857 (vn_reference_lookup): Likewise.
859 2014-05-20 Richard Biener <rguenther@suse.de>
861 * config.gcc: Remove need_64bit_hwint.
862 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
863 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
865 * config.in: Regenerate.
866 * configure: Likewise.
868 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
870 * doc/extend.texi: Create Label Attributes section,
871 move all label attributes into it and reference it.
873 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
875 * arm.c (thumb1_reorg): When scanning backwards skip anything
876 that's not a proper insn.
878 2014-05-19 Richard Biener <rguenther@suse.de>
880 PR tree-optimization/61221
881 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
882 Do nothing for unreachable blocks.
883 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
884 Improve unreachability detection.
886 2014-05-19 Richard Biener <rguenther@suse.de>
888 PR tree-optimization/61209
889 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
891 2014-05-19 Nick Clifton <nickc@redhat.com>
893 * except.c (init_eh): Fix computation of builtin setjmp buffer
894 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
896 2014-05-19 Richard Biener <rguenther@suse.de>
898 PR tree-optimization/61184
899 * tree-vrp.c (is_negative_overflow_infinity): Use
900 TREE_OVERFLOW_P and do that check first.
901 (is_positive_overflow_infinity): Likewise.
902 (is_overflow_infinity): Likewise.
903 (vrp_operand_equal_p): Properly treat operands with
904 differing overflow as not equal.
906 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
908 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
909 shift simplification where it was intended.
911 2014-05-19 Christian Bruel <christian.bruel@st.com>
914 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
916 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
919 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
922 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
924 * reg-notes.def (CROSSING_JUMP): Likewise.
925 * rtl.h (rtx_def): Update comment for jump flag.
926 (CROSSING_JUMP_P): Define.
927 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
928 of a REG_CROSSING_JUMP note.
929 * cfghooks.c (tidy_fallthru_edges): Likewise.
930 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
931 * emit-rtl.c (try_split): Likewise.
932 * haifa-sched.c (sched_create_recovery_edges): Likewise.
933 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
934 * jump.c (redirect_jump_2): Likewise.
935 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
936 (relax_delay_slots): Likewise.
937 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
939 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
940 * config/sh/sh.md (jump_compact): Likewise.
941 * bb-reorder.c (rotate_loop): Likewise.
942 (pass_duplicate_computed_gotos::execute): Likewise.
943 (add_reg_crossing_jump_notes): Rename to...
944 (update_crossing_jump_flags): ...this.
945 (pass_partition_blocks::execute): Update accordingly.
947 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
949 * tree.h: Remove extraneous template <>.
951 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
953 * ipa.c (symtab_remove_unreachable_nodes): Remove
954 symbol from comdat group if its body was eliminated.
955 (comdat_can_be_unshared_p_1): Static symbols can always
957 * symtab.c (symtab_remove_from_same_comdat_group): Break out
959 (symtab_unregister_node): ... this one.
960 (verify_symtab_base): More strict checking of comdats.
961 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
963 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
965 * tree-pass.h (make_pass_ipa_comdats): New pass.
966 * timevar.def (TV_IPA_COMDATS): New timevar.
967 * passes.def (pass_ipa_comdats): Add.
968 * Makefile.in (OBJS): Add ipa-comdats.o
969 * ipa-comdats.c: New file.
971 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
973 * ipa.c (update_visibility_by_resolution_info): New function.
974 (function_and_variable_visibility): Use it.
976 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
978 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
980 (FOR_EACH_DEFINED_SYMBOL): New macro.
981 (varpool_first_static_initializer, varpool_next_static_initializer,
982 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
983 (symtab_in_same_comdat_p): Correctly deal with inline functions.
985 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
987 * ggc-page.c (ggc_handle_finalizers): Add comment.
989 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
991 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
992 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
993 (ggc_internal_cleared_alloc): Likewise.
994 * ggc-page.c (finalizer): New class.
995 (vec_finalizer): Likewise.
996 (globals::finalizers): New member.
997 (globals::vec_finalizers): Likewise.
998 (ggc_internal_alloc): Record the finalizer if any for the block being
1000 (ggc_handle_finalizers): New function.
1001 (ggc_collect): Call ggc_handle_finalizers.
1002 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1004 (ggc_internal_cleared_alloc): Likewise.
1005 (finalize): New function.
1006 (need_finalization_p): Likewise.
1007 (ggc_alloc): Install the type's destructor as the finalizer if it
1009 (ggc_cleared_alloc): Likewise.
1010 (ggc_vec_alloc): Likewise.
1011 (ggc_cleared_vec_alloc): Likewise.
1013 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1015 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1017 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1019 * alias.c (record_alias_subset): Adjust.
1020 * bitmap.c (bitmap_element_allocate): Likewise.
1021 (bitmap_gc_alloc_stat): Likewise.
1022 * cfg.c (init_flow): Likewise.
1023 (alloc_block): Likewise.
1024 (unchecked_make_edge): Likewise.
1025 * cfgloop.c (alloc_loop): Likewise.
1026 (flow_loops_find): Likewise.
1027 (rescan_loop_exit): Likewise.
1028 * cfgrtl.c (init_rtl_bb_info): Likewise.
1029 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1030 (cgraph_allocate_node): Likewise.
1031 (cgraph_create_edge_1): Likewise.
1032 (cgraph_allocate_init_indirect_info): Likewise.
1033 * cgraphclones.c (cgraph_clone_edge): Likewise.
1034 * cgraphunit.c (add_asm_node): Likewise.
1035 (init_lowered_empty_function): Likewise.
1036 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1037 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1038 (alpha_use_linkage): Likewise.
1039 * config/arc/arc.c (arc_init_machine_status): Likewise.
1040 * config/arm/arm.c (arm_init_machine_status): Likewise.
1041 * config/avr/avr.c (avr_init_machine_status): Likewise.
1042 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1043 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1044 * config/cris/cris.c (cris_init_machine_status): Likewise.
1045 * config/darwin.c (machopic_indirection_name): Likewise.
1046 (darwin_build_constant_cfstring): Likewise.
1047 (darwin_enter_string_into_cfstring_table): Likewise.
1048 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1049 * config/frv/frv.c (frv_init_machine_status): Likewise.
1050 * config/i386/i386.c (get_dllimport_decl): Likewise.
1051 (ix86_init_machine_status): Likewise.
1052 (assign_386_stack_local): Likewise.
1053 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1054 (i386_pe_maybe_record_exported_symbol): Likewise.
1055 (i386_pe_record_stub): Likewise.
1056 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1057 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1058 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1059 (m32c_note_pragma_address): Likewise.
1060 * config/mep/mep.c (mep_init_machine_status): Likewise.
1061 (mep_note_pragma_flag): Likewise.
1062 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1063 (mips16_local_alias): Likewise.
1064 (mips_init_machine_status): Likewise.
1065 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1066 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1067 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1068 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1069 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1070 * config/pa/pa.c (pa_init_machine_status): Likewise.
1071 (pa_get_deferred_plabel): Likewise.
1072 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1073 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1074 (rs6000_init_machine_status): Likewise.
1075 (output_toc): Likewise.
1076 * config/s390/s390.c (s390_init_machine_status): Likewise.
1077 * config/score/score.c (score_output_external): Likewise.
1078 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1079 * config/spu/spu.c (spu_init_machine_status): Likewise.
1080 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1081 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1082 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1083 * coverage.c (coverage_end_function): Likewise.
1084 * dbxout.c (dbxout_init): Likewise.
1085 * doc/gty.texi: Don't mention variable_size attribute.
1086 * dwarf2cfi.c (new_cfi): Adjust.
1087 (new_cfi_row): Likewise.
1088 (copy_cfi_row): Likewise.
1089 (create_cie_data): Likewise.
1090 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1091 (new_loc_descr): Likewise.
1092 (find_AT_string_in_table): Likewise.
1093 (add_addr_table_entry): Likewise.
1094 (new_die): Likewise.
1095 (add_var_loc_to_decl): Likewise.
1096 (clone_die): Likewise.
1097 (clone_as_declaration): Likewise.
1098 (break_out_comdat_types): Likewise.
1099 (new_loc_list): Likewise.
1100 (add_loc_descr_to_each): Likewise.
1101 (add_location_or_const_value_attribute): Likewise.
1102 (add_linkage_name): Likewise.
1103 (lookup_filename): Likewise.
1104 (dwarf2out_var_location): Likewise.
1105 (new_line_info_table): Likewise.
1106 (dwarf2out_init): Likewise.
1107 (mem_loc_descriptor): Likewise.
1108 (loc_descriptor): Likewise.
1109 (add_const_value_attribute): Likewise.
1110 (tree_add_const_value_attribute): Likewise.
1111 (comp_dir_string): Likewise.
1112 (dwarf2out_vms_debug_main_pointer): Likewise.
1113 (string_cst_pool_decl): Likewise.
1114 * emit-rtl.c (set_mem_attrs): Likewise.
1115 (get_reg_attrs): Likewise.
1116 (start_sequence): Likewise.
1117 (init_emit): Likewise.
1118 (init_emit_regs): Likewise.
1119 * except.c (init_eh_for_function): Likewise.
1120 (gen_eh_region): Likewise.
1121 (gen_eh_region_catch): Likewise.
1122 (gen_eh_landing_pad): Likewise.
1123 (add_call_site): Likewise.
1124 * function.c (add_frame_space): Likewise.
1125 (insert_temp_slot_address): Likewise.
1126 (assign_stack_temp_for_type): Likewise.
1127 (get_hard_reg_initial_val): Likewise.
1128 (allocate_struct_function): Likewise.
1129 (prepare_function_start): Likewise.
1130 (types_used_by_var_decl_insert): Likewise.
1131 * gengtype.c (variable_size_p): Remove function.
1132 (enum alloc_quantity): Remove enum.
1133 (write_typed_alloc_def): Remove function.
1134 (write_typed_struct_alloc_def): Likewise.
1135 (write_typed_typedef_alloc_def): Likewise.
1136 (write_typed_alloc_defns): Likewise.
1138 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1139 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1140 * ggc.h (ggc_alloc): new function.
1141 (ggc_cleared_alloc): Likewise.
1142 (ggc_vec_alloc): Template on type of vector element, and remove
1143 element size argument.
1144 (ggc_cleared_vec_alloc): Likewise.
1145 * gimple.c (gimple_build_omp_for): Adjust.
1146 (gimple_copy): Likewise.
1147 * ipa-cp.c (get_replacement_map): Likewise.
1148 (find_aggregate_values_for_callers_subset): Likewise.
1149 (known_aggs_to_agg_replacement_list): Likewise.
1150 * ipa-devirt.c (get_odr_type): Likewise.
1151 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1152 (read_agg_replacement_chain): Likewise.
1153 * loop-iv.c (get_simple_loop_desc): Likewise.
1154 * lto-cgraph.c (input_node_opt_summary): Likewise.
1155 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1156 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1157 (input_eh_region): Likewise.
1158 (input_eh_lp): Likewise.
1159 (input_cfg): Likewise.
1160 * optabs.c (set_optab_libfunc): Likewise.
1161 (init_tree_optimization_optabs): Likewise.
1162 (set_conv_libfunc): Likewise.
1163 * passes.c (do_per_function_toporder): Likewise.
1164 * rtl.h: Don't use variable_size gty attribute.
1165 * sese.c (if_region_set_false_region): Adjust.
1166 * stringpool.c (gt_pch_save_stringpool): Likewise.
1167 * target-globals.c (save_target_globals): Likewise.
1168 * toplev.c (general_init): Likewise.
1169 * trans-mem.c (record_tm_replacement): Likewise.
1170 (split_bb_make_tm_edge): Likewise.
1171 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1172 * tree-data-ref.h (lambda_vector_new): Likewise.
1173 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1174 * tree-iterator.c (tsi_link_before): Likewise.
1175 (tsi_link_after): Likewise.
1176 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1177 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1178 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1179 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1180 * tree-ssa.c (init_tree_ssa): Adjust.
1181 * tree-ssanames.c (set_range_info): Likewise.
1182 (get_ptr_info): Likewise.
1183 (duplicate_ssa_name_ptr_info): Likewise.
1184 (duplicate_ssa_name_range_info): Likewise.
1185 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1186 (unpack_ts_fixed_cst_value_fields): Likewise.
1187 * tree.c (build_fixed): Likewise.
1188 (build_real): Likewise.
1189 (build_string): Likewise.
1190 (decl_priority_info): Likewise.
1191 (decl_debug_expr_insert): Likewise.
1192 (decl_value_expr_insert): Likewise.
1193 (decl_debug_args_insert): Likewise.
1194 (type_hash_add): Likewise.
1195 (build_omp_clause): Likewise.
1196 * ubsan.c (decl_for_type_insert): Likewise.
1197 * varasm.c (get_unnamed_section): Likewise.
1198 (get_noswitch_section): Likewise.
1199 (get_section): Likewise.
1200 (get_block_for_section): Likewise.
1201 (create_block_symbol): Likewise.
1202 (build_constant_desc): Likewise.
1203 (create_constant_pool): Likewise.
1204 (force_const_mem): Likewise.
1205 (record_tm_clone_pair): Likewise.
1206 * varpool.c (varpool_create_empty_node): Likewise.
1208 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1210 * dwarf2out.c (tree_add_const_value_attribute): Call
1211 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1212 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1213 instead of ggc_internal_<x>alloc_stat.
1214 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1215 (ggc_realloc): Likewise.
1216 * ggc-none.c (ggc_internal_alloc): Likewise.
1217 (ggc_internal_cleared_alloc): Likewise.
1218 * ggc-page.c: Likewise.
1219 * ggc.h (ggc_internal_alloc_stat): Likewise.
1220 (ggc_internal_alloc): Remove macro.
1221 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1222 (ggc_internal_cleared_alloc): Remove macro.
1223 (GGC_RESIZEVEC): Adjust.
1224 (ggc_resizevar): Remove macro.
1225 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1226 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1227 (ggc_internal_vec_cleared_alloc): Remove macro.
1228 (ggc_alloc_atomic_stat): Drop _stat suffix.
1229 (ggc_alloc_atomic): Remove macro.
1230 (ggc_alloc_cleared_atomic): Remove macro.
1231 (ggc_alloc_string_stat): Drop _stat suffix.
1232 (ggc_alloc_string): Remove macro.
1233 (ggc_alloc_rtx_def_stat): Adjust.
1234 (ggc_alloc_tree_node_stat): Likewise.
1235 (ggc_alloc_cleared_tree_node_stat): Likewise.
1236 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1237 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1238 * gimple.c (gimple_build_omp_for): Likewise.
1239 (gimple_copy): Likewise.
1240 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1241 * toplev.c (realloc_for_line_map): Adjust.
1242 * tree-data-ref.h (lambda_vector_new): Likewise.
1243 * tree-phinodes.c (allocate_phi_node): Likewise.
1244 * tree.c (grow_tree_vec_stat): Likewise.
1245 * vec.h (va_gc::reserve): Adjust.
1247 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1249 * config/microblaze/microblaze.c (break_handler): New Declaration.
1250 (microblaze_break_function_p,microblaze_is_break_handler): New.
1251 (compute_frame_size): Use microblaze_break_function_p.
1252 Add the test of break_handler.
1253 (microblaze_function_prologue) : Add the test of variable
1254 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1255 (microblaze_function_epilogue) : Add the test of break_handler.
1256 (microblaze_globalize_label) : Add the test of break_handler.
1257 Check the name by BREAK_HANDLER_NAME.
1259 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1261 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1262 microblaze_is_break_handler test.
1263 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1264 Use SYMBOL_REF_DECL.
1266 * config/microblaze/microblaze-protos.h
1267 (microblaze_break_function_p,microblaze_is_break_handler):
1270 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1271 new MicroBlaze break_handler functions.
1273 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1275 * doc/extend.texi (Size of an asm): Move node text according
1276 to its @menu entry position.
1278 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1280 PR tree-optimization/61140
1281 PR tree-optimization/61150
1282 PR tree-optimization/61197
1283 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1285 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1287 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1289 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1291 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1292 __SIZEOF_INT128__ is defined.
1294 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1296 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1297 (rs6000_delegitimize_address): Use it.
1299 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1301 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1302 inplace argument. Store the new address in the original MEM when true.
1303 * emit-rtl.c (change_address_1): Likewise.
1304 (adjust_address_1, adjust_automodify_address_1, offset_address):
1306 * rtl.h (plus_constant): Add an inplace argument.
1307 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1308 when true. Avoid generating (plus X (const_int 0)).
1309 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1310 in-place. Pass true to plus_constant.
1311 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1313 2014-05-16 Dehao Chen <dehao@google.com>
1315 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1317 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1320 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1322 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1324 2014-05-16 Dehao Chen <dehao@google.com>
1326 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1327 optimize_function_for_size_p.
1328 * regs.h (REG_FREQ_FROM_BB): Likewise.
1330 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1333 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1334 negt_reg_operand cases.
1335 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1337 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1339 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1341 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1344 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1347 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1349 * tree-cfg.c (dump_function_to_file): Dump the return type of
1350 functions, in a line to itself before the function body, mimicking
1351 the layout of a C function.
1353 2014-05-16 Dehao Chen <dehao@google.com>
1355 * cfghooks.c (make_forwarder_block): Use direct computation to
1356 get fall-through edge's count and frequency.
1358 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1360 * config/arc/arc.c (arc_init): Fix typo in error message.
1361 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1362 (split_stack_prologue_scratch_regno): Likewise.
1363 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1364 word from error message.
1366 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1368 * ira-costs.c: Fix typo in comment.
1370 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1372 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1374 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1376 * varpool.c (dump_varpool_node): Dump write-only flag.
1377 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1379 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1380 write-only variables.
1381 * ipa.c (process_references): New function.
1382 (set_readonly_bit): New function.
1383 (set_writeonly_bit): New function.
1384 (clear_addressable_bit): New function.
1385 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1386 fix handling of aliases.
1387 * cgraph.h (struct varpool_node): Add writeonly flag.
1389 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1391 PR rtl-optimization/60969
1392 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1393 Calculate costs for this case.
1395 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1397 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1398 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1400 2014-05-16 Richard Biener <rguenther@suse.de>
1402 PR tree-optimization/61194
1403 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1404 bool patterns ending in a COND_EXPR.
1406 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1408 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1410 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1412 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1413 where we were unable to cost an RTX.
1415 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1417 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
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): Cost TRUNCATE.
1425 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1426 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1428 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1429 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1431 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1432 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1434 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1437 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1438 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1440 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1443 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1444 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1446 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1447 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1449 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1450 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1452 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1455 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1456 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1458 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1459 ZERO_EXTEND and SIGN_EXTEND better.
1461 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1462 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1464 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1467 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1468 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1470 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1471 costs when costing loads and stores to memory.
1473 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1474 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1476 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1479 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1481 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1483 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1484 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1486 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1488 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1490 (aarch64_rtx_mult_cost): New.
1491 (aarch64_rtx_costs): Use it, refactor as appropriate.
1493 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1494 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1496 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1497 emit instructions, return number of instructions which would
1499 (aarch64_add_constant): Update call to aarch64_build_constant.
1500 (aarch64_output_mi_thunk): Likewise.
1501 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1504 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1506 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1507 (TARGET_RTX_COSTS): Call it.
1509 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1511 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1512 (cortexa57_vector_cost): Likewise.
1513 (cortexa57_tunings): Use them.
1515 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1517 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1518 (cpu_addrcost_table): Use it.
1519 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1520 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1523 2014-05-16 Richard Biener <rguenther@suse.de>
1525 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1526 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1527 (visit_phi): Ignore edges marked as not executable.
1528 (class cond_dom_walker): New.
1529 (cond_dom_walker::before_dom_children): Value-number
1530 control statements and mark successor edges as not
1531 executable if possible.
1532 (run_scc_vn): First walk all control statements in
1533 dominator order, marking edges as not executable.
1534 * tree-inline.c (copy_edges_for_bb): Be not confused
1535 about random edge flags.
1537 2014-05-16 Richard Biener <rguenther@suse.de>
1539 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1541 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1544 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1545 (__TM_simple_begin): Use it.
1546 (__TM_begin): Likewise.
1548 2014-05-15 Martin Jambor <mjambor@suse.cz>
1551 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1552 type_preserved flag when the indirect edge is polymorphic.
1554 2014-05-15 Martin Jambor <mjambor@suse.cz>
1556 PR tree-optimization/61090
1557 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1558 build_ref_for_model.
1560 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1562 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1563 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1565 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1567 PR tree-optimization/61158
1568 * fold-const.c (fold_binary_loc): If X is zero-extended and
1569 shiftc >= prec, make sure zerobits is all ones instead of
1570 invoking undefined behavior.
1572 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1574 * regcprop.h: New file.
1575 * regcprop.c (skip_debug_insn_p): New decl.
1576 (replace_oldest_value_reg): Check skip_debug_insn_p.
1577 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1578 * shrink-wrap.c: Include regcprop.h.
1579 (prepare_shrink_wrap): Call
1580 copyprop_hardreg_forward_bb_without_debug_insn.
1582 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1584 * shrink-wrap.h: Update comment.
1585 * shrink-wrap.c: Update comment.
1586 (next_block_for_reg): Rename to live_edge_for_reg.
1587 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1588 (move_insn_for_shrink_wrap): Split live_edge.
1589 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1591 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1593 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1595 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1596 * config/sparc/sparc.md (fptype_ut699): New attribute.
1597 (in_branch_delay): Return false if -mfix-ut699 is specified and
1598 fptype_ut699 is set to single.
1599 (truncdfsf2): Add fptype_ut699 attribute.
1600 (fix_truncdfsi2): Likewise.
1601 (floatsisf2): Change fptype attribute.
1602 (fix_truncsfsi2): Likewise.
1603 (negtf2_notv9): Delete.
1604 (negtf2_v9): Likewise.
1605 (negtf2_hq): New instruction.
1606 (negtf2): New instruction and splitter.
1607 (negdf2_notv9): Rewrite.
1608 (abstf2_notv9): Delete.
1609 (abstf2_hq_v9): Likewise.
1610 (abstf2_v9): Likewise.
1611 (abstf2_hq): New instruction.
1612 (abstf2): New instruction and splitter.
1613 (absdf2_notv9): Rewrite.
1615 2014-05-14 Cary Coutant <ccoutant@google.com>
1618 * opts.c (common_handle_option): Don't special-case "-g".
1619 (set_debug_level): Default to at least level 2 with "-g".
1621 2014-05-14 DJ Delorie <dj@redhat.com>
1623 * config/msp430/msp430.c (msp430_builtin): Add
1624 MSP430_BUILTIN_DELAY_CYCLES.
1625 (msp430_init_builtins): Register void __delay_cycles(long long).
1626 (msp430_builtin_decl): Add it.
1627 (cg_magic_constant): New.
1628 (msp430_expand_delay_cycles): New.
1629 (msp430_expand_builtin): Call it.
1630 (msp430_print_operand_raw): Change integer printing from "int" to
1632 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1633 (delay_cycles_start): New.
1634 (delay_cycles_end): New.
1635 (delay_cycles_32): New.
1636 (delay_cycles_32x): New.
1637 (delay_cycles_16): New.
1638 (delay_cycles_16x): New.
1639 (delay_cycles_2): New.
1640 (delay_cycles_1): New.
1641 * doc/extend.texi: Document __delay_cycles().
1643 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1645 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1646 length attribute computation.
1648 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1651 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1653 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1656 * config/sparc/sparc.md: Fix types of low and high in DI constant
1657 splitter. Use gen_int_mode in some other splitters.
1659 2014-05-14 Martin Jambor <mjambor@suse.cz>
1662 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1664 2014-05-14 James Norris <jnorris@codesourcery.com>
1666 * omp-low.c (expand_parallel_call): Remove shadow variable.
1667 (expand_omp_taskreg): Likewise.
1669 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1671 * common/config/i386/i386-common.c
1672 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1673 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1674 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1675 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1676 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1677 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1678 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1679 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1680 xsavecintrin.h, xsavesintrin.h.
1681 (x86_64-*-*): Ditto.
1682 * config/i386/clflushoptintrin.h: New.
1683 * config/i386/xsavecintrin.h: Ditto.
1684 * config/i386/xsavesintrin.h: Ditto.
1685 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1686 (bit_XSAVES): Ditto.
1687 (bit_XSAVES): Ditto.
1688 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1689 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1691 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1692 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1693 OPTION_MASK_ISA_XSAVES.
1694 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1696 (PTA_CLFLUSHOPT) Define.
1697 (PTA_XSAVEC): Ditto.
1698 (PTA_XSAVES): Ditto.
1699 (ix86_option_override_internal): Handle new options.
1700 (ix86_valid_target_attribute_inner_p): Ditto.
1701 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1702 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1703 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1704 (bdesc_special_args): Add __builtin_ia32_xsaves,
1705 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1706 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1707 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1708 (ix86_expand_builtin): Handle new builtins.
1709 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1710 (TARGET_CLFLUSHOPT_P): Ditto.
1711 (TARGET_XSAVEC): Ditto.
1712 (TARGET_XSAVEC_P): Ditto.
1713 (TARGET_XSAVES): Ditto.
1714 (TARGET_XSAVES_P): Ditto.
1715 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1716 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1717 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1719 (ANY_XRSTOR64): Ditto.
1721 (xrstor): Change into <xrstor>.
1722 (xrstor_rex64): Change into <xrstor>_rex64.
1723 (xrstor64): Change into <xrstor>64
1725 * config/i386/i386.opt (mclflushopt): New.
1728 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1730 * doc/invoke.texi: Document new options.
1732 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1734 PR rtl-optimization/60866
1735 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1736 Default it to -1. Pass it down to init_simplejump_data.
1737 (init_simplejump_data): New parameter old_seqno. Pass it down
1738 to get_seqno_for_a_jump.
1739 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1740 initializing new jump seqno as a last resort. Add comment.
1741 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1742 jump and pass it down to sel_init_new_insn.
1743 (sel_redirect_edge_and_branch_force): Likewise.
1745 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1747 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1748 shifted values to avoid build warning.
1750 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1752 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1753 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1754 (cfg_layout_merge_blocks): Likewise.
1755 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1757 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1759 PR rtl-optimization/60901
1760 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1761 bb predecessor belongs to the same scheduling region. Adjust comment.
1763 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1765 * doc/sourcebuild.texi: (dfp_hw): Document.
1766 (p8vector_hw): Likewise.
1767 (powerpc_eabi_ok): Likewise.
1768 (powerpc_elfv2): Likewise.
1769 (powerpc_htm_ok): Likewise.
1770 (ppc_recip_hw): Likewise.
1773 2014-05-13 Cary Coutant <ccoutant@google.com>
1775 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1777 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1779 * gengtype-parse.c (require3): Eliminate in favor of...
1781 (require_template_declaration): Update to support optional single *
1784 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1785 (create_user_defined_type): Handle a single level of explicit
1786 pointerness within template arguments.
1787 (struct write_types_data): Add field "kind".
1788 (filter_type_name): Handle "*" character.
1789 (write_user_func_for_structure_ptr): Require a write_types_data
1790 rather than just a prefix string, so that we can look up the kind
1791 of the wtd and use it as an index into wrote_user_func_for_ptr,
1792 ensuring that such functions are written at most once. Support
1793 subclasses by invoking the marking function of the ultimate base class.
1794 (write_user_func_for_structure_body): Require a write_types_data
1795 rather than just a prefix string, so that we can pass this to
1796 write_user_func_for_structure_ptr.
1797 (write_func_for_structure): Likewise.
1798 (ggc_wtd): Add initializer of new "kind" field.
1799 (pch_wtd): Likewise.
1801 * gengtype.h (enum write_types_kinds): New.
1802 (struct type): Add field wrote_user_func_for_ptr to the "s"
1805 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1807 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1808 instead of const_binop.
1809 (fold_binary_loc): Likewise.
1811 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1813 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1814 calculation to match get_ref_base_and_extent.
1816 2014-05-13 Catherine Moore <clm@codesourcery.com>
1817 Sandra Loosemore <sandra@codesourcery.com>
1819 * configure.ac: Fix assembly for explicit JALR relocation check.
1820 * configure: Regenerate.
1822 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1824 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1825 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1826 Remove associated type declarations and initialisations.
1827 (arm_expand_neon_builtin): Likewise.
1828 (neon_emit_pair_result_insn): Delete.
1829 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1830 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1831 (neon_vzip<mode>): Likewise.
1832 (neon_vuzp<mode>): Likewise.
1834 2014-05-13 Richard Biener <rguenther@suse.de>
1837 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1838 it needs revisiting whether the call still may be tail-called.
1840 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1842 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1843 * rtl.h (block_symbol): Reduce number of fields to 2.
1844 (rtx_def): Add u2.symbol_ref_flags.
1845 (SYMBOL_REF_FLAGS): Use it.
1846 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1847 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1848 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1849 Lower index of SYMBOL_REF_DATA.
1850 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1851 Print SYMBOL_REF_FLAGS at the same time.
1852 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1854 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1856 * rtl.def (VAR_LOCATION): Remove "i" field.
1857 * rtl.h (rtx_def): Add u2.var_location_status.
1858 (PAT_VAR_LOCATION_STATUS): Use it.
1859 (gen_rtx_VAR_LOCATION): Declare.
1860 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1861 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1862 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1864 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1866 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1867 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1869 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1871 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1872 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1873 * rtl.h (rtx_def): Add insn_uid to u2 field.
1874 (RTX_FLAG_CHECK8): Delete in favor of...
1875 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1876 (INSN_DELETED_P): Update accordingly.
1877 (INSN_UID): Use u2.insn_uid.
1878 (INSN_CHAIN_CODE_P): Define.
1879 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1880 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1881 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1882 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1883 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1884 indices accordingly.
1885 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1886 Update indices for insn-chain rtxes.
1887 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1888 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1889 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1890 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1891 * combine.c (try_combine): Likewise.
1892 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1894 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1896 * rtl.def (REG): Remove middle field.
1897 * rtl.h (rtx_def): Add orignal_regno to u2.
1898 (ORIGINAL_REGNO): Use it instead of field 1.
1899 (REG_ATTRS): Lower field index accordingly.
1900 * gengtype.c (adjust_field_rtx_def): Remove handling of
1901 ORIGINAL_REGNO. Move REG_ATTRS index down.
1902 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1903 code that prints the REGNO.
1905 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1907 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1910 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1912 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1914 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1916 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1917 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1919 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1921 * config/aarch64/aarch64-protos.h
1922 (aarch64_hard_regno_caller_save_mode): New prototype.
1923 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1925 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1927 2014-05-13 Christian Bruel <christian.bruel@st.com>
1929 * target.def (mode_switching): New hook vector.
1930 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1931 (mode_exit, modepriority_to_mode): Likewise.
1932 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1933 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1934 * target.h: Include tm.h and hard-reg-set.h.
1935 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1936 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1937 * doc/tm.texi Regenerate.
1938 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1939 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1940 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1941 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1942 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1943 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1944 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1945 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1946 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1947 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1948 (ix86_emit_mode_set): Hookify.
1949 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1951 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1952 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1953 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1954 (epiphany_mode_priority_to_mode): Remove declaration.
1955 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1956 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1957 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1959 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1960 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1961 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1963 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1966 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1967 is const0_rtx, return immediately. Don't test count == 0 when
1970 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1972 * Makefile.in: add shrink-wrap.o.
1973 * config/i386/i386.c: include "shrink-wrap.h"
1974 * function.c: Likewise.
1975 (requires_stack_frame_p, next_block_for_reg,
1976 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1977 dup_block_and_redirect): Move to shrink-wrap.c
1978 (thread_prologue_and_epilogue_insns): Extract three code segments
1979 as functions in shrink-wrap.c
1980 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1982 * shrink-wrap.c: New file.
1983 * shrink-wrap.h: New file.
1985 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1987 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1988 reference to Solaris.
1990 2014-05-12 Mike Stump <mikestump@comcast.net>
1993 * genattrtab.c (filename): Add.
1994 (convert_set_attr_alternative): Improve error message.
1995 (check_defs): Restore read_md_filename for error messages.
1996 (gen_insn): Save filename.
1998 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2000 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2001 -fno-local-ivars and -fivar-visibility.
2002 * c-family/c.opt: Make -Wshadow also implicitly enable
2005 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2007 * doc/tm.texi: Remove reference to deleted macro.
2008 * doc/tm.texi.in: Likewise.
2010 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2013 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2016 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2018 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2019 * config/arm/aout.h (License): Same.
2020 * config/arm/bpabi.h (License): Same.
2021 * config/arm/elf.h (License): Same.
2022 * config/arm/linux-elf.h (License): Same.
2023 * config/arm/linux-gas.h (License): Same.
2024 * config/arm/netbsd-elf.h (License): Same.
2025 * config/arm/uclinux-eabi.h (License): Same.
2026 * config/arm/uclinux-elf.h (License): Same.
2027 * config/arm/vxworks.h (License): Same.
2029 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2031 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2032 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2033 number of operands to 3.
2034 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2035 * tree-nested.c (convert_nonlocal_omp_clauses,
2036 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2037 * gimplify.c (gimplify_scan_omp_clauses): Handle
2038 OMP_CLAUSE_LINEAR_STMT.
2039 * omp-low.c (lower_rec_input_clauses): Fix typo.
2040 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2041 cast between Fortran boolean_type_node and C _Bool if
2044 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2046 PR tree-optimization/61136
2047 * wide-int.h (multiple_of_p): Define a version that doesn't return
2049 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2050 integer_zerop/const_binop pair.
2051 (multiple_of_p): Likewise, converting both operands to widest_int
2054 2014-05-09 Teresa Johnson <tejohnson@google.com>
2056 * cgraphunit.c (analyze_functions): Use correct dump file.
2058 2014-05-09 Florian Weimer <fweimer@redhat.com>
2060 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2062 (stack_protect_return_slot_p): New function.
2063 (expand_used_vars): Call stack_protect_decl_p and
2064 stack_protect_return_slot_p for -fstack-protector-strong.
2066 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2069 * fold-const.c (fold_binary_loc): Changed width of mask.
2071 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2073 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2074 unsigned int initializers for regno_in, regno_out.
2076 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2079 * config/avr/avr.md (cc): Add new attribute set_vzn.
2080 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2081 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2082 with INC, DEC or NEG.
2083 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2084 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2085 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2087 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2090 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2092 * wide-int.cc (UTItype): Define.
2093 (UDWtype): Define for appropriate W_TYPE_SIZE.
2095 2014-05-09 Richard Biener <rguenther@suse.de>
2097 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2098 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2099 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2100 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2101 ssa_propagate): Adjust.
2103 2014-05-08 Jeff Law <law@redhat.com>
2105 PR tree-optimization/61009
2106 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2107 tri-state rather than a boolean. When a block is too big to
2108 thread through, inform caller via negative return value.
2109 (thread_across_edge): If a block was too big for normal threading,
2110 then it's too big for a joiner too, so remove temporary equivalences
2111 and return immediately.
2113 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2114 Matthias Klose <doko@ubuntu.com>
2117 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2119 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2122 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2124 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2127 * config/alpha/alpha.c: Include gimple-iterator.h.
2128 (alpha_gimple_fold_builtin): New function. Move
2129 ALPHA_BUILTIN_UMULH folding from ...
2130 (alpha_fold_builtin): ... here.
2131 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2133 2014-05-08 Wei Mi <wmi@google.com>
2136 * config/i386/i386.c (ix86_compute_frame_layout): Update
2137 preferred_stack_boundary for call, expanded from tls descriptor.
2138 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2139 to depend on SP register.
2140 (*tls_local_dynamic_base_32_gnu): Ditto.
2141 (*tls_local_dynamic_32_once): Ditto.
2142 (tls_global_dynamic_64_<mode>): Set
2143 ix86_tls_descriptor_calls_expanded_in_cfun.
2144 (tls_local_dynamic_base_64_<mode>): Ditto.
2145 (tls_global_dynamic_32): Set
2146 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2147 to depend on SP register.
2148 (tls_local_dynamic_base_32): Ditto.
2150 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2152 * config/arm/arm_neon.h: Update comment.
2153 * config/arm/neon-docgen.ml: Delete.
2154 * config/arm/neon-gen.ml: Delete.
2155 * doc/arm-neon-intrinsics.texi: Update comment.
2157 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2159 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2161 (vand, vorr, veor, vorn, vbic): Remove.
2162 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2164 (neon_vsub_unspec): Likewise.
2165 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2167 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2169 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2170 (vadd_s16): Likewise.
2171 (vadd_s32): Likewise.
2172 (vadd_f32): Likewise.
2173 (vadd_u8): Likewise.
2174 (vadd_u16): Likewise.
2175 (vadd_u32): Likewise.
2176 (vadd_s64): Likewise.
2177 (vadd_u64): Likewise.
2178 (vaddq_s8): Likewise.
2179 (vaddq_s16): Likewise.
2180 (vaddq_s32): Likewise.
2181 (vaddq_s64): Likewise.
2182 (vaddq_f32): Likewise.
2183 (vaddq_u8): Likewise.
2184 (vaddq_u16): Likewise.
2185 (vaddq_u32): Likewise.
2186 (vaddq_u64): Likewise.
2187 (vmul_s8): Likewise.
2188 (vmul_s16): Likewise.
2189 (vmul_s32): Likewise.
2190 (vmul_f32): Likewise.
2191 (vmul_u8): Likewise.
2192 (vmul_u16): Likewise.
2193 (vmul_u32): Likewise.
2194 (vmul_p8): Likewise.
2195 (vmulq_s8): Likewise.
2196 (vmulq_s16): Likewise.
2197 (vmulq_s32): Likewise.
2198 (vmulq_f32): Likewise.
2199 (vmulq_u8): Likewise.
2200 (vmulq_u16): Likewise.
2201 (vmulq_u32): Likewise.
2202 (vsub_s8): Likewise.
2203 (vsub_s16): Likewise.
2204 (vsub_s32): Likewise.
2205 (vsub_f32): Likewise.
2206 (vsub_u8): Likewise.
2207 (vsub_u16): Likewise.
2208 (vsub_u32): Likewise.
2209 (vsub_s64): Likewise.
2210 (vsub_u64): Likewise.
2211 (vsubq_s8): Likewise.
2212 (vsubq_s16): Likewise.
2213 (vsubq_s32): Likewise.
2214 (vsubq_s64): Likewise.
2215 (vsubq_f32): Likewise.
2216 (vsubq_u8): Likewise.
2217 (vsubq_u16): Likewise.
2218 (vsubq_u32): Likewise.
2219 (vsubq_u64): Likewise.
2220 (vand_s8): Likewise.
2221 (vand_s16): Likewise.
2222 (vand_s32): Likewise.
2223 (vand_u8): Likewise.
2224 (vand_u16): Likewise.
2225 (vand_u32): Likewise.
2226 (vand_s64): Likewise.
2227 (vand_u64): Likewise.
2228 (vandq_s8): Likewise.
2229 (vandq_s16): Likewise.
2230 (vandq_s32): Likewise.
2231 (vandq_s64): Likewise.
2232 (vandq_u8): Likewise.
2233 (vandq_u16): Likewise.
2234 (vandq_u32): Likewise.
2235 (vandq_u64): Likewise.
2236 (vorr_s8): Likewise.
2237 (vorr_s16): Likewise.
2238 (vorr_s32): Likewise.
2239 (vorr_u8): Likewise.
2240 (vorr_u16): Likewise.
2241 (vorr_u32): Likewise.
2242 (vorr_s64): Likewise.
2243 (vorr_u64): Likewise.
2244 (vorrq_s8): Likewise.
2245 (vorrq_s16): Likewise.
2246 (vorrq_s32): Likewise.
2247 (vorrq_s64): Likewise.
2248 (vorrq_u8): Likewise.
2249 (vorrq_u16): Likewise.
2250 (vorrq_u32): Likewise.
2251 (vorrq_u64): Likewise.
2252 (veor_s8): Likewise.
2253 (veor_s16): Likewise.
2254 (veor_s32): Likewise.
2255 (veor_u8): Likewise.
2256 (veor_u16): Likewise.
2257 (veor_u32): Likewise.
2258 (veor_s64): Likewise.
2259 (veor_u64): Likewise.
2260 (veorq_s8): Likewise.
2261 (veorq_s16): Likewise.
2262 (veorq_s32): Likewise.
2263 (veorq_s64): Likewise.
2264 (veorq_u8): Likewise.
2265 (veorq_u16): Likewise.
2266 (veorq_u32): Likewise.
2267 (veorq_u64): Likewise.
2268 (vbic_s8): Likewise.
2269 (vbic_s16): Likewise.
2270 (vbic_s32): Likewise.
2271 (vbic_u8): Likewise.
2272 (vbic_u16): Likewise.
2273 (vbic_u32): Likewise.
2274 (vbic_s64): Likewise.
2275 (vbic_u64): Likewise.
2276 (vbicq_s8): Likewise.
2277 (vbicq_s16): Likewise.
2278 (vbicq_s32): Likewise.
2279 (vbicq_s64): Likewise.
2280 (vbicq_u8): Likewise.
2281 (vbicq_u16): Likewise.
2282 (vbicq_u32): Likewise.
2283 (vbicq_u64): Likewise.
2284 (vorn_s8): Likewise.
2285 (vorn_s16): Likewise.
2286 (vorn_s32): Likewise.
2287 (vorn_u8): Likewise.
2288 (vorn_u16): Likewise.
2289 (vorn_u32): Likewise.
2290 (vorn_s64): Likewise.
2291 (vorn_u64): Likewise.
2292 (vornq_s8): Likewise.
2293 (vornq_s16): Likewise.
2294 (vornq_s32): Likewise.
2295 (vornq_s64): Likewise.
2296 (vornq_u8): Likewise.
2297 (vornq_u16): Likewise.
2298 (vornq_u32): Likewise.
2299 (vornq_u64): Likewise.
2301 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2303 * wide-int.cc (UTItype): Define.
2304 (UDWtype): Define for appropriate W_TYPE_SIZE.
2306 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2308 PR tree-optimization/59100
2309 * tree-ssa-phiopt.c: Include tree-inline.h.
2310 (neutral_element_p, absorbing_element_p): New functions.
2311 (value_replacement): Handle conditional binary operations with a
2312 neutral or absorbing element.
2314 2014-05-08 Richard Biener <rguenther@suse.de>
2316 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2317 folding the expression.
2318 (valueize_expr): Remove.
2319 (visit_reference_op_load): Do not valueize the result of
2321 (simplify_binary_expression): Likewise.
2322 (simplify_unary_expression): Likewise.
2324 2014-05-08 Richard Biener <rguenther@suse.de>
2326 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2327 looking at TYPE_ARG_TYPES.
2329 2014-05-08 Richard Biener <rguenther@suse.de>
2331 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2332 pointer propagation special-case.
2334 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2336 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2337 core part of address expressions.
2339 2014-05-08 Alan Modra <amodra@gmail.com>
2342 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2343 loads and stores when -mno-strict-align at any alignment.
2344 (expand_block_clear): Similarly. Also correct calculation of
2347 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2350 * tree-complex.c (expand_complex_move): Keep line info when expanding
2352 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2353 of complex expression. Use new argument to display correct location
2354 for values coming from phi statement.
2355 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2356 (warn_uninitialized_phi): Pass location of phi argument to
2358 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2359 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2361 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2363 * config/rs6000/predicates.md (indexed_address_mem): New.
2364 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2365 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2366 fpstore_ux, fpstore_u.
2367 (sign_extend, indexed, update): New.
2368 (cell_micro): Adjust.
2369 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2370 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2371 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2372 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2373 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2374 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2375 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2376 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2377 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2378 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2379 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2380 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2381 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2382 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2383 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2385 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2386 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2387 *vsx_extract_<mode>_store): Adjust.
2388 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2389 is_cracked_insn, insn_must_be_first_in_group,
2390 insn_must_be_last_in_group): Adjust.
2392 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2394 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2395 ppc440-fpstore): Adjust.
2396 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2397 ppc476-fpstore): Adjust.
2398 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2399 ppc601-fpstore): Adjust.
2400 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2402 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2404 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2405 ppc7450-fpstore): Adjust.
2406 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2407 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2408 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2410 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2411 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2412 cell-fpstore, cell-fpstore-update): Adjust.
2413 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2414 ppce300c3_store, ppce300c3_fpstore): Adjust.
2415 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2416 e500mc_fpstore): Adjust.
2417 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2418 e500mc64_store, e500mc64_fpstore): Adjust.
2419 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2420 e5500_fpstore): Adjust.
2421 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2422 e6500_fpstore): Adjust.
2423 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2425 * config/rs6000/power4.md (power4-load, power4-load-ext,
2426 power4-load-ext-update, power4-load-ext-update-indexed,
2427 power4-load-update-indexed, power4-load-update, power4-fpload,
2428 power4-fpload-update, power4-store, power4-store-update,
2429 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2431 * config/rs6000/power5.md (power5-load, power5-load-ext,
2432 power5-load-ext-update, power5-load-ext-update-indexed,
2433 power5-load-update-indexed, power5-load-update, power5-fpload,
2434 power5-fpload-update, power5-store, power5-store-update,
2435 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2437 * config/rs6000/power6.md (power6-load, power6-load-ext,
2438 power6-load-update, power6-load-update-indexed,
2439 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2440 power6-fpload-update, power6-store, power6-store-update,
2441 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2443 * config/rs6000/power7.md (power7-load, power7-load-ext,
2444 power7-load-update, power7-load-update-indexed,
2445 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2446 power7-fpload-update, power7-store, power7-store-update,
2447 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2449 * config/rs6000/power8.md (power8-load, power8-load-update,
2450 power8-load-ext, power8-load-ext-update, power8-fpload,
2451 power8-fpload-update, power8-store, power8-store-update-indexed,
2452 power8-fpstore, power8-fpstore-update): Adjust.
2453 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2455 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2456 titan_lsu_store, titan_lsu_fpstore): Adjust.
2457 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2459 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2462 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2463 unrolled byte insns. Emit address increments after move insns.
2465 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2467 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2468 const_gimple, rather than a gimple.
2469 (gimple_call_builtin_p): Likewise, for the three variants.
2471 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2472 (gimple_call_builtin_p): Likewise, for the three variants.
2474 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2476 PR tree-optimization/61095
2477 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2479 2014-05-07 Richard Biener <rguenther@suse.de>
2481 PR tree-optimization/61034
2482 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2483 (maybe_skip_until): Use translate to take into account
2484 lattices when trying to do disambiguations.
2485 (get_continuation_for_phi_1): Likewise.
2486 (get_continuation_for_phi): Adjust for added translate arguments.
2487 (walk_non_aliased_vuses): Likewise.
2488 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2489 (walk_non_aliased_vuses): Likewise.
2490 (call_may_clobber_ref_p_1): Declare.
2491 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2492 calls. Stop early if we are only supposed to disambiguate.
2493 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2495 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2497 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2498 Emit an error when the function has arguments.
2500 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2502 * cfgloop.h (unswitch_loops): Remove.
2503 * doc/passes.texi: Remove references to loop-unswitch.c
2504 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2506 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2508 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2509 check for loads group of length 3.
2510 (vect_permute_load_chain): New permutations for loads group of
2512 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2513 of vec_perm_shuffle for the new permutations.
2515 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2517 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2518 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2519 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2520 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2521 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2522 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2523 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2524 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2526 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2528 * loop-unswitch.c: Delete.
2530 2014-05-07 Richard Biener <rguenther@suse.de>
2532 * config.gcc: Always set need_64bit_hwint to yes.
2534 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2536 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2537 of using optimize_size.
2539 2014-05-06 Mike Stump <mikestump@comcast.net>
2541 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2543 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2545 * config/i386/sse.md (*mov<mode>_internal)
2546 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2547 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2548 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2549 (*<code><mode>3, *andnot<mode>3<mask_name>)
2550 (<mask_codefor><code><mode>3<mask_name>): Only consider
2551 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2553 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2556 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2558 * lra-constraints.c (valid_address_p): Move earlier in file.
2559 Add a constraint argument to the address_info version.
2560 (satisfies_memory_constraint_p): New function.
2561 (satisfies_address_constraint_p): Likewise.
2562 (process_alt_operands, curr_insn_transform): Use them.
2563 (process_address): Pass the constraint to valid_address_p when
2564 checking address operands.
2566 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2568 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2569 to their respective blocks. Fix inadvertent use of "node".
2571 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2573 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2575 (init_emit_once): ...here.
2576 * rtl.h (init_derived_machine_modes): Declare.
2577 * toplev.c (do_compile): Call it even if no_backend.
2579 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2580 Mike Stump <mikestump@comcast.net>
2581 Richard Sandiford <rdsandiford@googlemail.com>
2582 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2584 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2585 (rtx_equal_for_memref_p): Update comment.
2586 (adjust_offset_for_component_ref): Use wide-int interfaces.
2587 * builtins.c (get_object_alignment_2): Likewise.
2588 (c_readstr): Likewise.
2589 (target_char_cast): Add comment.
2590 (determine_block_size): Use wide-int interfaces.
2591 (expand_builtin_signbit): Likewise.
2592 (fold_builtin_int_roundingfn): Likewise.
2593 (fold_builtin_bitop): Likewise.
2594 (fold_builtin_bswap): Likewise.
2595 (fold_builtin_logarithm): Use signop.
2596 (fold_builtin_pow): Likewise.
2597 (fold_builtin_memory_op): Use wide-int interfaces.
2598 (fold_builtin_object_size): Likewise.
2599 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2600 nb_iterations_estimate.
2601 (record_niter_bound): Use wide-int interfaces.
2602 (get_estimated_loop_iterations_int): Likewise.
2603 (get_estimated_loop_iterations): Likewise.
2604 (get_max_loop_iterations): Likewise.
2605 * cfgloop.h: Include wide-int.h.
2606 (struct nb_iter_bound): Change bound to widest_int.
2607 (struct loop): Change nb_iterations_upper_bound and
2608 nb_iterations_estimate to widest_int.
2609 (record_niter_bound): Switch to use widest_int.
2610 (get_estimated_loop_iterations): Likewise.
2611 (get_max_loop_iterations): Likewise.
2612 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2613 update for wide-int.
2614 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2615 * combine.c (try_combine): Likewise.
2616 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2617 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2619 (aarch64_float_const_representable_p): Likewise.
2620 * config/arc/arc.c: Include wide-int.h.
2621 (arc_can_use_doloop_p): Use wide-int interfaces.
2622 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2623 (vfp3_const_double_index): Likewise.
2624 * config/avr/avr.c (avr_out_round): Likewise.
2625 (avr_fold_builtin): Likewise.
2626 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2627 (bfin_can_use_doloop_p): Likewise.
2628 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2629 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2630 * config/i386/i386.c: Include wide-int.h.
2631 (ix86_data_alignment): Use wide-int interfaces.
2632 (ix86_local_alignment): Likewise.
2633 (ix86_emit_swsqrtsf): Update real_from_integer.
2634 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2635 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2636 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2637 (zero_constant): Likewise.
2638 (input_operand): Likewise.
2639 (splat_input_operand): Likewise.
2640 (non_logical_cint_operand): Change const_double to const_wide_int.
2641 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2642 (easy_altivec_constant): Remove comment.
2643 (paired_expand_vector_init): Use CONSTANT_P.
2644 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2645 (rs6000_emit_move): Update checks.
2646 (rs6000_aggregate_candidate): Use wide-int interfaces.
2647 (rs6000_expand_ternop_builtin): Likewise.
2648 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2649 (rs6000_assemble_integer): Likewise.
2650 (rs6000_hash_constant): Likewise.
2651 (output_toc): Likewise.
2652 (rs6000_rtx_costs): Likewise.
2653 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2654 * config/rs6000/rs6000-c.c: Include wide-int.h.
2655 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2656 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2657 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2658 Handle CONST_WIDE_INT.
2659 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2660 Use tree_fits_uhwi_p.
2661 * config/sparc/sparc.c: Include wide-int.h.
2662 (sparc_fold_builtin): Use wide-int interfaces.
2663 * config/vax/vax.c: Include wide-int.h.
2664 (vax_float_literal): Use real_from_integer.
2665 * coretypes.h (struct hwivec_def): New.
2667 (const_hwivec): New.
2668 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2669 (equiv_constant): Handle CONST_WIDE_INT.
2670 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2671 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2672 * dbxout.c (stabstr_U): Use wide-int interfaces.
2673 (dbxout_type): Update to use cst_fits_shwi_p.
2674 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2675 (TARGET_SUPPORTS_WIDE_INT): Add default.
2676 * dfp.c: Include wide-int.h.
2677 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2678 decimal_real_to_integer.
2679 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2680 decimal_real_to_integer.
2681 * doc/generic.texi (Constant expressions): Update for wide_int.
2682 * doc/rtl.texi (const_double): Likewise.
2683 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2684 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2685 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2686 (REAL_VALUE_FROM_INT): Remove.
2687 (TARGET_SUPPORTS_WIDE_INT): New.
2688 * doc/tm.texi: Regenerate.
2689 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2690 * double-int.h: Include wide-int.h.
2691 (struct wi::int_traits): New.
2692 * dwarf2out.c (get_full_len): New.
2693 (dw_val_equal_p): Add case dw_val_class_wide_int.
2694 (size_of_loc_descr): Likewise.
2695 (output_loc_operands): Likewise.
2696 (insert_double): Remove.
2697 (insert_wide_int): New.
2699 (print_die): Add case dw_val_class_wide_int.
2700 (attr_checksum): Likewise.
2701 (attr_checksum_ordered): Likewise.
2702 (same_dw_val_p): Likewise.
2703 (size_of_die): Likewise.
2704 (value_format): Likewise.
2705 (output_die): Likewise.
2706 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2708 (clz_loc_descriptor): Use wide-int interfaces.
2709 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2710 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2711 (round_up_to_align): Use wide-int interfaces.
2712 (field_byte_offset): Likewise.
2713 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2714 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2715 CONST_DOUBLE handling. Use wide-int interfaces.
2716 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2717 (gen_enumeration_type_die): Use add_AT_wide.
2718 (hash_loc_operands): Add case dw_val_class_wide_int.
2719 (compare_loc_operands): Likewise.
2720 * dwarf2out.h: Include wide-int.h.
2721 (wide_int_ptr): New.
2722 (enum dw_val_class): Add dw_val_class_wide_int.
2723 (struct dw_val_struct): Add val_wide.
2724 * emit-rtl.c (const_wide_int_htab): New.
2725 (const_wide_int_htab_hash): New.
2726 (const_wide_int_htab_eq): New.
2727 (lookup_const_wide_int): New.
2728 (const_double_htab_hash): Use wide-int interfaces.
2729 (const_double_htab_eq): Likewise.
2730 (rtx_to_double_int): Conditionally compile for wide-int.
2731 (immed_double_int_const): Rename to immed_wide_int_const and
2732 update for wide-int.
2733 (immed_double_const): Conditionally compile for wide-int.
2734 (init_emit_once): Use wide-int interfaces.
2735 * explow.c (plus_constant): Likewise.
2736 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2737 (lshift_value): Use wide-int interfaces.
2738 (expand_mult): Likewise.
2739 (choose_multiplier): Likewise.
2740 (expand_smod_pow2): Likewise.
2741 (make_tree): Likewise.
2742 * expr.c (convert_modes): Consolidate handling of constants.
2743 Use wide-int interfaces.
2744 (emit_group_load_1): Add note.
2745 (store_expr): Update comment.
2746 (get_inner_reference): Use wide-int interfaces.
2747 (expand_constructor): Update comment.
2748 (expand_expr_real_2): Use wide-int interfaces.
2749 (expand_expr_real_1): Likewise.
2750 (reduce_to_bit_field_precision): Likewise.
2751 (const_vector_from_tree): Likewise.
2752 * final.c: Include wide-int-print.h.
2753 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2754 * fixed-value.c: Include wide-int.h.
2755 (fixed_from_string): Use wide-int interfaces.
2756 (fixed_to_decimal): Likewise.
2757 (fixed_convert_from_real): Likewise.
2758 (real_convert_from_fixed): Likewise.
2759 * fold-const.h (mem_ref_offset): Return an offset_int.
2760 (div_if_zero_remainder): Remove code parameter.
2761 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2762 Use wide-int interfaces.
2763 (may_negate_without_overflow_p): Use wide-int interfaces.
2764 (negate_expr_p): Likewise.
2765 (fold_negate_expr): Likewise.
2766 (int_const_binop_1): Likewise.
2767 (const_binop): Likewise.
2768 (fold_convert_const_int_from_int): Likewise.
2769 (fold_convert_const_int_from_real): Likewise.
2770 (fold_convert_const_int_from_fixed): Likewise.
2771 (fold_convert_const_fixed_from_int): Likewise.
2772 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2773 (sign_bit_p): Use wide-int interfaces.
2774 (make_range_step): Likewise.
2775 (build_range_check): Likewise. Pass an integer of the correct type
2776 instead of using integer_one_node.
2777 (range_predecessor): Pass an integer of the correct type instead
2778 of using integer_one_node.
2779 (range_successor): Likewise.
2780 (merge_ranges): Likewise.
2781 (unextend): Use wide-int interfaces.
2782 (extract_muldiv_1): Likewise.
2783 (fold_div_compare): Likewise.
2784 (fold_single_bit_test): Likewise.
2785 (fold_sign_changed_comparison): Likewise.
2786 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2787 (fold_plusminus_mult_expr): Use wide-int interfaces.
2788 (native_encode_int): Likewise.
2789 (native_interpret_int): Likewise.
2790 (fold_unary_loc): Likewise.
2791 (pointer_may_wrap_p): Likewise.
2792 (size_low_cst): Likewise.
2793 (mask_with_tz): Likewise.
2794 (fold_binary_loc): Likewise.
2795 (fold_ternary_loc): Likewise.
2796 (multiple_of_p): Likewise.
2797 (tree_call_nonnegative_warnv_p): Update calls to
2798 tree_int_cst_min_precision and real_from_integer.
2799 (fold_negate_const): Use wide-int interfaces.
2800 (fold_abs_const): Likewise.
2801 (fold_relational_const): Use tree_int_cst_lt.
2802 (round_up_loc): Use wide-int interfaces.
2803 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2804 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2805 * gengtype.c: Remove include of double-int.h.
2806 (do_typedef): Use wide-int interfaces.
2807 (open_base_files): Add wide-int.h.
2808 (main): Add offset_int and widest_int typedefs.
2809 * gengtype-lex.l: Handle "^".
2810 (CXX_KEYWORD): Add "static".
2811 * gengtype-parse.c (require3): New.
2812 (require_template_declaration): Handle constant template arguments
2813 and nested templates.
2814 * gengtype-state.c: Don't include "double-int.h".
2815 * genpreds.c (write_one_predicate_function): Update comment.
2816 (write_tm_constrs_h): Add check for hval and lval use in
2818 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2819 (add_to_sequence): Likewise.
2820 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2821 and const_double_operand.
2822 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2824 * gimple-fold.c (get_base_constructor): Likewise.
2825 (fold_array_ctor_reference): Likewise.
2826 (fold_nonarray_ctor_reference): Likewise.
2827 (fold_const_aggregate_ref_1): Likewise.
2828 (gimple_val_nonnegative_real_p): Likewise.
2829 (gimple_fold_indirect_ref): Likewise.
2830 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2831 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2832 (struct slsr_cand_d): Change index to be widest_int.
2833 (struct incr_info_d): Change incr to be widest_int.
2834 (alloc_cand_and_find_basis): Use wide-int interfaces.
2835 (slsr_process_phi): Likewise.
2836 (backtrace_base_for_ref): Likewise. Return a widest_int.
2837 (restructure_reference): Take a widest_int instead of a double_int.
2838 (slsr_process_ref): Use wide-int interfaces.
2839 (create_mul_ssa_cand): Likewise.
2840 (create_mul_imm_cand): Likewise.
2841 (create_add_ssa_cand): Likewise.
2842 (create_add_imm_cand): Take a widest_int instead of a double_int.
2843 (slsr_process_add): Use wide-int interfaces.
2844 (slsr_process_cast): Likewise.
2845 (slsr_process_copy): Likewise.
2846 (dump_candidate): Likewise.
2847 (dump_incr_vec): Likewise.
2848 (replace_ref): Likewise.
2849 (cand_increment): Likewise. Return a widest_int.
2850 (cand_abs_increment): Likewise.
2851 (replace_mult_candidate): Take a widest_int instead of a double_int.
2852 (replace_unconditional_candidate): Use wide-int interfaces.
2853 (incr_vec_index): Take a widest_int instead of a double_int.
2854 (create_add_on_incoming_edge): Likewise.
2855 (create_phi_basis): Use wide-int interfaces.
2856 (replace_conditional_candidate): Likewise.
2857 (record_increment): Take a widest_int instead of a double_int.
2858 (record_phi_increments): Use wide-int interfaces.
2859 (phi_incr_cost): Take a widest_int instead of a double_int.
2860 (lowest_cost_path): Likewise.
2861 (total_savings): Likewise.
2862 (analyze_increments): Use wide-int interfaces.
2863 (ncd_with_phi): Take a widest_int instead of a double_int.
2864 (ncd_of_cand_and_phis): Likewise.
2865 (nearest_common_dominator_for_cands): Likewise.
2866 (insert_initializers): Use wide-int interfaces.
2867 (all_phi_incrs_profitable): Likewise.
2868 (replace_one_candidate): Likewise.
2869 (replace_profitable_candidates): Likewise.
2870 * godump.c: Include wide-int-print.h.
2871 (go_output_typedef): Use wide-int interfaces.
2872 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2873 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2874 (build_loop_iteration_domains): Likewise.
2875 * hooks.h: Include wide-int.h rather than double-int.h.
2876 (hook_bool_dint_dint_uint_bool_true): Delete.
2877 (hook_bool_wint_wint_uint_bool_true): Declare.
2878 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2879 (hook_bool_wint_wint_uint_bool_true): New.
2880 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2882 (ubsan_expand_si_overflow_mul_check): Likewise.
2883 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2884 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2885 (get_ancestor_addr_info): Likewise.
2886 (ipa_modify_call_arguments): Likewise.
2887 * loop-doloop.c (doloop_modify): Likewise.
2888 (doloop_optimize): Likewise.
2889 * loop-iv.c (iv_number_of_iterations): Likewise.
2890 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2891 (unroll_loop_constant_iterations): Likewise.
2892 (decide_unroll_runtime_iterations): Likewise.
2893 (unroll_loop_runtime_iterations): Likewise.
2894 (decide_peel_simple): Likewise.
2895 (decide_unroll_stupid): Likewise.
2896 * lto-streamer-in.c (streamer_read_wi): Add.
2897 (input_cfg): Use wide-int interfaces.
2898 (lto_input_tree_1): Likewise.
2899 * lto-streamer-out.c (streamer_write_wi): Add.
2900 (hash_tree): Use wide-int interfaces.
2901 (output_cfg): Likewise.
2902 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2903 (GTFILES): Add wide-int.h and signop.h.
2904 (TAGS): Look for .cc files too.
2905 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2906 * optabs.c (expand_subword_shift): Likewise.
2907 (expand_doubleword_shift): Likewise.
2908 (expand_absneg_bit): Likewise.
2909 (expand_copysign_absneg): Likewise.
2910 (expand_copysign_bit): Likewise.
2911 * postreload.c (reload_cse_simplify_set): Likewise.
2912 * predict.c (predict_iv_comparison): Likewise.
2913 * pretty-print.h: Include wide-int-print.h.
2915 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2916 * print-tree.c: Include wide-int-print.h.
2917 (print_node_brief): Use wide-int interfaces.
2918 (print_node): Likewise.
2919 * read-rtl.c (validate_const_wide_int): New.
2920 (read_rtx_code): Add CONST_WIDE_INT case.
2921 * real.c: Include wide-int.h.
2922 (real_to_integer2): Delete.
2923 (real_to_integer): New function, returning a wide_int.
2924 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2925 (ten_to_ptwo): Update call to real_from_integer.
2926 (real_digit): Likewise.
2927 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2928 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2929 (REAL_VALUE_TO_INT): Delete.
2930 (real_to_integer): Declare a wide-int form.
2931 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2932 * recog.c (const_int_operand): Improve comment.
2933 (const_scalar_int_operand): New.
2934 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2935 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2936 (split_double): Likewise.
2937 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2938 (rtx_size): Likewise.
2939 (rtx_alloc_stat_v): New.
2940 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2941 (cwi_output_hex): New.
2942 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2943 (cwi_check_failed_bounds): New.
2944 * rtl.def (CONST_WIDE_INT): New.
2945 * rtl.h: Include <utility> and wide-int.h.
2946 (struct hwivec_def): New.
2947 (CWI_GET_NUM_ELEM): New.
2948 (CWI_PUT_NUM_ELEM): New.
2949 (struct rtx_def): Add num_elem and hwiv.
2950 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2951 (CASE_CONST_UNIQUE): Likewise.
2952 (CASE_CONST_ANY): Likewise.
2953 (CONST_SCALAR_INT_P): Likewise.
2954 (CONST_WIDE_INT_P): New.
2956 (HWIVEC_CHECK): New.
2957 (cwi_check_failed_bounds): New.
2959 (HWIVEC_CHECK): New.
2960 (CONST_WIDE_INT_VEC) New.
2961 (CONST_WIDE_INT_NUNITS) New.
2962 (CONST_WIDE_INT_ELT) New.
2963 (rtx_mode_t): New type.
2964 (wi::int_traits <rtx_mode_t>): New.
2966 (wi::min_value): New.
2967 (wi::max_value): New.
2969 (const_wide_int_alloc): New.
2970 (immed_wide_int_const): New.
2971 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2972 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2973 * signop.h: New file.
2974 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2975 (simplify_const_unary_operation): Use wide-int interfaces.
2976 (simplify_binary_operation_1): Likewise.
2977 (simplify_const_binary_operation): Likewise.
2978 (simplify_const_relational_operation): Likewise.
2979 (simplify_immed_subreg): Likewise.
2980 * stmt.c (expand_case): Likewise.
2981 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2982 signop rather than a bool.
2983 * stor-layout.c (layout_type): Use wide-int interfaces.
2984 (initialize_sizetypes): Update calls to
2985 set_min_and_max_values_for_integral_type.
2986 (set_min_and_max_values_for_integral_type): Take a signop rather
2987 than a bool. Use wide-int interfaces.
2988 (fixup_signed_type): Update accordingly. Remove
2989 HOST_BITS_PER_DOUBLE_INT limit.
2990 (fixup_unsigned_type): Likewise.
2991 * system.h (STATIC_CONSTANT_P): New.
2992 (STATIC_ASSERT): New.
2993 * target.def (can_use_doloop_p): Take widest_ints rather than
2995 * target.h: Include wide-int.h rather than double-int.h.
2996 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2998 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2999 rather than INT_CST_LT_UNSIGNED.
3000 (can_use_doloop_if_innermost): Take widest_ints rather than
3002 * tree-affine.c: Include wide-int-print.h.
3003 (double_int_ext_for_comb): Delete.
3004 (wide_int_ext_for_comb): New.
3005 (aff_combination_zero): Use wide-int interfaces.
3006 (aff_combination_const): Take a widest_int instead of a double_int.
3007 (aff_combination_elt): Use wide-int interfaces.
3008 (aff_combination_scale): Take a widest_int instead of a double_int.
3009 (aff_combination_add_elt): Likewise.
3010 (aff_combination_add_cst): Likewise.
3011 (aff_combination_add): Use wide-int interfaces.
3012 (aff_combination_convert): Likewise.
3013 (tree_to_aff_combination): Likewise.
3014 (add_elt_to_tree): Take a widest_int instead of a double_int.
3015 (aff_combination_to_tree): Use wide-int interfaces.
3016 (aff_combination_remove_elt): Likewise.
3017 (aff_combination_add_product): Take a widest_int instead of
3019 (aff_combination_mult): Use wide-int interfaces.
3020 (aff_combination_expand): Likewise.
3021 (double_int_constant_multiple_p): Delete.
3022 (wide_int_constant_multiple_p): New.
3023 (aff_combination_constant_multiple_p): Take a widest_int pointer
3024 instead of a double_int pointer.
3025 (print_aff): Use wide-int interfaces.
3026 (get_inner_reference_aff): Take a widest_int pointer
3027 instead of a double_int pointer.
3028 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3029 * tree-affine.h: Include wide-int.h.
3030 (struct aff_comb_elt): Change type of coef to widest_int.
3031 (struct affine_tree_combination): Change type of offset to widest_int.
3032 (double_int_ext_for_comb): Delete.
3033 (wide_int_ext_for_comb): New.
3034 (aff_combination_const): Use widest_int instead of double_int.
3035 (aff_combination_scale): Likewise.
3036 (aff_combination_add_elt): Likewise.
3037 (aff_combination_constant_multiple_p): Likewise.
3038 (get_inner_reference_aff): Likewise.
3039 (aff_comb_cannot_overlap_p): Likewise.
3040 (aff_combination_zero_p): Use wide-int interfaces.
3041 * tree.c: Include tree.h.
3042 (init_ttree): Use make_int_cst.
3043 (tree_code_size): Removed code for INTEGER_CST case.
3044 (tree_size): Add INTEGER_CST case.
3045 (make_node_stat): Update comment.
3046 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3047 (build_int_cst_type): Use wide-int interfaces.
3048 (double_int_to_tree): Likewise.
3049 (double_int_fits_to_tree_p): Delete.
3050 (force_fit_type_double): Delete.
3051 (force_fit_type): New.
3052 (int_cst_hash_hash): Use wide-int interfaces.
3053 (int_cst_hash_eq): Likewise.
3054 (build_int_cst_wide): Delete.
3055 (wide_int_to_tree): New.
3056 (cache_integer_cst): Use wide-int interfaces.
3057 (build_low_bits_mask): Likewise.
3058 (cst_and_fits_in_hwi): Likewise.
3059 (real_value_from_int_cst): Likewise.
3060 (make_int_cst_stat): New.
3061 (integer_zerop): Use wide_int interfaces.
3062 (integer_onep): Likewise.
3063 (integer_all_onesp): Likewise.
3064 (integer_pow2p): Likewise.
3065 (integer_nonzerop): Likewise.
3066 (tree_log2): Likewise.
3067 (tree_floor_log2): Likewise.
3068 (tree_ctz): Likewise.
3069 (int_size_in_bytes): Likewise.
3070 (mem_ref_offset): Return an offset_int rather than a double_int.
3071 (build_type_attribute_qual_variant): Use wide_int interfaces.
3072 (type_hash_eq): Likewise
3073 (tree_int_cst_equal): Likewise.
3074 (tree_int_cst_lt): Delete.
3075 (tree_int_cst_compare): Likewise.
3076 (tree_fits_shwi_p): Use wide_int interfaces.
3077 (tree_fits_uhwi_p): Likewise.
3078 (tree_int_cst_sign_bit): Likewise.
3079 (tree_int_cst_sgn): Likewise.
3080 (tree_int_cst_min_precision): Take a signop rather than a bool.
3081 (simple_cst_equal): Use wide_int interfaces.
3082 (compare_tree_int): Likewise.
3083 (iterative_hash_expr): Likewise.
3084 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3086 (get_type_static_bounds): Use wide_int interfaces.
3087 (tree_int_cst_elt_check_failed): New.
3088 (build_common_tree_nodes): Reordered to set prec before filling in
3090 (int_cst_value): Check cst_and_fits_in_hwi.
3091 (widest_int_cst_value): Use wide_int interfaces.
3092 (upper_bound_in_type): Likewise.
3093 (lower_bound_in_type): Likewise.
3094 (num_ending_zeros): Likewise.
3095 (drop_tree_overflow): Likewise.
3096 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3097 (gen_conditions_for_pow_cst_base): Likewise.
3098 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3099 (group_case_labels_stmt): Use wide-int interfaces.
3100 (verify_gimple_assign_binary): Likewise.
3101 (print_loop): Likewise.
3102 * tree-chrec.c (tree_fold_binomial): Likewise.
3103 * tree-core.h (struct tree_base): Add int_length.
3104 (struct tree_int_cst): Change rep of value.
3105 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3106 (dr_may_alias_p): Likewise.
3107 (max_stmt_executions_tree): Likewise.
3108 * tree.def (INTEGER_CST): Update comment.
3109 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3110 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3111 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3112 (dequeue_and_dump): Use wide-int interfaces.
3113 * tree.h: Include wide-int.h.
3114 (NULL_TREE): Moved to earlier loc in file.
3115 (TREE_INT_CST_ELT_CHECK): New.
3116 (tree_int_cst_elt_check_failed): New.
3118 (TREE_INT_CST): Delete.
3119 (TREE_INT_CST_LOW): Use wide-int interfaces.
3120 (TREE_INT_CST_HIGH): Delete.
3121 (TREE_INT_CST_NUNITS): New.
3122 (TREE_INT_CST_EXT_NUNITS): Likewise.
3123 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3124 (TREE_INT_CST_ELT): Likewise.
3125 (INT_CST_LT): Delete.
3126 (tree_int_cst_elt_check): New (two forms).
3127 (type_code_size): Update comment.
3128 (make_int_cst_stat, make_int_cst): New.
3129 (tree_to_double_int): Delete.
3130 (double_int_fits_to_tree_p): Delete.
3131 (force_fit_type_double): Delete.
3132 (build_int_cstu): Replace with out-of-line function.
3133 (build_int_cst_wide): Delete.
3134 (tree_int_cst_lt): Define inline.
3135 (tree_int_cst_le): New.
3136 (tree_int_cst_compare): Define inline.
3137 (tree_int_cst_min_precision): Take a signop rather than a bool.
3138 (wi::int_traits <const_tree>): New.
3139 (wi::int_traits <tree>): New.
3140 (wi::extended_tree): New.
3141 (wi::int_traits <wi::extended_tree>): New.
3142 (wi::to_widest): New.
3143 (wi::to_offset): New.
3144 (wi::fits_to_tree_p): New.
3145 (wi::min_value): New.
3146 (wi::max_value): New.
3147 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3148 (copy_tree_body_r): Likewise.
3149 * tree-object-size.c (compute_object_offset): Likewise.
3150 (addr_object_size): Likewise.
3151 * tree-predcom.c: Include wide-int-print.h.
3152 (struct dref_d): Change type of offset to widest_int.
3153 (dump_dref): Call wide-int printer.
3154 (aff_combination_dr_offset): Use wide-int interfaces.
3155 (determine_offset): Take a widest_int pointer rather than a
3157 (split_data_refs_to_components): Use wide-int interfaces.
3158 (suitable_component_p): Likewise.
3159 (order_drefs): Likewise.
3160 (add_ref_to_chain): Likewise.
3161 (valid_initializer_p): Likewise.
3162 (determine_roots_comp): Likewise.
3163 * tree-pretty-print.c: Include wide-int-print.h.
3164 (dump_generic_node): Use wide-int interfaces.
3165 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3166 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3167 (move_fixed_address_to_symbol): Likewise.
3168 (move_hint_to_base): Likewise.
3169 (move_pointer_to_base): Likewise.
3170 (move_variant_to_index): Likewise.
3171 (most_expensive_mult_to_index): Likewise.
3172 (addr_to_parts): Likewise.
3173 (copy_ref_info): Likewise.
3174 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3175 (indirect_refs_may_alias_p): Likewise.
3176 (stmt_kills_ref_p_1): Likewise.
3177 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3178 * tree-ssa-ccp.c: Update comment at top of file. Include
3180 (struct prop_value_d): Change type of mask to widest_int.
3181 (extend_mask): New function.
3182 (dump_lattice_value): Use wide-int interfaces.
3183 (get_default_value): Likewise.
3184 (set_constant_value): Likewise.
3185 (set_value_varying): Likewise.
3186 (valid_lattice_transition): Likewise.
3187 (set_lattice_value): Likewise.
3188 (value_to_double_int): Delete.
3189 (value_to_wide_int): New.
3190 (get_value_from_alignment): Use wide-int interfaces.
3191 (get_value_for_expr): Likewise.
3192 (do_dbg_cnt): Likewise.
3193 (ccp_finalize): Likewise.
3194 (ccp_lattice_meet): Likewise.
3195 (bit_value_unop_1): Use widest_ints rather than double_ints.
3196 (bit_value_binop_1): Likewise.
3197 (bit_value_unop): Use wide-int interfaces.
3198 (bit_value_binop): Likewise.
3199 (bit_value_assume_aligned): Likewise.
3200 (evaluate_stmt): Likewise.
3201 (ccp_fold_stmt): Likewise.
3202 (visit_cond_stmt): Likewise.
3203 (ccp_visit_stmt): Likewise.
3204 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3205 (constant_pointer_difference): Likewise.
3206 (associate_pointerplus): Likewise.
3207 (combine_conversions): Likewise.
3208 * tree-ssa-loop.h: Include wide-int.h.
3209 (struct tree_niter_desc): Change type of max to widest_int.
3210 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3211 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3212 (remove_redundant_iv_tests): Likewise.
3213 (canonicalize_loop_induction_variables): Likewise.
3214 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3215 (constant_multiple_of): Take a widest_int pointer instead of
3216 a double_int pointer.
3217 (get_computation_aff): Use wide-int interfaces.
3218 (ptr_difference_cost): Likewise.
3219 (difference_cost): Likewise.
3220 (get_loop_invariant_expr_id): Likewise.
3221 (get_computation_cost_at): Likewise.
3222 (iv_elimination_compare_lt): Likewise.
3223 (may_eliminate_iv): Likewise.
3224 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3225 instead of double_int.
3226 (max_loop_iterations): Likewise.
3227 (max_stmt_executions): Likewise.
3228 (estimated_stmt_executions): Likewise.
3229 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3230 (split_to_var_and_offset): Use wide-int interfaces.
3231 (determine_value_range): Likewise.
3232 (bound_difference_of_offsetted_base): Likewise.
3233 (bounds_add): Take a widest_int instead of a double_int.
3234 (number_of_iterations_ne_max): Use wide-int interfaces.
3235 (number_of_iterations_ne): Likewise.
3236 (number_of_iterations_lt_to_ne): Likewise.
3237 (assert_loop_rolls_lt): Likewise.
3238 (number_of_iterations_lt): Likewise.
3239 (number_of_iterations_le): Likewise.
3240 (number_of_iterations_cond): Likewise.
3241 (number_of_iterations_exit): Likewise.
3242 (finite_loop_p): Likewise.
3243 (derive_constant_upper_bound_assign): Likewise.
3244 (derive_constant_upper_bound): Return a widest_int.
3245 (derive_constant_upper_bound_ops): Likewise.
3246 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3247 (record_estimate): Take a widest_int rather than a double_int.
3248 (record_nonwrapping_iv): Use wide-int interfaces.
3249 (double_int_cmp): Delete.
3250 (wide_int_cmp): New.
3251 (bound_index): Take a widest_int rather than a double_int.
3252 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3253 (maybe_lower_iteration_bound): Likewise.
3254 (estimate_numbers_of_iterations_loop): Likewise.
3255 (estimated_loop_iterations): Take a widest_int pointer than than
3256 a double_int pointer.
3257 (estimated_loop_iterations_int): Use wide-int interfaces.
3258 (max_loop_iterations): Take a widest_int pointer than than
3259 a double_int pointer.
3260 (max_loop_iterations_int): Use wide-int interfaces.
3261 (max_stmt_executions): Take a widest_int pointer than than
3262 a double_int pointer.
3263 (estimated_stmt_executions): Likewise.
3264 (n_of_executions_at_most): Use wide-int interfaces.
3265 (scev_probably_wraps_p): Likewise.
3266 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3268 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3270 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3271 double_ints. Adjust for trailing_wide_ints <3> representation.
3272 (set_nonzero_bits): Likewise.
3273 (get_range_info): Return wide_ints rather than double_ints.
3274 Adjust for trailing_wide_ints <3> representation.
3275 (get_nonzero_bits): Likewise.
3276 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3278 * tree-ssanames.h (struct range_info_def): Replace min, max and
3279 nonzero_bits with a trailing_wide_ints <3>.
3280 (set_range_info): Use wide_int_refs rather than double_ints.
3281 (set_nonzero_bits): Likewise.
3282 (get_range_info): Return wide_ints rather than double_ints.
3283 (get_nonzero_bits): Likewise.
3284 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3285 * tree-ssa-pre.c (phi_translate_1): Likewise.
3286 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3287 (acceptable_pow_call): Likewise.
3288 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3290 (vn_reference_fold_indirect): Likewise.
3291 (vn_reference_maybe_forwprop_address): Likewise.
3292 (valueize_refs_1): Likewise.
3293 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3294 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3295 tree_int_cst_lt and tree_int_cst_le.
3296 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3298 (streamer_alloc_tree): Likewise.
3299 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3300 (streamer_write_tree_header): Likewise.
3301 (streamer_write_integer_cst): Likewise.
3302 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3303 (build_constructors): Likewise.
3304 (array_value_type): Likewise.
3305 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3306 (vect_check_gather): Likewise.
3307 * tree-vect-generic.c (build_replicated_const): Likewise.
3308 (expand_vector_divmod): Likewise.
3309 * tree-vect-loop.c (vect_transform_loop): Likewise.
3310 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3311 (vect_do_peeling_for_alignment): Likewise.
3312 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3313 * tree-vrp.c: Include wide-int.h.
3314 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3315 (extract_range_from_assert): Use wide-int interfaces.
3316 (vrp_int_const_binop): Likewise.
3317 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3318 double_int pointers.
3319 (ranges_from_anti_range): Use wide-int interfaces.
3320 (quad_int_cmp): Delete.
3321 (quad_int_pair_sort): Likewise.
3322 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3323 (extract_range_from_unary_expr_1): Likewise.
3324 (adjust_range_with_scev): Likewise.
3325 (masked_increment): Take and return wide_ints rather than double_ints.
3326 (register_edge_assert_for_2): Use wide-int interfaces.
3327 (check_array_ref): Likewise.
3328 (search_for_addr_array): Likewise.
3329 (maybe_set_nonzero_bits): Likewise.
3330 (union_ranges): Pass an integer of the correct type instead of
3331 using integer_one_node.
3332 (intersect_ranges): Likewise.
3333 (simplify_truth_ops_using_ranges): Likewise.
3334 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3335 (range_fits_type_p): Likewise.
3336 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3338 (simplify_conversion_using_ranges): Use wide-int interfaces.
3339 (simplify_float_conversion_using_ranges): Likewise.
3340 (vrp_finalize): Likewise.
3341 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3342 (gimple_stringops_transform): Likewise.
3343 * varasm.c (decode_addr_const): Likewise.
3344 (const_hash_1): Likewise.
3345 (const_rtx_hash_1): Likewise
3346 (output_constant): Likewise.
3347 (array_size_for_constructor): Likewise.
3348 (output_constructor_regular_field): Likewise.
3349 (output_constructor_bitfield): Likewise.
3350 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3351 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3353 * gencheck.c: Define BITS_PER_UNIT.
3356 * wide-int-print.cc: New.
3357 * wide-int-print.h: New.
3359 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3361 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3363 2014-05-06 Richard Biener <rguenther@suse.de>
3365 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3366 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3367 (TODO_verify_all): Adjust.
3368 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3369 TODO_verify_stmts and TODO_verify_rtl_sharing.
3370 * bb-reorder.c: Likewise.
3371 * cfgexpand.c: Likewise.
3372 * cprop.c: Likewise.
3374 * function.c: Likewise.
3375 * fwprop.c: Likewise.
3377 * gimple-ssa-isolate-paths.c: Likewise.
3378 * gimple-ssa-strength-reduction.c: Likewise.
3379 * ipa-split.c: Likewise.
3380 * loop-init.c: Likewise.
3381 * loop-unroll.c: Likewise.
3382 * lower-subreg.c: Likewise.
3383 * modulo-sched.c: Likewise.
3384 * postreload-gcse.c: Likewise.
3385 * predict.c: Likewise.
3386 * recog.c: Likewise.
3387 * sched-rgn.c: Likewise.
3388 * store-motion.c: Likewise.
3389 * tracer.c: Likewise.
3390 * trans-mem.c: Likewise.
3391 * tree-call-cdce.c: Likewise.
3392 * tree-cfg.c: Likewise.
3393 * tree-cfgcleanup.c: Likewise.
3394 * tree-complex.c: Likewise.
3395 * tree-eh.c: Likewise.
3396 * tree-emutls.c: Likewise.
3397 * tree-if-conv.c: Likewise.
3398 * tree-into-ssa.c: Likewise.
3399 * tree-loop-distribution.c: Likewise.
3400 * tree-object-size.c: Likewise.
3401 * tree-parloops.c: Likewise.
3402 * tree-pass.h: Likewise.
3403 * tree-sra.c: Likewise.
3404 * tree-ssa-ccp.c: Likewise.
3405 * tree-ssa-copy.c: Likewise.
3406 * tree-ssa-copyrename.c: Likewise.
3407 * tree-ssa-dce.c: Likewise.
3408 * tree-ssa-dom.c: Likewise.
3409 * tree-ssa-dse.c: Likewise.
3410 * tree-ssa-forwprop.c: Likewise.
3411 * tree-ssa-ifcombine.c: Likewise.
3412 * tree-ssa-loop-ch.c: Likewise.
3413 * tree-ssa-loop-ivcanon.c: Likewise.
3414 * tree-ssa-loop.c: Likewise.
3415 * tree-ssa-math-opts.c: Likewise.
3416 * tree-ssa-phiopt.c: Likewise.
3417 * tree-ssa-phiprop.c: Likewise.
3418 * tree-ssa-pre.c: Likewise.
3419 * tree-ssa-reassoc.c: Likewise.
3420 * tree-ssa-sink.c: Likewise.
3421 * tree-ssa-strlen.c: Likewise.
3422 * tree-ssa-tail-merge.c: Likewise.
3423 * tree-ssa-uncprop.c: Likewise.
3424 * tree-switch-conversion.c: Likewise.
3425 * tree-tailcall.c: Likewise.
3426 * tree-vect-generic.c: Likewise.
3427 * tree-vectorizer.c: Likewise.
3428 * tree-vrp.c: Likewise.
3430 * var-tracking.c: Likewise.
3431 * bt-load.c: Likewise.
3432 * cfgcleanup.c: Likewise.
3433 * combine-stack-adj.c: Likewise.
3434 * combine.c: Likewise.
3435 * compare-elim.c: Likewise.
3436 * config/epiphany/resolve-sw-modes.c: Likewise.
3437 * config/i386/i386.c: Likewise.
3438 * config/mips/mips.c: Likewise.
3439 * config/s390/s390.c: Likewise.
3440 * config/sh/sh_treg_combine.cc: Likewise.
3441 * config/sparc/sparc.c: Likewise.
3444 * final.c: Likewise.
3445 * ifcvt.c: Likewise.
3446 * mode-switching.c: Likewise.
3447 * passes.c: Likewise.
3448 * postreload.c: Likewise.
3450 * reg-stack.c: Likewise.
3451 * regcprop.c: Likewise.
3452 * regrename.c: Likewise.
3455 2014-05-06 Richard Biener <rguenther@suse.de>
3458 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3459 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3461 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3464 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3466 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3467 Tom de Vries <tom@codesourcery.com>
3469 * target.def (call_fusage_contains_non_callee_clobbers): New
3471 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3473 (@node Miscellaneous Register Hooks): New node.
3474 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3475 * doc/tm.texi: Regenerate.
3477 2014-05-05 Marek Polacek <polacek@redhat.com>
3480 * opts.c (common_handle_option): Call error_at instead of warning_at.
3482 2014-05-05 Richard Biener <rguenther@suse.de>
3484 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3485 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3486 under the TODO_verify_il umbrella.
3488 2014-05-05 Richard Biener <rguenther@suse.de>
3490 * passes.c (execute_function_todo): Move TODO_verify_flow under
3491 the TODO_verify_ul umbrella.
3493 2014-05-05 Richard Biener <rguenther@suse.de>
3496 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3497 X & CST away from a CST that is the mask of a mode.
3499 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3501 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3502 int argument to enum machine_mode.
3503 (picochip_class_max_nregs): Ditto.
3504 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3505 (picochip_class_max_nregs): Ditto.
3507 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3509 * target.def: Add new target hook.
3510 * doc/tm.texi: Regenerate.
3511 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3512 * targhooks.c (default_keep_leaf_when_profiled): New function.
3514 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3515 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3517 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3519 PR tree-optimization/60363
3520 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3521 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3522 (update_destination_phis): New parameter.
3523 (create_edge_and_update_destination_phis): Ditto.
3524 (ssa_fix_duplicate_block_edges): Pass new arguments.
3525 (thread_single_edge): Ditto.
3527 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3529 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3530 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3531 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3532 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3533 Use RS6000_BTM_HARD_FLOAT.
3534 (BU_MISC_2): Likewise.
3535 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3536 RS6000_BTM_HARD_FLOAT.
3537 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3539 (rs6000_invalid_builtin): Add hard floating builtin support.
3540 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3541 hard float builtins.
3542 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3544 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3546 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3547 Add missing function* argument.
3549 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3551 * lra-constraints.c (valid_address_p): Move earlier in file.
3552 Add a constraint argument to the address_info version.
3553 (satisfies_memory_constraint_p): New function.
3554 (satisfies_address_constraint_p): Likewise.
3555 (process_alt_operands, curr_insn_transform): Use them.
3556 (process_address): Pass the constraint to valid_address_p when
3557 checking address operands.
3559 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3561 * config/mips/mips.c (mips_isa_rev): New variable.
3562 (mips_set_architecture): Set it.
3563 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3565 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3566 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3567 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3568 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3569 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3570 conditions in terms of mips_isa_rev.
3571 (mips_isa_rev): Declare.
3573 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3575 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3576 (prob_unlikely, prob_likely): Make variables const.
3578 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3580 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3582 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3584 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3586 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3588 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3589 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3590 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3592 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3593 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3595 Replace usage of ROUND_REG with sh_round_reg.
3596 Use CEIL instead of ROUND_ADVANCE.
3598 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3601 * config/sh/sh.c: Include stdlib headers before everything else.
3603 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3605 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3606 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3607 (gimplify_adjust_omp_clauses): Simd region is never
3608 directly nested in combined parallel. Instead, for linear
3609 with copyin/copyout, if in combined for simd loop, make decl
3610 firstprivate/lastprivate on OMP_FOR.
3611 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3612 expand_omp_for_static_chunk): When setting endvar, also set
3613 fd->loop.v to the same value.
3615 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3617 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3619 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3621 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3624 2014-05-02 Marek Polacek <polacek@redhat.com>
3626 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3628 2014-05-02 Kito Cheng <kito@0xlab.org>
3630 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3631 to a C expression marco.
3632 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3633 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3634 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3635 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3636 HONOR_REG_ALLOC_ORDER.
3637 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3639 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3641 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3643 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3645 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3647 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3649 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3650 (convert_scalar_cond_reduction): Likewise.
3651 (predicate_scalar_phi): Add recognition and transformation
3652 of simple conditioanl reduction to be vectorizable.
3654 2014-05-01 Marek Polacek <polacek@redhat.com>
3657 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3659 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3661 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3662 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3663 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3664 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3665 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3666 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3667 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3668 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3670 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3672 * config/arc/arc.opt (mlra): Move comment above option name
3673 to avoid mis-parsing as language options.
3675 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3677 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3678 * config/sol2.h: ... here.
3679 * config/sol2-10.h: Remove.
3681 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3682 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3683 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3684 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3685 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3686 * config/sol2.h: ... here.
3687 (SECTION_NAME_FORMAT): Don't redefine.
3688 (STARTFILE_ARCH32_SPEC): Rename to ...
3689 (STARTFILE_ARCH_SPEC): ... this.
3690 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3691 * config/sparc/sol2.h: ... here.
3692 (SECTION_NAME_FORMAT): Don't undef.
3693 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3694 (SUBTARGET_EXTRA_SPECS): Remove.
3695 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3697 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3698 (MD_STARTFILE_PREFIX): Remove.
3699 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3700 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3701 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3702 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3703 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3704 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3705 * config/i386/sol2.h: ... here.
3706 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3707 * config/i386/sol2-bi.h: Remove.
3708 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3709 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3711 * config/i386/t-sol2-64: Rename to ...
3712 * config/i386/t-sol2: ... this.
3713 * config/sparc/t-sol2-64: Rename to ...
3714 * config/sparc/t-sol2: ... this.
3716 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3717 sol2_tm_file_head, sol2_tm_file_tail.
3718 Include ${cpu_type}/sol2.h before sol2.h.
3720 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3721 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3722 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3723 Reflect i386/t-sol2-64 renaming.
3724 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3725 Reflect sparc/t-sol2-64 renaming.
3727 2014-04-30 Richard Biener <rguenther@suse.de>
3729 * passes.c (execute_function_todo): Move TODO_verify_stmts
3730 and TODO_verify_ssa under the TODO_verify_il umbrella.
3731 * tree-ssa.h (verify_ssa): Adjust prototype.
3732 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3733 we should verify SSA operands.
3734 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3735 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3736 whether we should verify whether not throwing stmts have EH info.
3737 * graphite-scop-detection.c (create_sese_edges): Adjust.
3738 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3739 * tree-eh.c (lower_try_finally_switch): Do not add the
3740 default case label twice.
3742 2014-04-30 Marek Polacek <polacek@redhat.com>
3744 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3745 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3746 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3747 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3749 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3751 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3752 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3753 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3754 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3755 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3756 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3757 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3758 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3760 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3762 * tree-cfg.c (dump_function_to_file): Dump the return type of
3763 functions, in a line to itself before the function body, mimicking
3764 the layout of a C function.
3766 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3768 PR tree-optimization/60971
3769 * tree-tailcall.c (process_assignment): Reject conversions which
3772 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3774 * calls.c (initialize_argument_information): Always treat
3775 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3776 (expand_call): Likewise.
3777 (emit_library_call_calue_1): Likewise.
3778 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3779 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3782 2014-04-29 Nick Clifton <nickc@redhat.com>
3784 * config/msp430/msp430.md (umulsidi): Fix typo.
3785 (mulhisi3): Enable even inside interrupt handlers.
3786 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3787 bigger return address pushed in large mode.
3789 2014-04-29 Nick Clifton <nickc@redhat.com>
3791 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3792 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3794 * config/m32r/m32r.c (init_reg_tables): Likewise.
3795 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3796 enum to hold the modes.
3798 2014-04-29 Richard Biener <rguenther@suse.de>
3800 * dominance.c (free_dominance_info): Add overload with
3802 (dom_info_state): Likewise.
3803 (dom_info_available_p): Likewise.
3804 * basic-block.h (free_dominance_info, dom_info_state,
3805 dom_info_available_p): Declare overloads.
3806 * passes.c (execute_function_todo): Verify that verifiers
3807 don't change dominator info state. Drop dominator info
3808 for IPA pass invocations.
3809 * cgraph.c (release_function_body): Restore asserts that
3810 dominator information is released.
3812 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3814 * doc/invoke.texi: Fix typo.
3815 * tree-vrp.c: Fix typos.
3816 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3818 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3820 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3822 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3824 * config/aarch64/aarch64-builtins.c
3825 (aarch64_types_storestruct_lane_qualifiers): New.
3826 (TYPES_STORESTRUCT_LANE): Likewise.
3827 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3828 (st3_lane): Likewise.
3829 (st4_lane): Likewise.
3830 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3831 (vec_store_lanesci_lane<mode>): Likewise.
3832 (vec_store_lanesxi_lane<mode>): Likewise.
3833 (aarch64_st2_lane<VQ:mode>): Likewise.
3834 (aarch64_st3_lane<VQ:mode>): Likewise.
3835 (aarch64_st4_lane<VQ:mode>): Likewise.
3836 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3837 * config/aarch64/arm_neon.h
3838 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3839 use new macro arguments.
3840 (__ST3_LANE_FUNC): Likewise.
3841 (__ST4_LANE_FUNC): Likewise.
3842 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3843 (V_THREE_ELEM): Likewise.
3844 (V_FOUR_ELEM): Likewise.
3846 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3848 * doc/gimple.texi: Replace the description of the now-defunct
3849 union gimple_statement_d with a diagram showing the
3850 gimple_statement_base class hierarchy and its relationships to
3851 the GSS_ and GIMPLE_ enums.
3853 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3855 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3856 * config/aarch64/aarch64.c
3857 (aarch64_cannot_change_mode_class): Weaken conditions.
3858 (aarch64_modes_tieable_p): New.
3859 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3861 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3863 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3864 (loadsync_<mode>): Change mode.
3865 (load_quadpti, store_quadpti): New.
3866 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3867 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3869 2014-04-28 Martin Jambor <mjambor@suse.cz>
3871 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3872 same alias type as the original statement.
3873 (subreplacement_assignment_data): New type.
3874 (handle_unscalarized_data_in_subtree): New type of parameter,
3875 generate new memory accesses with same alias type as the original
3877 (load_assign_lhs_subreplacements): Likewise.
3878 (sra_modify_constructor_assign): Generate new memory accesses with
3879 same alias type as the original statement.
3881 2014-04-28 Richard Biener <rguenther@suse.de>
3883 * tree-pass.h (TODO_verify_il): Define.
3884 (TODO_verify_all): Complete properly.
3885 * passes.c (execute_function_todo): Move existing loop-closed
3886 SSA verification under TODO_verify_il.
3887 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3888 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3889 Fix tree sharing issue.
3891 2014-04-28 Richard Biener <rguenther@suse.de>
3894 * builtins.def (DEF_C11_BUILTIN): Add.
3895 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3896 * coretypes.h (enum function_class): Add function_c11_misc.
3897 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3898 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3899 (call_may_clobber_ref_p_1): Likewise.
3900 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3901 (mark_all_reaching_defs_necessary_1): Likewise.
3902 (propagate_necessity): Likewise.
3903 (eliminate_unnecessary_stmts): Likewise.
3904 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3906 2014-04-28 Richard Biener <rguenther@suse.de>
3908 * tree-vrp.c (vrp_var_may_overflow): Remove.
3909 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3910 with overflow immediately bump to one before that value and
3911 let iteration figure out overflow status.
3913 2014-04-28 Richard Biener <rguenther@suse.de>
3915 * configure.ac: Do valgrind header checks unconditionally.
3916 Add --enable-valgrind-annotations.
3917 * system.h: Guard valgrind header inclusion with
3918 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3919 * alloc-pool.c (pool_alloc, pool_free): Use
3920 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3921 to guard possibly dead code.
3922 * config.in: Regenerated.
3923 * configure: Likewise.
3925 2014-04-28 Jeff Law <law@redhat.com>
3927 PR tree-optimization/60902
3928 * tree-ssa-threadedge.c
3929 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3930 over real defs when invalidating outputs from statements that do not
3931 produce useful outputs for threading.
3933 2014-04-28 Richard Biener <rguenther@suse.de>
3935 PR tree-optimization/60979
3936 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3937 SCOPs that end in a block with a successor with abnormal
3940 2014-04-28 Richard Biener <rguenther@suse.de>
3942 * tree-pass.h (execute_pass_list): Adjust prototype.
3943 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3944 (do_per_function): Change callback signature, push all actual
3945 work to the callbals.
3946 (do_per_function_toporder): Likewise.
3947 (execute_function_dump): Adjust.
3948 (execute_function_todo): Likewise.
3949 (clear_last_verified): Likewise.
3950 (verify_curr_properties): Likewise.
3951 (update_properties_after_pass): Likewise.
3952 (execute_pass_list_1): Split out from ...
3953 (execute_pass_list): ... here. Adjust.
3954 (execute_ipa_pass_list): Likewise.
3955 * cgraphunit.c (cgraph_add_new_function): Adjust.
3956 (analyze_function): Likewise.
3957 (expand_function): Likewise.
3958 * cgraph.c (release_function_body): Free dominance info
3959 here instead of asserting it was magically freed elsewhere.
3961 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3963 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3964 * configure: Regenerate.
3965 * config/sparc/sparc.opt (muser-mode): New option.
3966 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3968 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3969 * doc/invoke.texi (SPARC options): Document -muser-mode.
3971 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3973 * cselib.c (find_slot_memmode): Delete.
3974 (cselib_hasher): Change compare_type to a struct.
3975 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3977 (preserve_constants_and_equivs): Adjust for new compare_type.
3978 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3979 (wrap_constant): Delete.
3980 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3982 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3984 * doc/install.texi (Building with profile feedback): Remove
3987 2014-04-26 Tom de Vries <tom@codesourcery.com>
3989 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3992 2014-04-25 Cary Coutant <ccoutant@google.com>
3995 * dwarf2out.c (should_move_die_to_comdat): A type definition
3996 can contain a subprogram definition, but don't move it to a
3998 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3999 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4001 (clone_tree_hash): Rename to...
4002 (clone_tree_partial): ...this; change callers. Copy
4003 DW_TAG_subprogram DIEs as declarations.
4004 (copy_decls_walk): Don't copy children of a declaration into a
4007 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4010 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4013 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4015 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4016 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4018 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4021 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4023 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4025 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4027 PR tree-optimization/60930
4028 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4029 creating a multiply candidate by folding two constant
4030 multiplicands when the result overflows.
4032 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4034 PR tree-optimization/60960
4035 * tree-vect-generic.c (expand_vector_operation): Only call
4036 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4038 2014-04-25 Tom de Vries <tom@codesourcery.com>
4040 * expr.c (clobber_reg_mode): New function.
4041 * expr.h (clobber_reg): New function.
4043 2014-04-25 Tom de Vries <tom@codesourcery.com>
4045 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4048 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4049 Tom de Vries <tom@codesourcery.com>
4051 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4053 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4054 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4055 new argument to find_all_hard_reg_sets call.
4057 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4059 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4060 Use HOST_WIDE_INT_C for mask literal.
4061 (aarch_rev16_shleft_mask_imm_p): Likewise.
4063 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4066 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4068 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4070 PR preprocessor/56540
4071 * config/i386/i386-c.c (ix86_target_macros): Define
4072 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4074 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4076 * configure.ac (tga_func): Remove.
4077 (LIB_TLS_SPEC): Remove.
4078 * configure: Regenerate.
4079 * config.in: Regenerate.
4080 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4082 2014-04-25 Richard Biener <rguenther@suse.de>
4085 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4086 call stmt use/clobber sets during stmt walk instead of
4087 walking the possibly incomplete set of caller edges.
4089 2014-04-25 Richard Biener <rguenther@suse.de>
4092 * passes.c (apply_ipa_transforms): Inline into only caller ...
4093 (execute_one_pass): ... here. Properly bring in function
4094 bodies for nodes we want to apply IPA transforms to.
4096 2014-04-24 Cong Hou <congh@google.com>
4098 PR tree-optimization/60896
4099 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4100 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4101 (vect_mark_pattern_stmts): Set the def type of all statements in
4102 PATTERN_DEF_SEQ as vect_internal_def.
4104 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4106 * doc/extend.texi (PowerPC Built-in Functions): Document new
4107 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4108 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4110 * config/rs6000/predicates.md (const_0_to_3_operand): New
4111 predicate to match 0..3 integer constants.
4113 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4114 to support adding miscellaneous builtin functions.
4115 (BU_DFP_MISC_2): Likewise.
4116 (BU_P7_MISC_1): Likewise.
4117 (BU_P7_MISC_2): Likewise.
4118 (BU_P8V_MISC_3): Likewise.
4119 (BU_MISC_1): Likewise.
4120 (BU_MISC_2): Likewise.
4121 (DIVWE): Add extended divide builtin functions.
4124 (DIVWEUO): Likewise.
4128 (DIVDEUO): Likewise.
4129 (DXEX): Add decimal floating-point builtin functions.
4132 (DDEDPDQ): Likewise.
4134 (DENBCDQ): Likewise.
4141 (CDTBCD): Add new BCD builtin functions.
4145 (BCDADD_LT): Likewise.
4146 (BCDADD_EQ): Likewise.
4147 (BCDADD_GT): Likewise.
4148 (BCDADD_OV): Likewise.
4150 (BCDSUB_LT): Likewise.
4151 (BCDSUB_EQ): Likewise.
4152 (BCDSUB_GT): Likewise.
4153 (BCDSUB_OV): Likewise.
4154 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4155 (UNPACK_TD): Likewise.
4156 (PACK_TF): Likewise.
4157 (UNPACK_TF): Likewise.
4158 (UNPACK_TF_0): Likewise.
4159 (UNPACK_TF_1): Likewise.
4160 (PACK_V1TI): Likewise.
4161 (UNPACK_V1TI): Likewise.
4163 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4164 support for decimal floating point builtin functions.
4165 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4166 functions that take constant arguments.
4167 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4168 (rs6000_init_builtins): Setup long double, _Decimal64, and
4169 _Decimal128 types for new builtin functions.
4170 (builtin_function_type): Set the unsigned flags appropriately for
4171 the new builtin functions.
4172 (rs6000_opt_masks): Add support for decimal floating point builtin
4175 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4176 floating point builtin functions.
4177 (RS6000_BTM_COMMON): Likewise.
4178 (RS6000_BTI_long_double): Likewise.
4179 (RS6000_BTI_dfloat64): Likewise.
4180 (RS6000_BTI_dfloat128): Likewise.
4181 (long_double_type_internal_node): Likewise.
4182 (dfloat64_type_internal_node): Likewise.
4183 (dfloat128_type_internal_node): Likewise.
4185 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4186 2.07 bcd arithmetic instructions.
4187 (UNSPEC_BCDSUB): Likewise.
4188 (UNSPEC_BCD_OVERFLOW): Likewise.
4189 (UNSPEC_BCD_ADD_SUB): Likewise.
4190 (bcd_add_sub): Likewise.
4191 (BCD_TEST): Likewise.
4192 (bcd<bcd_add_sub>): Likewise.
4193 (bcd<bcd_add_sub>_test): Likewise.
4194 (bcd<bcd_add_sub>_test2): Likewise.
4195 (bcd<bcd_add_sub>_<code>): Likewise.
4196 (peephole2 for combined bcd ops): Likewise.
4198 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4199 decimal floating point builtin functions.
4200 (UNSPEC_DENBCD): Likewise.
4201 (UNSPEC_DXEX): Likewise.
4202 (UNSPEC_DIEX): Likewise.
4203 (UNSPEC_DSCLI): Likewise.
4204 (UNSPEC_DSCRI): Likewise.
4205 (D64_D128): Likewise.
4206 (dfp_suffix): Likewise.
4207 (dfp_ddedpd_<mode>): Likewise.
4208 (dfp_denbcd_<mode>): Likewise.
4209 (dfp_dxex_<mode>): Likewise.
4210 (dfp_diex_<mode>): Likewise.
4211 (dfp_dscli_<mode>): Likewise.
4212 (dfp_dscri_<mode>): Likewise.
4214 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4216 (UNSPEC_CDTBCD): Likewise.
4217 (UNSPEC_CBCDTD): Likewise.
4218 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4219 (UNSPEC_DIVEO): Likewise.
4220 (UNSPEC_DIVEU): Likewise.
4221 (UNSPEC_DIVEUO): Likewise.
4222 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4223 pack/unpack 128-bit types.
4224 (UNSPEC_PACK_128BIT): Likewise.
4225 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4226 (udiv<mode>3): Use idiv_ldiv mode attribute.
4227 (div<mode>3): Likewise.
4228 (addg6s): Add new BCD builtin functions.
4231 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4232 (div_extend): Likewise.
4233 (div<div_extend>_<mode>"): Likewise.
4234 (FP128_64): Add support for new builtin functions to pack/unpack
4236 (unpack<mode>): Likewise.
4237 (unpacktf_0): Likewise.
4238 (unpacktf_1): Likewise.
4239 (unpack<mode>_dm): Likewise.
4240 (unpack<mode>_nodm): Likewise.
4241 (pack<mode>): Likewise.
4242 (unpackv1ti): Likewise.
4243 (packv1ti): Likewise.
4245 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4247 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4250 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4252 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4253 * gimplify.c (omp_is_private): Change last argument's type to int.
4254 Only diagnose lastprivate if the simd argument is 1, only diagnose
4255 linear if the simd argument is 2.
4256 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4257 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4258 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4259 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4260 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4261 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4262 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4263 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4264 * tree-nested.c (convert_nonlocal_omp_clauses,
4265 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4267 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4270 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4273 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4275 * flag-types.h (enum ivar_visibility): Add.
4277 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4279 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4280 function * argument.
4282 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4284 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4286 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4287 Tom de Vries <tom@codesourcery.com>
4289 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4290 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4292 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4293 * emit-rtl.c (try_split): Same.
4295 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4296 Tom de Vries <tom@codesourcery.com>
4298 * common.opt (fuse-caller-save): New option.
4300 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4302 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4303 elements for big-endian.
4305 2014-04-24 Richard Biener <rguenther@suse.de>
4307 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4308 during TER and instead use the sepops interface for expanding
4309 non-GIMPLE_SINGLE_RHS.
4311 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4313 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4314 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4316 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4318 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4319 assembler 64-bit option.
4320 * configure: Regenerate.
4322 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4324 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4325 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4326 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4327 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4328 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4330 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4332 * config/aarch64/aarch64-builtins.c
4333 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4334 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4335 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4336 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4338 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4339 (Vrevsuff): New mode attribute.
4341 2014-04-24 Terry Guo <terry.guo@arm.com>
4343 * config/arm/arm.h (machine_function): Define variable
4344 after_arm_reorg here.
4345 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4346 (arm_split_constant): Update the way to access variable
4349 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4351 2014-04-23 Tom de Vries <tom@codesourcery.com>
4353 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4355 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4357 * is-a.h: Update comments to reflect the following changes to the
4358 "pointerness" of the API, making the template parameter match the
4359 return type, allowing use of is-a.h with typedefs of pointers.
4360 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4361 that the return type matches the parameter to the is_a_helper.
4363 (dyn_cast): Likewise.
4365 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4366 pointer from the is-a.h API.
4368 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4369 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4371 (is_a_helper <varpool_node>::test): Likewise, convert to...
4372 (is_a_helper <varpool_node *>::test): ...this.
4374 (varpool_first_variable): Update for removal of implicit pointer
4375 from the is-a.h API.
4376 (varpool_next_variable): Likewise.
4377 (varpool_first_static_initializer): Likewise.
4378 (varpool_next_static_initializer): Likewise.
4379 (varpool_first_defined_variable): Likewise.
4380 (varpool_next_defined_variable): Likewise.
4381 (cgraph_first_defined_function): Likewise.
4382 (cgraph_next_defined_function): Likewise.
4383 (cgraph_first_function): Likewise.
4384 (cgraph_next_function): Likewise.
4385 (cgraph_first_function_with_gimple_body): Likewise.
4386 (cgraph_next_function_with_gimple_body): Likewise.
4387 (cgraph_alias_target): Likewise.
4388 (varpool_alias_target): Likewise.
4389 (cgraph_function_or_thunk_node): Likewise.
4390 (varpool_variable_node): Likewise.
4391 (symtab_real_symbol_p): Likewise.
4392 * cgraphunit.c (referred_to_p): Likewise.
4393 (analyze_functions): Likewise.
4394 (handle_alias_pairs): Likewise.
4395 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4396 * gimple-ssa.h (gimple_vuse_op): Likewise.
4397 (gimple_vdef_op): Likewise.
4398 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4399 * gimple.c (gimple_build_asm_1): Likewise.
4400 (gimple_build_try): Likewise.
4401 (gimple_build_resx): Likewise.
4402 (gimple_build_eh_dispatch): Likewise.
4403 (gimple_build_omp_for): Likewise.
4404 (gimple_omp_for_set_clauses): Likewise.
4406 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4407 (is_a_helper <gimple_statement_asm *>::test): ...this.
4408 (is_a_helper <gimple_statement_bind>::test): Convert to...
4409 (is_a_helper <gimple_statement_bind *>::test): ...this.
4410 (is_a_helper <gimple_statement_call>::test): Convert to...
4411 (is_a_helper <gimple_statement_call *>::test): ...this.
4412 (is_a_helper <gimple_statement_catch>::test): Convert to...
4413 (is_a_helper <gimple_statement_catch *>::test): ...this.
4414 (is_a_helper <gimple_statement_resx>::test): Convert to...
4415 (is_a_helper <gimple_statement_resx *>::test): ...this.
4416 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4417 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4418 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4419 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4420 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4421 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4422 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4423 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4424 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4425 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4426 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4427 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4428 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4429 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4430 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4431 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4432 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4433 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4434 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4435 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4436 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4437 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4438 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4439 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4440 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4441 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4442 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4443 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4444 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4445 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4446 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4447 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4448 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4449 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4450 (is_a_helper <gimple_statement_phi>::test): Convert to...
4451 (is_a_helper <gimple_statement_phi *>::test): ...this.
4452 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4453 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4454 (is_a_helper <gimple_statement_try>::test): Convert to...
4455 (is_a_helper <gimple_statement_try *>::test): ...this.
4456 (is_a_helper <gimple_statement_wce>::test): Convert to...
4457 (is_a_helper <gimple_statement_wce *>::test): ...this.
4458 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4459 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4460 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4461 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4462 (is_a_helper <const gimple_statement_call>::test): Convert to...
4463 (is_a_helper <const gimple_statement_call *>::test): ...this.
4464 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4465 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4466 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4467 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4468 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4469 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4470 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4471 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4472 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4474 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4476 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4478 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4480 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4481 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4482 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4484 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4485 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4487 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4488 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4489 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4490 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4491 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4492 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4494 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4495 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4496 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4497 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4499 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4500 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4501 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4502 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4503 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4504 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4505 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4506 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4507 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4508 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4509 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4510 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4511 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4512 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4513 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4514 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4516 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4518 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4519 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4521 (gimple_use_ops): Update for removal of implicit pointer from the
4523 (gimple_set_use_ops): Likewise.
4524 (gimple_vuse): Likewise.
4525 (gimple_vdef): Likewise.
4526 (gimple_vuse_ptr): Likewise.
4527 (gimple_vdef_ptr): Likewise.
4528 (gimple_set_vuse): Likewise.
4529 (gimple_set_vdef): Likewise.
4530 (gimple_omp_return_set_lhs): Likewise.
4531 (gimple_omp_return_lhs): Likewise.
4532 (gimple_omp_return_lhs_ptr): Likewise.
4533 (gimple_call_fntype): Likewise.
4534 (gimple_call_set_fntype): Likewise.
4535 (gimple_call_set_internal_fn): Likewise.
4536 (gimple_call_use_set): Likewise.
4537 (gimple_call_clobber_set): Likewise.
4538 (gimple_bind_vars): Likewise.
4539 (gimple_bind_set_vars): Likewise.
4540 (gimple_bind_body_ptr): Likewise.
4541 (gimple_bind_set_body): Likewise.
4542 (gimple_bind_add_stmt): Likewise.
4543 (gimple_bind_block): Likewise.
4544 (gimple_bind_set_block): Likewise.
4545 (gimple_asm_ninputs): Likewise.
4546 (gimple_asm_noutputs): Likewise.
4547 (gimple_asm_nclobbers): Likewise.
4548 (gimple_asm_nlabels): Likewise.
4549 (gimple_asm_input_op): Likewise.
4550 (gimple_asm_input_op_ptr): Likewise.
4551 (gimple_asm_output_op): Likewise.
4552 (gimple_asm_output_op_ptr): Likewise.
4553 (gimple_asm_set_output_op): Likewise.
4554 (gimple_asm_clobber_op): Likewise.
4555 (gimple_asm_set_clobber_op): Likewise.
4556 (gimple_asm_label_op): Likewise.
4557 (gimple_asm_set_label_op): Likewise.
4558 (gimple_asm_string): Likewise.
4559 (gimple_catch_types): Likewise.
4560 (gimple_catch_types_ptr): Likewise.
4561 (gimple_catch_handler_ptr): Likewise.
4562 (gimple_catch_set_types): Likewise.
4563 (gimple_catch_set_handler): Likewise.
4564 (gimple_eh_filter_types): Likewise.
4565 (gimple_eh_filter_types_ptr): Likewise.
4566 (gimple_eh_filter_failure_ptr): Likewise.
4567 (gimple_eh_filter_set_types): Likewise.
4568 (gimple_eh_filter_set_failure): Likewise.
4569 (gimple_eh_must_not_throw_fndecl): Likewise.
4570 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4571 (gimple_eh_else_n_body_ptr): Likewise.
4572 (gimple_eh_else_e_body_ptr): Likewise.
4573 (gimple_eh_else_set_n_body): Likewise.
4574 (gimple_eh_else_set_e_body): Likewise.
4575 (gimple_try_eval_ptr): Likewise.
4576 (gimple_try_cleanup_ptr): Likewise.
4577 (gimple_try_set_eval): Likewise.
4578 (gimple_try_set_cleanup): Likewise.
4579 (gimple_wce_cleanup_ptr): Likewise.
4580 (gimple_wce_set_cleanup): Likewise.
4581 (gimple_phi_capacity): Likewise.
4582 (gimple_phi_num_args): Likewise.
4583 (gimple_phi_result): Likewise.
4584 (gimple_phi_result_ptr): Likewise.
4585 (gimple_phi_set_result): Likewise.
4586 (gimple_phi_arg): Likewise.
4587 (gimple_phi_set_arg): Likewise.
4588 (gimple_resx_region): Likewise.
4589 (gimple_resx_set_region): Likewise.
4590 (gimple_eh_dispatch_region): Likewise.
4591 (gimple_eh_dispatch_set_region): Likewise.
4592 (gimple_omp_critical_name): Likewise.
4593 (gimple_omp_critical_name_ptr): Likewise.
4594 (gimple_omp_critical_set_name): Likewise.
4595 (gimple_omp_for_clauses): Likewise.
4596 (gimple_omp_for_clauses_ptr): Likewise.
4597 (gimple_omp_for_set_clauses): Likewise.
4598 (gimple_omp_for_collapse): Likewise.
4599 (gimple_omp_for_index): Likewise.
4600 (gimple_omp_for_index_ptr): Likewise.
4601 (gimple_omp_for_set_index): Likewise.
4602 (gimple_omp_for_initial): Likewise.
4603 (gimple_omp_for_initial_ptr): Likewise.
4604 (gimple_omp_for_set_initial): Likewise.
4605 (gimple_omp_for_final): Likewise.
4606 (gimple_omp_for_final_ptr): Likewise.
4607 (gimple_omp_for_set_final): Likewise.
4608 (gimple_omp_for_incr): Likewise.
4609 (gimple_omp_for_incr_ptr): Likewise.
4610 (gimple_omp_for_set_incr): Likewise.
4611 (gimple_omp_for_pre_body_ptr): Likewise.
4612 (gimple_omp_for_set_pre_body): Likewise.
4613 (gimple_omp_parallel_clauses): Likewise.
4614 (gimple_omp_parallel_clauses_ptr): Likewise.
4615 (gimple_omp_parallel_set_clauses): Likewise.
4616 (gimple_omp_parallel_child_fn): Likewise.
4617 (gimple_omp_parallel_child_fn_ptr): Likewise.
4618 (gimple_omp_parallel_set_child_fn): Likewise.
4619 (gimple_omp_parallel_data_arg): Likewise.
4620 (gimple_omp_parallel_data_arg_ptr): Likewise.
4621 (gimple_omp_parallel_set_data_arg): Likewise.
4622 (gimple_omp_task_clauses): Likewise.
4623 (gimple_omp_task_clauses_ptr): Likewise.
4624 (gimple_omp_task_set_clauses): Likewise.
4625 (gimple_omp_task_child_fn): Likewise.
4626 (gimple_omp_task_child_fn_ptr): Likewise.
4627 (gimple_omp_task_set_child_fn): Likewise.
4628 (gimple_omp_task_data_arg): Likewise.
4629 (gimple_omp_task_data_arg_ptr): Likewise.
4630 (gimple_omp_task_set_data_arg): Likewise.
4631 (gimple_omp_taskreg_clauses): Likewise.
4632 (gimple_omp_taskreg_clauses_ptr): Likewise.
4633 (gimple_omp_taskreg_set_clauses): Likewise.
4634 (gimple_omp_taskreg_child_fn): Likewise.
4635 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4636 (gimple_omp_taskreg_set_child_fn): Likewise.
4637 (gimple_omp_taskreg_data_arg): Likewise.
4638 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4639 (gimple_omp_taskreg_set_data_arg): Likewise.
4640 (gimple_omp_task_copy_fn): Likewise.
4641 (gimple_omp_task_copy_fn_ptr): Likewise.
4642 (gimple_omp_task_set_copy_fn): Likewise.
4643 (gimple_omp_task_arg_size): Likewise.
4644 (gimple_omp_task_arg_size_ptr): Likewise.
4645 (gimple_omp_task_set_arg_size): Likewise.
4646 (gimple_omp_task_arg_align): Likewise.
4647 (gimple_omp_task_arg_align_ptr): Likewise.
4648 (gimple_omp_task_set_arg_align): Likewise.
4649 (gimple_omp_single_clauses): Likewise.
4650 (gimple_omp_single_clauses_ptr): Likewise.
4651 (gimple_omp_single_set_clauses): Likewise.
4652 (gimple_omp_target_clauses): Likewise.
4653 (gimple_omp_target_clauses_ptr): Likewise.
4654 (gimple_omp_target_set_clauses): Likewise.
4655 (gimple_omp_target_child_fn): Likewise.
4656 (gimple_omp_target_child_fn_ptr): Likewise.
4657 (gimple_omp_target_set_child_fn): Likewise.
4658 (gimple_omp_target_data_arg): Likewise.
4659 (gimple_omp_target_data_arg_ptr): Likewise.
4660 (gimple_omp_target_set_data_arg): Likewise.
4661 (gimple_omp_teams_clauses): Likewise.
4662 (gimple_omp_teams_clauses_ptr): Likewise.
4663 (gimple_omp_teams_set_clauses): Likewise.
4664 (gimple_omp_sections_clauses): Likewise.
4665 (gimple_omp_sections_clauses_ptr): Likewise.
4666 (gimple_omp_sections_set_clauses): Likewise.
4667 (gimple_omp_sections_control): Likewise.
4668 (gimple_omp_sections_control_ptr): Likewise.
4669 (gimple_omp_sections_set_control): Likewise.
4670 (gimple_omp_for_set_cond): Likewise.
4671 (gimple_omp_for_cond): Likewise.
4672 (gimple_omp_atomic_store_set_val): Likewise.
4673 (gimple_omp_atomic_store_val): Likewise.
4674 (gimple_omp_atomic_store_val_ptr): Likewise.
4675 (gimple_omp_atomic_load_set_lhs): Likewise.
4676 (gimple_omp_atomic_load_lhs): Likewise.
4677 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4678 (gimple_omp_atomic_load_set_rhs): Likewise.
4679 (gimple_omp_atomic_load_rhs): Likewise.
4680 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4681 (gimple_omp_continue_control_def): Likewise.
4682 (gimple_omp_continue_control_def_ptr): Likewise.
4683 (gimple_omp_continue_set_control_def): Likewise.
4684 (gimple_omp_continue_control_use): Likewise.
4685 (gimple_omp_continue_control_use_ptr): Likewise.
4686 (gimple_omp_continue_set_control_use): Likewise.
4687 (gimple_transaction_body_ptr): Likewise.
4688 (gimple_transaction_label): Likewise.
4689 (gimple_transaction_label_ptr): Likewise.
4690 (gimple_transaction_set_body): Likewise.
4691 (gimple_transaction_set_label): Likewise.
4693 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4694 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4695 * ipa-ref.c (ipa_record_reference): Likewise.
4696 * ipa-reference.c (analyze_function): Likewise.
4697 (ipa_reference_write_optimization_summary): Likewise.
4698 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4699 (address_taken_from_non_vtable_p): Likewise.
4700 (comdat_can_be_unshared_p_1): Likewise.
4701 * lto-cgraph.c (lto_output_ref): Likewise.
4702 (add_references): Likewise.
4703 (compute_ltrans_boundary): Likewise.
4704 (output_symtab): Likewise.
4705 (input_ref): Likewise.
4706 (input_cgraph_1): Likewise.
4707 (output_cgraph_opt_summary): Likewise.
4708 * lto-streamer-out.c (lto_output): Likewise.
4709 (output_symbol_p): Likewise.
4710 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4711 (lsei_start_function_in_partition): Likewise.
4712 (lsei_next_variable_in_partition): Likewise.
4713 (lsei_start_variable_in_partition): Likewise.
4714 * symtab.c (insert_to_assembler_name_hash): Likewise.
4715 (unlink_from_assembler_name_hash): Likewise.
4716 (symtab_unregister_node): Likewise.
4717 (symtab_remove_node): Likewise.
4718 (dump_symtab_node): Likewise.
4719 (verify_symtab_base): Likewise.
4720 (verify_symtab_node): Likewise.
4721 (symtab_make_decl_local): Likewise.
4722 (symtab_alias_ultimate_target): Likewise.
4723 (symtab_resolve_alias): Likewise.
4724 (symtab_get_symbol_partitioning_class): Likewise.
4725 * tree-phinodes.c (allocate_phi_node): Likewise.
4726 (reserve_phi_args_for_new_edge): Likewise.
4727 (remove_phi_args): Likewise.
4728 * varpool.c (varpool_node_for_asm): Likewise.
4729 (varpool_remove_unreferenced_decls): Likewise.
4731 2014-04-23 Jeff Law <law@redhat.com>
4733 PR tree-optimization/60902
4734 * tree-ssa-threadedge.c
4735 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4736 invalidate outputs from statements that do not produce useful
4737 outputs for threading.
4739 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4741 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4742 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4743 machine descriptions for Stack Smashing Protector.
4745 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4747 * aarch64.md (<optab>_rol<mode>3): New pattern.
4748 (<optab>_rolsi3_uxtw): Likewise.
4749 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4751 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4753 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4754 (arm_cortex_a12_tune): Likewise.
4756 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4758 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4760 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4762 * config/arm/arm.md (arm_rev16si2): New pattern.
4763 (arm_rev16si2_alt): Likewise.
4764 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4766 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4768 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4769 (rev16<mode>2_alt): Likewise.
4770 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4771 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4772 (aarch_rev16_shleft_mask_imm_p): Likewise.
4773 (aarch_rev16_p_1): Likewise.
4774 (aarch_rev16_p): Likewise.
4775 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4776 (aarch_rev16_shright_mask_imm_p): Likewise.
4777 (aarch_rev16_shleft_mask_imm_p): Likewise.
4779 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4781 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4782 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4784 (cortex_a53_extra_costs): Likewise.
4785 (cortex_a57_extra_costs): Likewise.
4786 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4787 (cortexa7_extra_costs): Likewise.
4788 (cortexa8_extra_costs): Likewise.
4789 (cortexa12_extra_costs): Likewise.
4790 (cortexa15_extra_costs): Likewise.
4791 (v7m_extra_costs): Likewise.
4792 (arm_new_rtx_costs): Handle BSWAP.
4794 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4796 * config/arm/arm.c (cortexa8_extra_costs): New table.
4797 (arm_cortex_a8_tune): New tuning struct.
4798 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4800 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4802 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4804 2014-04-23 Richard Biener <rguenther@suse.de>
4806 * Makefile.in (OBJS): Remove loop-unswitch.o.
4807 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4808 * passes.def (pass_rtl_unswitch): Likewise.
4809 * loop-init.c (gate_rtl_unswitch): Likewise.
4810 (rtl_unswitch): Likewise.
4811 (pass_data_rtl_unswitch): Likewise.
4812 (pass_rtl_unswitch): Likewise.
4813 (make_pass_rtl_unswitch): Likewise.
4814 * rtl.h (reversed_condition): Likewise.
4815 (compare_and_jump_seq): Likewise.
4816 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4818 * loop-unroll.c (compare_and_jump_seq): Likewise.
4820 2014-04-23 Richard Biener <rguenther@suse.de>
4822 PR tree-optimization/60903
4823 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4824 commented code block.
4825 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4826 loop flags to newly created BBs and edges.
4828 2014-04-23 Nick Clifton <nickc@redhat.com>
4830 * config/msp430/msp430.c (msp430_handle_option): Move function
4832 (msp430_option_override): Simplify mcu and mcpu option handling.
4833 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4834 support for -mhwmult command line option.
4835 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4836 -mhwmult command line option.
4837 (msp430_hwmult_enabled): Delete.
4838 (msp43o_output_labelref): Add support for -mhwmult command line option.
4839 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4840 (umulsidi3): Likewise.
4841 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4842 (mcpu, mlarge, msmall): Likewise.
4843 (mhwmult): New option.
4844 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4846 (msp430_is_f5_mcu): Remove prototype.
4847 (msp430_use_f5_series_hwmult): Add prototype.
4848 * config/msp430/msp430-opts.h: New file.
4849 * common/config/msp430: New directory.
4850 * common/config/msp430/msp430-common.c: New file.
4851 * config.gcc (msp430): Remove target_has_targetm_common.
4852 * doc/invoke.texi: Document -mhwmult command line option.
4854 2014-04-23 Nick Clifton <nickc@redhat.com>
4856 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4857 default-manifest.o if it can be found in the search path.
4858 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4860 2014-04-23 Terry Guo <terry.guo@arm.com>
4862 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4864 2014-04-23 Richard Biener <rguenther@suse.de>
4867 * tree-inline.c (declare_return_variable): Use mark_addressable.
4869 2014-04-23 Richard Biener <rguenther@suse.de>
4872 * loop-init.c (loop_optimizer_init): Make sure to apply
4873 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4875 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4878 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4880 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4881 if flag_sanitize_undefined_trap_on_error.
4882 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4883 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4884 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4885 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4886 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4887 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4888 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4889 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4890 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4891 * ubsan.c (ubsan_instrument_unreachable): Return
4892 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4893 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4894 if flag_sanitize_undefined_trap_on_error and
4895 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4896 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4897 instrument_bool_enum_load): Emit __builtin_trap () if
4898 flag_sanitize_undefined_trap_on_error and
4899 __builtin_handle_*_abort () if !flag_sanitize_recover.
4900 * doc/invoke.texi (-fsanitize-recover,
4901 -fsanitize-undefined-trap-on-error): Document.
4903 2014-04-22 Christian Bruel <christian.bruel@st.com>
4905 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4906 Force immediates to SImode.
4908 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4910 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4912 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4913 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4914 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4915 (nios2_fpu_insn): Add entry for round.
4916 (N2FPU_NO_ERRNO_P): Define.
4917 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4919 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4921 2014-04-22 Richard Henderson <rth@redhat.com>
4923 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4924 (add<GPI>3_compare0): Remove leading * from name.
4925 (add<GPI>3_carryin): Likewise.
4926 (sub<GPI>3_compare0): Likewise.
4927 (sub<GPI>3_carryin): Likewise.
4928 (<su_optab>mulditi3): New expander.
4929 (multi3): New expander.
4930 (madd<GPI>): Remove leading * from name.
4932 2014-04-22 Martin Jambor <mjambor@suse.cz>
4934 * cgraphclones.c (cgraph_function_versioning): Copy
4935 ipa_transforms_to_apply instead of asserting it is empty.
4937 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4940 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4941 on count_exp to get mode.
4943 2014-04-22 Andrew Pinski <apinski@cavium.com>
4945 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4946 Handle TLS for ILP32.
4947 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4948 (tlsie_small_<mode>): this and handle PTR.
4949 (tlsie_small_sidi): New pattern.
4950 (tlsle_small): Change to an expand to handle ILP32.
4951 (tlsle_small_<mode>): New pattern.
4952 (tlsdesc_small): Rename to ...
4953 (tlsdesc_small_<mode>): this and handle PTR.
4955 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4957 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4959 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4961 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4962 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4963 (aarch64_types_signed_poly_qualifiers): Likewise.
4964 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4965 (aarch64_types_poly_signed_qualifiers): Likewise.
4966 (TYPES_REINTERP_SS): Type macro added.
4967 (TYPES_REINTERP_SU): Likewise.
4968 (TYPES_REINTERP_SP): Likewise.
4969 (TYPES_REINTERP_US): Likewise.
4970 (TYPES_REINTERP_PS): Likewise.
4971 (aarch64_fold_builtin): New expression folding added.
4972 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4973 Declarations removed.
4974 (REINTERP_SS): Declarations added.
4975 (REINTERP_US): Likewise.
4976 (REINTERP_PS): Likewise.
4977 (REINTERP_SU): Likewise.
4978 (REINTERP_SP): Likewise.
4979 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4980 (vreinterpretq_p8_f64): Likewise.
4981 (vreinterpret_p16_f64): Likewise.
4982 (vreinterpretq_p16_f64): Likewise.
4983 (vreinterpret_f32_f64): Likewise.
4984 (vreinterpretq_f32_f64): Likewise.
4985 (vreinterpret_f64_f32): Likewise.
4986 (vreinterpret_f64_p8): Likewise.
4987 (vreinterpret_f64_p16): Likewise.
4988 (vreinterpret_f64_s8): Likewise.
4989 (vreinterpret_f64_s16): Likewise.
4990 (vreinterpret_f64_s32): Likewise.
4991 (vreinterpret_f64_s64): Likewise.
4992 (vreinterpret_f64_u8): Likewise.
4993 (vreinterpret_f64_u16): Likewise.
4994 (vreinterpret_f64_u32): Likewise.
4995 (vreinterpret_f64_u64): Likewise.
4996 (vreinterpretq_f64_f32): Likewise.
4997 (vreinterpretq_f64_p8): Likewise.
4998 (vreinterpretq_f64_p16): Likewise.
4999 (vreinterpretq_f64_s8): Likewise.
5000 (vreinterpretq_f64_s16): Likewise.
5001 (vreinterpretq_f64_s32): Likewise.
5002 (vreinterpretq_f64_s64): Likewise.
5003 (vreinterpretq_f64_u8): Likewise.
5004 (vreinterpretq_f64_u16): Likewise.
5005 (vreinterpretq_f64_u32): Likewise.
5006 (vreinterpretq_f64_u64): Likewise.
5007 (vreinterpret_s64_f64): Likewise.
5008 (vreinterpretq_s64_f64): Likewise.
5009 (vreinterpret_u64_f64): Likewise.
5010 (vreinterpretq_u64_f64): Likewise.
5011 (vreinterpret_s8_f64): Likewise.
5012 (vreinterpretq_s8_f64): Likewise.
5013 (vreinterpret_s16_f64): Likewise.
5014 (vreinterpretq_s16_f64): Likewise.
5015 (vreinterpret_s32_f64): Likewise.
5016 (vreinterpretq_s32_f64): Likewise.
5017 (vreinterpret_u8_f64): Likewise.
5018 (vreinterpretq_u8_f64): Likewise.
5019 (vreinterpret_u16_f64): Likewise.
5020 (vreinterpretq_u16_f64): Likewise.
5021 (vreinterpret_u32_f64): Likewise.
5022 (vreinterpretq_u32_f64): Likewise.
5024 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5026 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5027 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5028 (vreinterpret_p8_s8): Likewise.
5029 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5030 (vreinterpret_p8_s16): Likewise.
5031 (vreinterpret_p8_s32): Likewise.
5032 (vreinterpret_p8_s64): Likewise.
5033 (vreinterpret_p8_f32): Likewise.
5034 (vreinterpret_p8_u8): Likewise.
5035 (vreinterpret_p8_u16): Likewise.
5036 (vreinterpret_p8_u32): Likewise.
5037 (vreinterpret_p8_u64): Likewise.
5038 (vreinterpret_p8_p16): Likewise.
5039 (vreinterpretq_p8_s8): Likewise.
5040 (vreinterpretq_p8_s16): Likewise.
5041 (vreinterpretq_p8_s32): Likewise.
5042 (vreinterpretq_p8_s64): Likewise.
5043 (vreinterpretq_p8_f32): Likewise.
5044 (vreinterpretq_p8_u8): Likewise.
5045 (vreinterpretq_p8_u16): Likewise.
5046 (vreinterpretq_p8_u32): Likewise.
5047 (vreinterpretq_p8_u64): Likewise.
5048 (vreinterpretq_p8_p16): Likewise.
5049 (vreinterpret_p16_s8): Likewise.
5050 (vreinterpret_p16_s16): Likewise.
5051 (vreinterpret_p16_s32): Likewise.
5052 (vreinterpret_p16_s64): Likewise.
5053 (vreinterpret_p16_f32): Likewise.
5054 (vreinterpret_p16_u8): Likewise.
5055 (vreinterpret_p16_u16): Likewise.
5056 (vreinterpret_p16_u32): Likewise.
5057 (vreinterpret_p16_u64): Likewise.
5058 (vreinterpret_p16_p8): Likewise.
5059 (vreinterpretq_p16_s8): Likewise.
5060 (vreinterpretq_p16_s16): Likewise.
5061 (vreinterpretq_p16_s32): Likewise.
5062 (vreinterpretq_p16_s64): Likewise.
5063 (vreinterpretq_p16_f32): Likewise.
5064 (vreinterpretq_p16_u8): Likewise.
5065 (vreinterpretq_p16_u16): Likewise.
5066 (vreinterpretq_p16_u32): Likewise.
5067 (vreinterpretq_p16_u64): Likewise.
5068 (vreinterpretq_p16_p8): Likewise.
5069 (vreinterpret_f32_s8): Likewise.
5070 (vreinterpret_f32_s16): Likewise.
5071 (vreinterpret_f32_s32): Likewise.
5072 (vreinterpret_f32_s64): Likewise.
5073 (vreinterpret_f32_u8): Likewise.
5074 (vreinterpret_f32_u16): Likewise.
5075 (vreinterpret_f32_u32): Likewise.
5076 (vreinterpret_f32_u64): Likewise.
5077 (vreinterpret_f32_p8): Likewise.
5078 (vreinterpret_f32_p16): Likewise.
5079 (vreinterpretq_f32_s8): Likewise.
5080 (vreinterpretq_f32_s16): Likewise.
5081 (vreinterpretq_f32_s32): Likewise.
5082 (vreinterpretq_f32_s64): Likewise.
5083 (vreinterpretq_f32_u8): Likewise.
5084 (vreinterpretq_f32_u16): Likewise.
5085 (vreinterpretq_f32_u32): Likewise.
5086 (vreinterpretq_f32_u64): Likewise.
5087 (vreinterpretq_f32_p8): Likewise.
5088 (vreinterpretq_f32_p16): Likewise.
5089 (vreinterpret_s64_s8): Likewise.
5090 (vreinterpret_s64_s16): Likewise.
5091 (vreinterpret_s64_s32): Likewise.
5092 (vreinterpret_s64_f32): Likewise.
5093 (vreinterpret_s64_u8): Likewise.
5094 (vreinterpret_s64_u16): Likewise.
5095 (vreinterpret_s64_u32): Likewise.
5096 (vreinterpret_s64_u64): Likewise.
5097 (vreinterpret_s64_p8): Likewise.
5098 (vreinterpret_s64_p16): Likewise.
5099 (vreinterpretq_s64_s8): Likewise.
5100 (vreinterpretq_s64_s16): Likewise.
5101 (vreinterpretq_s64_s32): Likewise.
5102 (vreinterpretq_s64_f32): Likewise.
5103 (vreinterpretq_s64_u8): Likewise.
5104 (vreinterpretq_s64_u16): Likewise.
5105 (vreinterpretq_s64_u32): Likewise.
5106 (vreinterpretq_s64_u64): Likewise.
5107 (vreinterpretq_s64_p8): Likewise.
5108 (vreinterpretq_s64_p16): Likewise.
5109 (vreinterpret_u64_s8): Likewise.
5110 (vreinterpret_u64_s16): Likewise.
5111 (vreinterpret_u64_s32): Likewise.
5112 (vreinterpret_u64_s64): Likewise.
5113 (vreinterpret_u64_f32): Likewise.
5114 (vreinterpret_u64_u8): Likewise.
5115 (vreinterpret_u64_u16): Likewise.
5116 (vreinterpret_u64_u32): Likewise.
5117 (vreinterpret_u64_p8): Likewise.
5118 (vreinterpret_u64_p16): Likewise.
5119 (vreinterpretq_u64_s8): Likewise.
5120 (vreinterpretq_u64_s16): Likewise.
5121 (vreinterpretq_u64_s32): Likewise.
5122 (vreinterpretq_u64_s64): Likewise.
5123 (vreinterpretq_u64_f32): Likewise.
5124 (vreinterpretq_u64_u8): Likewise.
5125 (vreinterpretq_u64_u16): Likewise.
5126 (vreinterpretq_u64_u32): Likewise.
5127 (vreinterpretq_u64_p8): Likewise.
5128 (vreinterpretq_u64_p16): Likewise.
5129 (vreinterpret_s8_s16): Likewise.
5130 (vreinterpret_s8_s32): Likewise.
5131 (vreinterpret_s8_s64): Likewise.
5132 (vreinterpret_s8_f32): Likewise.
5133 (vreinterpret_s8_u8): Likewise.
5134 (vreinterpret_s8_u16): Likewise.
5135 (vreinterpret_s8_u32): Likewise.
5136 (vreinterpret_s8_u64): Likewise.
5137 (vreinterpret_s8_p8): Likewise.
5138 (vreinterpret_s8_p16): Likewise.
5139 (vreinterpretq_s8_s16): Likewise.
5140 (vreinterpretq_s8_s32): Likewise.
5141 (vreinterpretq_s8_s64): Likewise.
5142 (vreinterpretq_s8_f32): Likewise.
5143 (vreinterpretq_s8_u8): Likewise.
5144 (vreinterpretq_s8_u16): Likewise.
5145 (vreinterpretq_s8_u32): Likewise.
5146 (vreinterpretq_s8_u64): Likewise.
5147 (vreinterpretq_s8_p8): Likewise.
5148 (vreinterpretq_s8_p16): Likewise.
5149 (vreinterpret_s16_s8): Likewise.
5150 (vreinterpret_s16_s32): Likewise.
5151 (vreinterpret_s16_s64): Likewise.
5152 (vreinterpret_s16_f32): Likewise.
5153 (vreinterpret_s16_u8): Likewise.
5154 (vreinterpret_s16_u16): Likewise.
5155 (vreinterpret_s16_u32): Likewise.
5156 (vreinterpret_s16_u64): Likewise.
5157 (vreinterpret_s16_p8): Likewise.
5158 (vreinterpret_s16_p16): Likewise.
5159 (vreinterpretq_s16_s8): Likewise.
5160 (vreinterpretq_s16_s32): Likewise.
5161 (vreinterpretq_s16_s64): Likewise.
5162 (vreinterpretq_s16_f32): Likewise.
5163 (vreinterpretq_s16_u8): Likewise.
5164 (vreinterpretq_s16_u16): Likewise.
5165 (vreinterpretq_s16_u32): Likewise.
5166 (vreinterpretq_s16_u64): Likewise.
5167 (vreinterpretq_s16_p8): Likewise.
5168 (vreinterpretq_s16_p16): Likewise.
5169 (vreinterpret_s32_s8): Likewise.
5170 (vreinterpret_s32_s16): Likewise.
5171 (vreinterpret_s32_s64): Likewise.
5172 (vreinterpret_s32_f32): Likewise.
5173 (vreinterpret_s32_u8): Likewise.
5174 (vreinterpret_s32_u16): Likewise.
5175 (vreinterpret_s32_u32): Likewise.
5176 (vreinterpret_s32_u64): Likewise.
5177 (vreinterpret_s32_p8): Likewise.
5178 (vreinterpret_s32_p16): Likewise.
5179 (vreinterpretq_s32_s8): Likewise.
5180 (vreinterpretq_s32_s16): Likewise.
5181 (vreinterpretq_s32_s64): Likewise.
5182 (vreinterpretq_s32_f32): Likewise.
5183 (vreinterpretq_s32_u8): Likewise.
5184 (vreinterpretq_s32_u16): Likewise.
5185 (vreinterpretq_s32_u32): Likewise.
5186 (vreinterpretq_s32_u64): Likewise.
5187 (vreinterpretq_s32_p8): Likewise.
5188 (vreinterpretq_s32_p16): Likewise.
5189 (vreinterpret_u8_s8): Likewise.
5190 (vreinterpret_u8_s16): Likewise.
5191 (vreinterpret_u8_s32): Likewise.
5192 (vreinterpret_u8_s64): Likewise.
5193 (vreinterpret_u8_f32): Likewise.
5194 (vreinterpret_u8_u16): Likewise.
5195 (vreinterpret_u8_u32): Likewise.
5196 (vreinterpret_u8_u64): Likewise.
5197 (vreinterpret_u8_p8): Likewise.
5198 (vreinterpret_u8_p16): Likewise.
5199 (vreinterpretq_u8_s8): Likewise.
5200 (vreinterpretq_u8_s16): Likewise.
5201 (vreinterpretq_u8_s32): Likewise.
5202 (vreinterpretq_u8_s64): Likewise.
5203 (vreinterpretq_u8_f32): Likewise.
5204 (vreinterpretq_u8_u16): Likewise.
5205 (vreinterpretq_u8_u32): Likewise.
5206 (vreinterpretq_u8_u64): Likewise.
5207 (vreinterpretq_u8_p8): Likewise.
5208 (vreinterpretq_u8_p16): Likewise.
5209 (vreinterpret_u16_s8): Likewise.
5210 (vreinterpret_u16_s16): Likewise.
5211 (vreinterpret_u16_s32): Likewise.
5212 (vreinterpret_u16_s64): Likewise.
5213 (vreinterpret_u16_f32): Likewise.
5214 (vreinterpret_u16_u8): Likewise.
5215 (vreinterpret_u16_u32): Likewise.
5216 (vreinterpret_u16_u64): Likewise.
5217 (vreinterpret_u16_p8): Likewise.
5218 (vreinterpret_u16_p16): Likewise.
5219 (vreinterpretq_u16_s8): Likewise.
5220 (vreinterpretq_u16_s16): Likewise.
5221 (vreinterpretq_u16_s32): Likewise.
5222 (vreinterpretq_u16_s64): Likewise.
5223 (vreinterpretq_u16_f32): Likewise.
5224 (vreinterpretq_u16_u8): Likewise.
5225 (vreinterpretq_u16_u32): Likewise.
5226 (vreinterpretq_u16_u64): Likewise.
5227 (vreinterpretq_u16_p8): Likewise.
5228 (vreinterpretq_u16_p16): Likewise.
5229 (vreinterpret_u32_s8): Likewise.
5230 (vreinterpret_u32_s16): Likewise.
5231 (vreinterpret_u32_s32): Likewise.
5232 (vreinterpret_u32_s64): Likewise.
5233 (vreinterpret_u32_f32): Likewise.
5234 (vreinterpret_u32_u8): Likewise.
5235 (vreinterpret_u32_u16): Likewise.
5236 (vreinterpret_u32_u64): Likewise.
5237 (vreinterpret_u32_p8): Likewise.
5238 (vreinterpret_u32_p16): Likewise.
5239 (vreinterpretq_u32_s8): Likewise.
5240 (vreinterpretq_u32_s16): Likewise.
5241 (vreinterpretq_u32_s32): Likewise.
5242 (vreinterpretq_u32_s64): Likewise.
5243 (vreinterpretq_u32_f32): Likewise.
5244 (vreinterpretq_u32_u8): Likewise.
5245 (vreinterpretq_u32_u16): Likewise.
5246 (vreinterpretq_u32_u64): Likewise.
5247 (vreinterpretq_u32_p8): Likewise.
5248 (vreinterpretq_u32_p16): Likewise.
5250 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5252 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5254 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5256 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5257 (vqnegd_s64): Likewise.
5258 (vqabs_s64): Likewise.
5259 (vqabsd_s64): Likewise.
5261 2014-04-22 Richard Henderson <rth@redhat.com>
5263 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5264 computation to the top of the loop.
5266 2014-04-22 Renlin <renlin.li@arm.com>
5267 Jiong Wang <jiong.wang@arm.com>
5269 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5270 * config/aarch64/aarch64.c (aarch64_layout_frame)
5271 (aarch64_initial_elimination_offset): Likewise.
5273 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5275 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5278 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5280 * machmode.h (bitwise_mode_for_mode): Declare.
5281 * stor-layout.h (bitwise_type_for_mode): Likewise.
5282 * stor-layout.c (bitwise_mode_for_mode): New function.
5283 (bitwise_type_for_mode): Likewise.
5284 * builtins.c (fold_builtin_memory_op): Use it instead of
5285 int_mode_for_mode and build_nonstandard_integer_type.
5287 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5289 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5290 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5291 (*-*-solaris2*): Simplify.
5292 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5293 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5294 *-*-solaris2.9* handling.
5296 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5298 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5299 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5301 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5302 * configure: Regenerate.
5304 * config/i386/sol2-9.h: Remove.
5306 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5307 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5308 Remove Solaris 9 references.
5310 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5312 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5313 (floatuns<GPI:mode><GPF:mode>2): Remove.
5314 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5315 and floatuns conversions.
5316 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5317 and floatuns conversions.
5318 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5319 (w1,w2): New mode attributes for inequal width conversions.
5321 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5323 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5324 the output asm format.
5326 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5328 * config/aarch64/aarch64-simd.md
5329 (aarch64_cm<optab>di): Always split.
5330 (*aarch64_cm<optab>di): New.
5331 (aarch64_cmtstdi): Always split.
5332 (*aarch64_cmtstdi): New.
5334 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5336 PR tree-optimization/60823
5337 * omp-low.c (ipa_simd_modify_function_body): Go through
5338 all SSA_NAMEs and for those refering to vector arguments
5339 which are going to be replaced adjust SSA_NAME_VAR and,
5340 if it is a default definition, change it into a non-default
5341 definition assigned at the beginning of function from new_decl.
5342 (ipa_simd_modify_stmt_ops): Rewritten.
5343 * tree-dfa.c (set_ssa_default_def): When removing default def,
5344 check for NULL loc instead of NULL *loc.
5346 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5348 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5349 restrictions on core registers for DImode values in Thumb2.
5351 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5353 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5354 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5356 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5358 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5359 (*iordi_notzesidi_di): Likewise.
5360 (*iordi_notsesidi_di): Likewise.
5362 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5364 * config/arm/arm-protos.h (tune_params): New struct members.
5365 * config/arm/arm.c: Initialise tune_params per processor.
5366 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5367 for speed, based on new tune_params.
5369 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5371 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5372 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5373 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5374 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5375 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5376 (vrnda_f64): Likewise.
5377 (vrndi_f64): Likewise.
5378 (vrndm_f64): Likewise.
5379 (vrndn_f64): Likewise.
5380 (vrndp_f64): Likewise.
5381 (vrndx_f64): Likewise.
5383 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5385 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5386 GET_MODE_SIZE argument is enum machine_mode.
5388 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5391 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5392 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5394 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5397 * asan.c (asan_emit_stack_protection): Force the base to align to
5398 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5399 appropriate bits if STRICT_ALIGNMENT.
5400 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5402 (expand_used_vars): Leave a space in the stack frame for alignment
5403 if STRICT_ALIGNMENT.
5405 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5407 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5409 (gimple_store_p): Likewise.
5410 (gimple_assign_load_p): Likewise.
5411 (gimple_assign_cast_p): Likewise.
5412 (gimple_clobber_p): Likewise.
5414 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5415 rather than a gimple.
5416 (gimple_assign_cast_p): Likewise.
5418 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5421 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5422 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5424 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5425 more debug information for E500 if -mdebug=reg.
5427 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5430 * config/i386/i386.c (ix86_expand_builtin)
5431 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5432 register for target RTX.
5433 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5435 2014-04-18 Cong Hou <congh@google.com>
5437 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5438 the widen-mult pattern by handling two operands with different sizes,
5439 and operands whose size is smaller than half of the result type.
5441 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5443 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5444 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5445 (do_estimate_edge_time): Compute it.
5446 * ipa-inline.c (want_inline_small_function_p): Bypass
5447 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5449 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5451 * ipa-inline.c (spec_rem): New static variable.
5452 (dump_overall_stats): New function.
5453 (dump_inline_stats): New function.
5455 2014-04-18 Richard Henderson <rth@redhat.com>
5457 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5458 to GET_MODE_SIZE, not a reg_class_t.
5460 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5462 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5463 (vsx_xxmrglw_<mode>): Likewise.
5465 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5468 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5469 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5470 (rs6000_init_hard_regno_mode_ok): Likewise.
5472 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5474 * ipa-inline.c (inline_small_functions): Account only non-cold
5476 * doc/invoke.texi (inline-unit-growth): Update documentation.
5478 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5480 * config/rs6000/rs6000.md (addti3, subti3): New.
5482 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5485 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5486 comment. Check optimize_insn_for_size_p instead of
5487 optimize_insn_for_speed_p.
5489 2014-04-17 Martin Jambor <mjambor@suse.cz>
5491 * gimple-iterator.c (gsi_start_edge): New function.
5492 * gimple-iterator.h (gsi_start_edge): Declare.
5493 * tree-sra.c (single_non_eh_succ): New function.
5494 (disqualify_ops_if_throwing_stmt): Renamed to
5495 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5496 having one non-EH successor BB.
5497 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5498 generate loads into replacements.
5499 (sra_modify_assign): Likewise and and also use the simple path for
5501 (sra_modify_function_body): Commit statements on edges.
5503 2014-04-17 Richard Biener <rguenther@suse.de>
5506 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5507 comparison results and add clarifying comment.
5509 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5511 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5512 (blank_mode): Initialize it.
5513 (emit_mode_size_inline, emit_mode_nunits_inline,
5514 emit_mode_inner_inline): New functions.
5515 (emit_insn_modes_h): Call them and surround their output with
5516 #if GCC_VERSION >= 4001 ... #endif.
5517 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5518 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5519 mode_* arrays if the argument is __builtin_constant_p.
5520 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5521 is enum machine_mode.
5523 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5525 * passes.c (opt_pass::execute): Adjust.
5526 (pass_manager::execute_pass_mode_switching): Likewise.
5527 (early_local_passes::execute): Likewise.
5528 (execute_one_pass): Pass cfun to the pass's execute method.
5529 * tree-pass.h (opt_pass::execute): Add function * argument.
5530 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5531 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5532 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5533 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5534 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5535 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5536 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5537 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5538 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5539 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5540 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5541 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5542 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5543 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5544 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5545 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5546 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5547 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5548 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5549 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5550 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5551 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5552 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5553 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5554 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5555 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5556 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5557 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5558 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5561 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5563 * passes.c (opt_pass::gate): Take function * argument.
5564 (gate_all_early_local_passes): Merge into
5565 (early_local_passes::gate): this.
5566 (gate_all_early_optimizations): Merge into
5567 (all_early_optimizations::gate): this.
5568 (gate_all_optimizations): Mege into
5569 (all_optimizations::gate): this.
5570 (gate_all_optimizations_g): Merge into
5571 (all_optimizations_g::gate): this.
5572 (gate_rest_of_compilation): Mege into
5573 (rest_of_compilation::gate): this.
5574 (gate_postreload): Merge into
5575 (postreload::gate): this.
5576 (dump_one_pass): Pass cfun to the pass's gate method.
5577 (execute_ipa_summary_passes): Likewise.
5578 (execute_one_pass): Likewise.
5579 (ipa_write_summaries_2): Likewise.
5580 (ipa_write_optimization_summaries_1): Likewise.
5581 (ipa_read_summaries_1): Likewise.
5582 (ipa_read_optimization_summaries_1): Likewise.
5583 (execute_ipa_stmt_fixups): Likewise.
5584 * tree-pass.h (opt_pass::gate): Add function * argument.
5585 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5586 combine-stack-adj.c, combine.c, compare-elim.c,
5587 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5588 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5589 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5590 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5591 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5592 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5593 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5594 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5595 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5596 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5597 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5598 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5599 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5600 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5601 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5602 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5603 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5604 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5605 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5606 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5607 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5608 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5609 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5610 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5611 var-tracking.c, vtable-verify.c, web.c: Adjust.
5613 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5615 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5616 * configure: Regenerate.
5618 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5620 * passes.c (dump_one_pass): don't check pass->has_gate.
5621 (execute_ipa_summary_passes): Likewise.
5622 (execute_one_pass): Likewise.
5623 (ipa_write_summaries_2): Likewise.
5624 (ipa_write_optimization_summaries_1): Likewise.
5625 (ipa_read_optimization_summaries_1): Likewise.
5626 (execute_ipa_stmt_fixups): Likewise.
5627 * tree-pass.h (pass_data::has_gate): Remove.
5628 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5629 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5630 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5631 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5632 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5633 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5634 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5635 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5636 gimple-low.c, gimple-ssa-isolate-paths.c,
5637 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5638 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5639 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5640 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5641 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5642 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5643 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5644 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5645 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5646 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5647 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5648 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5649 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5650 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5651 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5652 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5653 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5654 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5655 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5656 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5657 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5658 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5661 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5663 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5665 * passes.c (pass_manager::register_dump_files_1): Merge into
5666 (pass_manager::register_dump_files): this, and remove its handling of
5667 properties since the pass always has the properties anyway.
5668 (pass_manager::pass_manager): Adjust.
5670 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5672 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5673 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5674 dealing with properties.
5675 (pass_manager::register_dump_files): Adjust.
5677 2014-03-20 Mark Wielaard <mjw@redhat.com>
5679 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5680 then represent the bound as normal constant value.
5682 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5685 Forward port from 4.8 branch
5686 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5688 * config/i386/bmiintrin.h (_blsi_u32): New.
5692 (_blsmsk_u32): Ditto.
5693 (_blsmsk_u64): Ditto.
5694 (_tzcnt_u32): Ditto.
5695 (_tzcnt_u64): Ditto.
5697 2014-04-17 Kito Cheng <kito@0xlab.org>
5699 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5701 2014-04-17 Richard Biener <rguenther@suse.de>
5704 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5705 boolean results for comparisons.
5707 2014-04-17 Richard Biener <rguenther@suse.de>
5709 PR tree-optimization/60836
5710 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5711 initial PHI args to be gimple values.
5713 2014-04-17 Richard Biener <rguenther@suse.de>
5715 PR tree-optimization/60841
5716 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5717 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5718 of stmts to SLP build.
5719 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5720 (vect_analyze_slp): Likewise.
5721 (vect_analyze_slp_instance): Likewise.
5722 (vect_build_slp_tree): Limit overall SLP tree growth.
5723 * tree-vectorizer.h (vect_analyze_data_refs,
5724 vect_analyze_slp): Adjust prototypes.
5726 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5728 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5731 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5733 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5734 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5735 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5736 for TARGET_SLOW_PSHUFB
5738 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5740 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5741 * config/i386/i386.c (intel_cost): Ditto.
5743 2014-04-17 Joey Ye <joey.ye@arm.com>
5745 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5747 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5749 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5752 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5754 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5755 (type_all_derivations_known_p): New predicate.
5756 (type_all_ctors_visible_p): New predicate.
5757 (type_possibly_instantiated_p): New predicate.
5758 (get_odr_type): Compute all_derivations_known.
5759 (dump_odr_type): Dump the flag.
5760 (maybe_record_type): Cleanup.
5761 (record_target_from_binfo): Add bases_to_consider array;
5762 record bases for types w/o instances and skip CXX destructor.
5763 (possible_polymorphic_call_targets_1): Add bases_to_consider
5764 and consider_construction parameters; check if type may have instance.
5765 (get_polymorphic_call_info): Set maybe_in_construction to true
5766 when we know nothing.
5767 (record_targets_from_bases): Skip CXX destructors; they are
5768 never called for types in construction.
5769 (possible_polymorphic_call_targets): Do not record target when
5770 type may not have instance.
5772 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5775 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5776 external aliases alive, too.
5778 2014-04-16 Andrew Pinski <apinski@cavium.com>
5780 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5783 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5785 * final.c (compute_alignments): Do not apply loop alignment to a block
5786 falling through to the exit.
5788 2014-04-16 Catherine Moore <clm@codesourcery.com>
5790 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5791 Adjust constraints for microMIPS store patterns.
5793 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5795 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5797 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5799 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5800 (append_use): Run at -O0.
5801 (append_vdef): Likewise.
5802 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5803 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5805 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5807 PR tree-optimization/60844
5808 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5809 (propagate_op_to_single_use, remove_visited_stmt_chain,
5810 linearize_expr, repropagate_negates, reassociate_bb): Use it
5811 instead of gsi_remove.
5813 2014-04-16 Martin Jambor <mjambor@suse.cz>
5815 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5816 ipa_transforms_to_apply.
5817 (cgraph_function_versioning): Assert that old_node has empty
5818 ipa_transforms_to_apply.
5819 * trans-mem.c (ipa_tm_create_version): Likewise.
5820 * tree-inline.c (tree_function_versioning): Do not duplicate
5821 ipa_transforms_to_apply.
5823 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5826 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5828 Pass necessary as flags on 64-bit Solaris/x86.
5829 Use lowercase relocs for x86_64-*-*.
5830 * configure: Regenerate.
5832 2014-04-15 Jan Hubicka <jh@suse.cz>
5834 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5835 (maybe_record_node, likely_target_p): Use it.
5837 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5840 Revert following patch
5842 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5845 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5846 software floating point or no floating point registers, do not
5847 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5848 in GPRs that occurs after we tested for GPRs that would never be
5851 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5852 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5853 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5854 specifically allow DDmode, since that does not use the SPE SIMD
5857 2014-03-21 Mark Wielaard <mjw@redhat.com>
5859 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5860 as unsigned or int depending on type and value used.
5862 2014-04-15 Richard Biener <rguenther@suse.de>
5864 PR rtl-optimization/56965
5865 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5866 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5868 * alias.c (true_dependence_1): Do not call
5869 nonoverlapping_component_refs_p.
5870 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5871 nonoverlapping_component_refs_p.
5872 (indirect_refs_may_alias_p): Likewise.
5874 2014-04-15 Teresa Johnson <tejohnson@google.com>
5876 * cfg.c (dump_bb_info): Fix flags check.
5877 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5879 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5881 PR rtl-optimization/60663
5882 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5885 2014-04-15 Richard Biener <rguenther@suse.de>
5887 * lto-streamer.h (LTO_major_version): Bump to 4.
5889 2014-04-15 Richard Biener <rguenther@suse.de>
5891 * common.opt (lto_partition_model): New enum.
5892 (flto-partition=): Merge separate options with a single with argument,
5893 add -flto-partition=one support.
5894 * flag-types.h (enum lto_partition_model): Declare.
5895 * opts.c (finish_options): Remove duplicate -flto-partition=
5897 * lto-wrapper.c (run_gcc): Adjust.
5899 2014-04-15 Richard Biener <rguenther@suse.de>
5901 * alias.c (ncr_compar): New function.
5902 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5904 2014-04-15 Richard Biener <rguenther@suse.de>
5906 * alias.c (record_component_aliases): Do not walk BINFOs.
5908 2014-04-15 Richard Biener <rguenther@suse.de>
5910 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5911 Add struct function argument and adjust.
5912 (find_func_aliases_for_call): Likewise.
5913 (find_func_aliases): Likewise.
5914 (find_func_clobbers): Likewise.
5915 (intra_create_variable_infos): Likewise.
5916 (compute_points_to_sets): Likewise.
5917 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5919 2014-04-15 Richard Biener <rguenther@suse.de>
5921 * tree.c (iterative_hash_expr): Use enum tree_code_class
5922 to store TREE_CODE_CLASS.
5923 (tree_block): Likewise.
5924 (tree_set_block): Likewise.
5925 * tree.h (fold_build_pointer_plus_loc): Use
5926 convert_to_ptrofftype_loc.
5928 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5931 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5934 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5936 * cfgloop.h (struct loop): Move force_vectorize down.
5937 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5938 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5939 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5940 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5941 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5942 * tree-core.h (enum annot_expr_kind): Add new kind values.
5943 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5944 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5946 * tree.def (ANNOTATE_EXPR): Tweak comment.
5948 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5950 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5953 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5955 * tree.h (TYPE_IDENTIFIER): Declare.
5956 * tree.c (subrange_type_for_debug_p): Use it.
5957 * godump.c (go_format_type): Likewise.
5958 * dwarf2out.c (is_cxx_auto, modified_type_die,
5959 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5960 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5962 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5965 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5967 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5969 * config/i386/i386.c (examine_argument): Return bool. Return true if
5970 parameter should be passed in memory.
5971 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5972 (construct_container): Update calls to examine_argument.
5973 (function_arg_advance_64): Ditto.
5974 (return_in_memory_32): Merge with ix86_return_in_memory.
5975 (return_in_memory_64): Ditto.
5976 (return_in_memory_ms_64): Ditto.
5978 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5980 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5981 * coverage.c (coverage_compute_profile_id): Handle externally visible
5984 2014-04-14 Martin Jambor <mjambor@suse.cz>
5986 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5987 DECL_DISREGARD_INLINE_LIMITS functions.
5989 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5992 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5994 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5997 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5998 optimize_insn_for_speed_p instead of
5999 optimize_function_for_speed_p.
6001 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6003 * doc/invoke.texi (free): Document AArch64.
6005 2014-04-14 Richard Biener <rguenther@suse.de>
6007 PR tree-optimization/60042
6008 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6009 (insert_into_preds_of_block): Do not prevent PHI insertion
6010 for REFERENCE exprs here ...
6011 (eliminate_dom_walker::before_dom_children): ... but prevent
6012 their use here under similar conditions when applied to the
6013 IL after PRE optimizations.
6015 2014-04-14 Richard Biener <rguenther@suse.de>
6017 * passes.def: Move early points-to after early SRA.
6019 2014-04-14 Richard Biener <rguenther@suse.de>
6021 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6022 check for which sign-changes we allow when forwarding
6023 a converted value into a switch.
6025 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6027 * stor-layout.c (place_field): Finalize non-constant offset for the
6030 2014-04-14 Richard Biener <rguenther@suse.de>
6032 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6034 (expand_switch_using_bit_tests_p): Likewise.
6035 (process_switch): Compute and pass on speed_p based on the
6037 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6038 optimize_bb_for_speed_p.
6040 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6042 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6043 * function.h (struct function): Rename has_force_vect_loops into
6044 has_force_vectorize_loops.
6045 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6046 (input_struct_function_base): Likewise.
6047 * lto-streamer-out.c (output_cfg): Likewise.
6048 (output_struct_function_base): Likewise.
6049 * omp-low.c (expand_omp_simd): Likewise.
6050 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6051 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6052 (version_loop_for_if_conversion): Likewise.
6053 (tree_if_conversion): Likewise.
6054 (main_tree_if_conversion): Likewise.
6055 (gate_tree_if_conversion): Likewise.
6056 * tree-inline.c (copy_loops): Likewise.
6057 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6058 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6059 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6060 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6061 * tree-vectorizer.c (vectorize_loops): Likewise.
6062 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6064 2014-04-14 Richard Biener <rguenther@suse.de>
6067 * lto-streamer-out.c (wrap_refs): New function.
6068 (lto_output): Wrap symbol references in global initializes in
6069 type-preserving MEM_REFs.
6071 2014-04-14 Christian Bruel <christian.bruel@st.com>
6073 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6075 2014-04-14 Christian Bruel <christian.bruel@st.com>
6077 * config/sh/sh.md (setmemqi): New expand pattern.
6078 * config/sh/sh.h (CLEAR_RATIO): Define.
6079 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6080 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6082 2014-04-14 Richard Biener <rguenther@suse.de>
6085 * fold-const.c (negate_expr_p): Don't negate directional rounding
6087 (fold_negate_expr): Likewise.
6089 2014-04-14 Richard Biener <rguenther@suse.de>
6091 PR tree-optimization/59817
6092 PR tree-optimization/60453
6093 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6094 recursion to catch all CHRECs in the scalar evolution and restrict
6095 the predicate for the remains appropriately.
6097 2014-04-12 Catherine Moore <clm@codesourcery.com>
6099 * config/mips/constraints.md: Add new register constraint "kb".
6100 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6101 (*movhi_internal): Likewise.
6102 (*movqi_internal): Likewise.
6103 * config/mips/mips.h (M16_STORE_REGS): New register class.
6104 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6105 (REG_CLASS_CONTENTS): Likewise.
6106 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6108 2014-04-11 Tobias Burnus <burnus@net-b.de>
6111 * doc/invoke.texi (-Wformat-signedness): Document it.
6112 (Wformat=2): Mention that this enables -Wformat-signedness.
6114 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6116 * common/config/epiphany/epiphany-common.c
6117 (epiphany_option_optimization_table): Enable section anchors by
6118 default at -O1 or higher.
6119 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6120 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6121 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6122 carries no extra cost.
6123 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6124 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6125 * config/epiphany/predicates.md (memclob_operand): New predicate.
6126 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6127 Use memclob_operand predicate and X constraint for operand 3.
6129 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6131 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6132 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6135 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6137 PR rtl-optimization/60651
6138 * mode-switching.c (optimize_mode_switching): Make sure to emit
6139 sets of a lower numbered entity before sets of a higher numbered
6140 entity to a mode of the same or lower priority.
6141 When creating a seginfo for a basic block that starts with a code
6142 label, move the insertion point past the code label.
6143 (new_seginfo): Document and enforce requirement that
6144 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6145 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6146 * doc/tm.texi: Regenerate.
6148 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6151 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6153 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6155 * BASE-VER: Set to 4.10.0.
6157 2014-04-11 Tobias Burnus <burnus@net-b.de>
6160 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6161 * doc/gcc.texi (Service): Update description in the @menu
6162 * doc/invoke.texi (Option Summary): Remove misplaced and
6165 2014-04-11 Steve Ellcey <sellcey@mips.com>
6166 Jakub Jelinek <jakub@redhat.com>
6169 * expr.c (convert_move): Use emit_store_flag_force instead of
6170 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6173 2014-04-11 Richard Biener <rguenther@suse.de>
6176 * varasm.c (assemble_alias): Avoid endless error reporting
6177 recursion by setting TREE_ASM_WRITTEN.
6179 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6181 * config/s390/s390.md: Add a splitter for NOT rtx.
6183 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6185 PR rtl-optimization/60663
6186 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6188 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6189 Jakub Jelinek <jakub@redhat.com>
6192 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6193 flag from decl_node to node.
6195 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6198 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6199 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6200 ameliorating the cases where it can be.
6202 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6205 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6207 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6208 (loadsync_<mode>): Change mode.
6209 (load_quadpti, store_quadpti): New.
6210 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6211 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6212 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6214 2014-04-09 Cong Hou <congh@google.com>
6217 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6220 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6222 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6223 instead of vnor to exploit possible fusion opportunity in the
6225 (altivec_expand_vec_perm_const_le): Likewise.
6227 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6229 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6230 (loadsync_<mode>): Change mode.
6231 (load_quadpti, store_quadpti): New.
6232 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6233 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6235 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6238 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6239 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6240 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6242 2014-04-08 Richard Biener <rguenther@suse.de>
6245 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6246 a 64bit widest int print double-int similar to on HWI64 hosts.
6248 2014-04-08 Richard Biener <rguenther@suse.de>
6250 PR tree-optimization/60785
6251 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6252 default defs properly.
6254 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6256 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6257 (Weffc++): Likewise.
6259 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6261 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6262 set completep to false rather than true.
6264 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6267 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6268 ARM_TARGET2_DWARF_FORMAT.
6270 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6273 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6274 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6277 2014-04-07 Richard Biener <rguenther@suse.de>
6279 PR tree-optimization/60766
6280 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6281 (may_eliminate_iv): Convert cand_value_at result to desired type.
6283 2014-04-07 Jason Merrill <jason@redhat.com>
6286 * common.opt (-fno-gnu-unique): Add.
6287 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6289 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6291 * haifa-sched.c: Fix outdated function reference and minor
6292 grammar errors in introductory comment.
6294 2014-04-07 Richard Biener <rguenther@suse.de>
6297 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6299 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6301 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6304 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6305 size accounting for thunks.
6306 (pa_asm_output_mi_thunk): Use final_start_function() and
6307 final_end_function() to output function start and end directives.
6309 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6311 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6312 device specific ISA/ feature information. Remove short_sp and
6313 errata_skip ds. Add avr_device_specific_features enum to have device
6315 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6316 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6317 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6318 updated device specific info.
6319 * config/avr/avr-mcus.def: Merge device specific details to
6320 dev_attribute field.
6321 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6323 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6324 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6325 assembler if RMW isa supported by current device.
6326 * config/avr/genmultilib.awk: Update as device info structure changed.
6327 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6329 2014-04-04 Cong Hou <congh@google.com>
6331 PR tree-optimization/60656
6332 * tree-vect-stmts.c (supportable_widening_operation):
6333 Fix a bug that elements in a vector with vect_used_by_reduction
6334 property are incorrectly reordered when the operation on it is not
6335 consistant with the one in reduction operation.
6337 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6339 PR rtl-optimization/60155
6340 * gcse.c (record_set_data): New function.
6341 (single_set_gcse): New function.
6342 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6343 (hoist_code): Likewise.
6344 (get_pressure_class_and_nregs): Likewise.
6346 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6348 * explow.c (probe_stack_range): Emit a final optimization blockage.
6350 2014-04-04 Anthony Green <green@moxielogic.com>
6352 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6355 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6358 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6359 flags are set only during streaming.
6360 * ipa.c (process_references, walk_polymorphic_call_targets,
6361 symtab_remove_unreachable_nodes): Drop bodies of always inline
6362 after early inlining.
6363 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6365 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6366 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6369 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6372 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6375 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6377 (cortex_a53_fdivd): Likewise.
6379 2014-04-04 Martin Jambor <mjambor@suse.cz>
6382 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6384 * cgraph.c (clone_of_p): Also return true if thunks match.
6385 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6386 cgraph_function_or_thunk_node and an obsolete comment.
6387 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6389 (build_function_decl_skip_args): Likewise.
6390 (set_new_clone_decl_and_node_flags): New function.
6391 (duplicate_thunk_for_node): Likewise.
6392 (redirect_edge_duplicating_thunks): Likewise.
6393 (cgraph_clone_node): New parameter args_to_skip, pass it to
6394 redirect_edge_duplicating_thunks which is called instead of
6395 cgraph_redirect_edge_callee.
6396 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6397 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6399 2014-04-04 Jeff Law <law@redhat.com>
6402 * config/arm/predicates.md (const_int_I_operand): New predicate.
6403 (const_int_M_operand): Similarly.
6404 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6406 (insv_t2, extv_reg, extzv_t2): Likewise.
6407 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6408 (pop_multiple_with_writeback_and_return): Likewise.
6409 (vfp_pop_multiple_with_writeback): Likewise
6411 2014-04-04 Richard Biener <rguenther@suse.de>
6414 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6415 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6417 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6418 * gimplify.c (gimple_add_tmp_var_fn): New function.
6419 * gimple-expr.h (create_tmp_reg_fn): Declare.
6420 * gimple-expr.c (create_tmp_reg_fn): New function.
6421 * gimple-low.c (record_vars_into): Don't change cfun.
6422 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6423 code generation without cfun.
6425 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6428 * Makefile.in (install-driver): Fix shell scripting.
6430 2014-04-03 Cong Hou <congh@google.com>
6432 PR tree-optimization/60505
6433 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6434 threshold of number of iterations below which no vectorization
6436 * tree-vect-loop.c (new_loop_vec_info):
6437 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6438 * tree-vect-loop.c (vect_analyze_loop_operations):
6439 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6440 * tree-vect-loop.c (vect_transform_loop):
6441 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6442 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6443 of iterations of the loop and see if we should build the epilogue.
6445 2014-04-03 Richard Biener <rguenther@suse.de>
6447 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6448 (streamer_tree_cache_create): Adjust.
6449 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6450 to allow optional nodes array.
6451 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6452 (streamer_tree_cache_append): Likewise.
6453 (streamer_tree_cache_create): Create nodes array optionally
6454 as specified by parameter.
6455 * lto-streamer-out.c (create_output_block): Avoid maintaining
6456 the node array in the writer cache.
6457 (DFS_write_tree): Remove assertion.
6458 (produce_asm_for_decls): Free the out decl state hash table early.
6459 * lto-streamer-in.c (lto_data_in_create): Adjust for
6460 streamer_tree_cache_create prototype change.
6462 2014-04-03 Richard Biener <rguenther@suse.de>
6464 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6465 set TREE_CHAIN to NULL_TREE.
6467 2014-04-03 Richard Biener <rguenther@suse.de>
6469 PR tree-optimization/60740
6470 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6471 over all GIMPLE_COND operands.
6473 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6475 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6476 (Weffc++): Remove Scott's numbering, merge lists and reference
6479 2014-04-03 Nick Clifton <nickc@redhat.com>
6481 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6484 2014-04-03 Martin Jambor <mjambor@suse.cz>
6486 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6487 mention gcc_unreachable before failing.
6488 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6491 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6494 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6495 inconsistent code and instead mark the context inconsistent.
6496 (possible_polymorphic_call_targets): For inconsistent contexts
6497 return empty complete list.
6499 2014-04-02 Anthony Green <green@moxielogic.com>
6501 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6502 (extendqisi2, extendhisi2): Define.
6503 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6504 (WCHAR_TYPE): Change to unsigned int.
6506 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6508 PR tree-optimization/60733
6509 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6510 insertion point for PHI candidates to be the end of the feeding
6511 block for the PHI argument.
6513 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6515 PR rtl-optimization/60650
6516 * lra-constraints.c (process_alt_operands): Decrease reject for
6517 earlyclobber matching.
6519 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6521 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6523 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6525 * config/spu/spu.c (pad_bb): Do not crash when the last
6526 insn is CODE_FOR_blockage.
6528 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6530 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6531 lies outside the target mode.
6533 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6536 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6537 software floating point or no floating point registers, do not
6538 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6539 in GPRs that occurs after we tested for GPRs that would never be
6542 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6543 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6544 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6545 specifically allow DDmode, since that does not use the SPE SIMD
6548 2014-04-02 Richard Biener <rguenther@suse.de>
6551 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6552 MODE_INTs. Properly use negv_optab.
6553 (expand_abs): Likewise.
6555 2014-04-02 Richard Biener <rguenther@suse.de>
6558 * Makefile.in (install-driver): Guard extra installs with special
6561 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6563 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6566 2014-04-01 Richard Henderson <rth@redhat.com>
6569 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6570 alternative enabled before register allocation.
6572 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6574 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6575 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6577 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6578 (nios2_got_address): Update nios2_large_got_address call site.
6579 (nios2_delegitimize_address): New function.
6580 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6581 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6582 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6584 2014-04-01 Martin Husemann <martin@duskware.de>
6586 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6589 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6591 PR rtl-optimization/60604
6592 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6593 check from register_operand.
6594 (register_operand): Redefine in terms of general_operand.
6595 (nonmemory_operand): Use register_operand for the non-constant cases.
6597 2014-04-01 Richard Biener <rguenther@suse.de>
6599 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6601 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6603 * doc/invoke.texi (mapp-regs): Clarify.
6605 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6607 * config/i386/avx512fintrin.h (__v32hi): Define type.
6608 (__v64qi): Likewise.
6609 (_mm512_set1_epi8): Define.
6610 (_mm512_set1_epi16): Define.
6611 (_mm512_set4_epi32): Define.
6612 (_mm512_set4_epi64): Define.
6613 (_mm512_set4_pd): Define.
6614 (_mm512_set4_ps): Define.
6615 (_mm512_setr4_epi64): Define.
6616 (_mm512_setr4_epi32): Define.
6617 (_mm512_setr4_pd): Define.
6618 (_mm512_setr4_ps): Define.
6619 (_mm512_setzero_epi32): Define.
6621 2014-03-31 Martin Jambor <mjambor@suse.cz>
6624 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6625 callsite_arguments_match_p. Updated all callers. Also check types of
6626 corresponding formal parameters and actual arguments.
6627 (not_all_callers_have_enough_arguments_p) Renamed to
6628 some_callers_have_mismatched_arguments_p.
6630 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6632 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6634 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6637 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6640 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6642 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6643 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6645 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6646 Use FMAMODE_NOVF512 mode iterator.
6647 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6648 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6649 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6650 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6652 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6653 Use VF_128_256 mode iterator.
6654 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6657 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6659 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6660 static chain if needed.
6662 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6665 * lra-constraints.c (index_part_to_reg): New.
6666 (process_address): Use it.
6668 2014-03-27 Jeff Law <law@redhat.com>
6669 Jakub Jelinek <jakub@redhat.com>
6672 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6673 gen_rtx_{PLUS,MULT} to build up the address expression.
6675 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6676 creating non-canonical RTL.
6678 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6681 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6682 functions; reorganize to make cheap checks first.
6683 (inline_small_functions): Do not estimate growth when dumping;
6685 * ipa-inline.h (inline_summary): Add min_size.
6686 (growth_likely_positive): New function.
6687 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6688 (set_cond_stmt_execution_predicate): Cleanup.
6689 (estimate_edge_size_and_time): Compute min_size.
6690 (estimate_calls_size_and_time): Likewise.
6691 (estimate_node_size_and_time): Likewise.
6692 (inline_update_overall_summary): Update min_size.
6693 (do_estimate_edge_time): Likewise.
6694 (do_estimate_edge_size): Update.
6695 (do_estimate_edge_hints): Update.
6696 (growth_likely_positive): New function.
6698 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6701 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6702 also if addr has VOIDmode.
6704 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6706 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6707 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6709 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6710 instructions as well as AdvancedSIMD loads.
6712 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6714 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6715 Use crypto_aese type.
6716 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6717 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6718 crypto_aese, crypto_aesmc. Move to types.md.
6719 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6721 * config/arm/iterators.md (crypto_type): Likewise.
6723 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6725 * cgraph.c: Include expr.h and tree-dfa.h.
6726 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6729 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6732 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6733 regs from checking multi-reg pseudos.
6735 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6737 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6739 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6741 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6742 if it would clobber the stack pointer, even temporarily.
6744 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6746 * mode-switching.c: Make small adjustments to the top comment.
6748 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6750 * config/rs6000/constraints.md (wD constraint): New constraint to
6751 match the constant integer to get the top DImode/DFmode out of a
6752 vector in a VSX register.
6754 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6755 match the constant integer to get the top DImode/DFmode out of a
6756 vector in a VSX register.
6758 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6761 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6764 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6765 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6767 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6768 Optimize vec_extract of 64-bit values, where the value being
6769 extracted is in the top word, where we can use scalar
6770 instructions. Add direct move and store support. Combine the big
6771 endian/little endian vector select load support into a single insn.
6772 (vsx_extract_<mode>_internal1): Likewise.
6773 (vsx_extract_<mode>_internal2): Likewise.
6774 (vsx_extract_<mode>_load): Likewise.
6775 (vsx_extract_<mode>_store): Likewise.
6776 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6777 combined into vsx_extract_<mode>_load.
6778 (vsx_extract_<mode>_one_le): Likewise.
6780 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6781 define the top 64-bit vector element.
6783 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6786 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6787 Document vec_vbpermq builtin.
6790 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6791 enable use of xxsldwi and xxpermdi builtin functions.
6792 (vec_xxpermdi): Likewise.
6794 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6795 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6797 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6799 PR rtl-optimization/60650
6800 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6802 (find_spills_for): New.
6803 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6804 Spill all pseudos on the second iteration.
6806 2014-03-27 Marek Polacek <polacek@redhat.com>
6809 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6812 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6814 * config/s390/s390.c (s390_can_use_return_insn): Check for
6815 call-saved FPRs on 31 bit.
6817 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6820 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6821 if they need regimplification, just drop them instead of
6822 calling gimple_regimplify_operands on them.
6824 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6827 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6828 (aarch64_frame_pointer_required): Adjust logic.
6829 (aarch64_can_eliminate): Adjust logic.
6830 (aarch64_override_options_after_change): Adjust logic.
6832 2014-03-27 Dehao Chen <dehao@google.com>
6834 * ipa-inline.c (early_inliner): Update node's inline info.
6836 2014-03-26 Dehao Chen <dehao@google.com>
6838 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6839 compiler inserted conditional jumps for NAN float check.
6841 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6843 * ubsan.h (ubsan_create_data): Change second argument's type
6844 to const location_t *.
6845 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6847 (ubsan_create_data): Change second argument to const location_t *PLOC.
6848 Create Loc field whenever PLOC is non-NULL.
6849 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6850 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6854 * real.c (real_to_integer2): Change type of low to UHWI.
6856 2014-03-26 Tobias Burnus <burnus@net-b.de>
6858 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6859 (CILK_SELF_SPECS): New define.
6860 (driver_self_specs): Use it.
6862 2014-03-26 Richard Biener <rguenther@suse.de>
6864 * tree-pretty-print.c (percent_K_format): Implement special
6865 case for LTO and its stripped down BLOCK tree.
6867 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6870 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6872 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6873 one range is range_int_cst_p, but not both, at least optimize
6874 addition/subtraction of 0 and multiplication by 0 or 1.
6875 * gimple-fold.c (gimple_fold_call): Fold
6876 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6877 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6878 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6880 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6882 PR rtl-optimization/60452
6883 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6884 <case REG>: Return 1 for invalid offsets from the frame pointer.
6886 2014-03-26 Marek Polacek <polacek@redhat.com>
6889 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6892 2014-03-26 Marek Polacek <polacek@redhat.com>
6895 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6898 2014-03-26 Marek Polacek <polacek@redhat.com>
6901 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6902 multiplication in unsigned type.
6904 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6906 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6908 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6910 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6912 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6915 * cif-code.def (UNREACHABLE) New code.
6916 * ipa-inline.c (inline_small_functions): Skip edges to
6917 __builtlin_unreachable.
6918 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6919 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6920 predicate to __bulitin_unreachable.
6921 (set_cond_stmt_execution_predicate): Fix issue when
6922 invert_tree_comparison returns ERROR_MARK.
6923 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6924 propagate to inline clones.
6925 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6927 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6928 * cgraphclones.c (cgraph_clone_node): If call destination is already
6929 ureachable, do not redirect it back.
6930 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6933 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6935 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6936 Do not modify inline clones.
6938 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6940 * config/i386/i386.md (general_sext_operand): New mode attr.
6941 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6942 don't generate (sign_extend (const_int)).
6943 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6944 operands[2]. Use We constraint instead of <i> and
6945 <general_sext_operand> predicate instead of <general_operand>.
6946 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6947 * config/i386/constraints.md (We): New constraint.
6948 * config/i386/predicates.md (x86_64_sext_operand,
6949 sext_operand): New predicates.
6951 2014-03-25 Martin Jambor <mjambor@suse.cz>
6954 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6955 inconsistent devirtualizations to __builtin_unreachable.
6957 2014-03-25 Marek Polacek <polacek@redhat.com>
6960 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6962 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6964 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6965 order of elements for big-endian.
6967 2014-03-25 Richard Biener <rguenther@suse.de>
6970 * gimplify-me.c (gimple_regimplify_operands): Update the
6973 2014-03-25 Martin Jambor <mjambor@suse.cz>
6976 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6977 (lto_output_varpool_node): Likewise.
6978 (input_overwrite_node): Likewise.
6979 (input_varpool_node): Likewise.
6981 2014-03-25 Richard Biener <rguenther@suse.de>
6983 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6984 (run_gcc): Likewise.
6986 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6988 * combine.c (simplify_compare_const): Add MODE argument.
6989 Handle mode_width 0 as very large mode_width.
6990 (try_combine, simplify_comparison): Adjust callers.
6992 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6993 type to avoid signed integer overflow.
6994 * explow.c (plus_constant): Likewise.
6996 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6998 * doc/generic.texi: Correct typos.
7000 2014-03-24 Tobias Burnus <burnus@net-b.de>
7002 * doc/invoke.texi (-flto): Expand section about
7003 using static libraries with LTO.
7005 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7007 PR rtl-optimization/60501
7008 * optabs.def (addptr3_optab): New optab.
7009 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7010 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7011 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7013 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7015 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7017 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7019 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7022 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7023 (_mm256_undefined_ps): Define.
7024 (_mm256_undefined_pd): Define.
7025 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7026 (_mm_undefined_pd): Define.
7027 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7028 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7029 (_mm512_undefined_ps): Define.
7030 (_mm512_undefined_pd): Define.
7031 Use _mm*_undefined_*.
7032 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7034 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7036 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7037 (lshr_simd): DI mode added.
7038 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7039 (aarch64_ushr_simddi): Likewise.
7040 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7041 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7042 (vshrd_n_u64): Likewise.
7044 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7046 * Makefile.in (s-macro_list): Depend on cc1.
7048 2014-03-23 Teresa Johnson <tejohnson@google.com>
7050 * ipa-utils.c (ipa_print_order): Use specified dump file.
7052 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7054 PR rtl-optimization/60601
7055 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7057 * gcc.c (eval_spec_function): Initialize save_growing_value.
7059 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7062 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7063 code == MINUS_EXPR, never swap op0 with op1.
7065 * toplev.c (init_local_tick): Avoid signed integer multiplication
7067 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7068 shift by first operand's bitsize.
7070 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7073 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7075 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7076 TARGET_ISA_64BIT_P(x).
7078 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7080 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7081 pattern for vector nor instead of subtract from splat(-1).
7082 (altivec_expand_vec_perm_const_le): Likewise.
7084 2014-03-21 Richard Henderson <rth@twiddle.net>
7087 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7088 related insns after epilogue_completed.
7090 2014-03-21 Martin Jambor <mjambor@suse.cz>
7093 * cgraph.h (symtab_node): New flag body_removed.
7094 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7095 when removing bodies.
7096 * symtab.c (dump_symtab_base): Dump body_removed flag.
7097 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7098 had their bodies removed.
7100 2014-03-21 Martin Jambor <mjambor@suse.cz>
7103 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7106 2014-03-21 Richard Biener <rguenther@suse.de>
7108 PR tree-optimization/60577
7109 * tree-core.h (struct tree_base): Document nothrow_flag use
7110 in VAR_DECL_NONALIASED.
7111 * tree.h (VAR_DECL_NONALIASED): New.
7112 (may_be_aliased): Adjust.
7113 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7115 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7117 * expr.c (expand_expr_real_1): Remove outdated comment.
7119 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7122 * ira.c (adjust_cleared_regs): Call copy_rtx on
7123 *reg_equiv[REGNO (loc)].src_p before passing it to
7124 simplify_replace_fn_rtx.
7127 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7128 into CONST, put pic register as first operand of PLUS. Use
7129 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7131 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7133 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7135 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7137 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7138 around for store forwarding issue in the FPU on the UT699.
7139 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7140 loads and operations if -mfix-ut699 is specified.
7141 (divtf3_hq): Tweak attribute.
7142 (sqrttf2_hq): Likewise.
7144 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7146 * calls.c (store_one_arg): Remove incorrect const qualification on the
7147 type of the temporary.
7148 * cfgexpand.c (expand_return): Likewise.
7149 * expr.c (expand_constructor): Likewise.
7150 (expand_expr_real_1): Likewise.
7152 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7154 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7157 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7160 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7162 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7164 * config/arm/aarch-common-protos.h
7165 (alu_cost_table): Fix spelling of "extend".
7166 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7168 2014-03-19 Richard Biener <rguenther@suse.de>
7171 * tree-core.h (tree_type_common): Re-order pointer members
7172 to reduce recursion depth during GC walks.
7174 2014-03-19 Marek Polacek <polacek@redhat.com>
7177 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7178 before accessing it.
7180 2014-03-19 Richard Biener <rguenther@suse.de>
7183 * lto-streamer-in.c (input_function): In WPA stage do not drop
7186 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7188 PR tree-optimization/60559
7189 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7190 with build_zero_cst assignment.
7192 2014-03-18 Kai Tietz <ktietz@redhat.com>
7194 PR rtl-optimization/56356
7195 * sdbout.c (sdbout_parms): Verify that parms'
7196 incoming argument is valid.
7197 (sdbout_reg_parms): Likewise.
7199 2014-03-18 Richard Henderson <rth@redhat.com>
7202 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7203 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7204 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7206 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7208 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7209 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7210 Italicize plugin event names in description. Explain that
7211 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7212 Remind that no GCC functions should be called after PLUGIN_FINISH.
7213 Explain what pragmas with expansion are.
7215 2014-03-18 Martin Liska <mliska@suse.cz>
7217 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7218 gimple call statement is update.
7219 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7220 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7222 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7225 * ubsan.c (ubsan_instrument_unreachable): Call
7226 initialize_sanitizer_builtins.
7227 (ubsan_pass): Likewise.
7230 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7231 varpool_finalize_decl instead of rest_of_decl_compilation.
7233 2014-03-18 Richard Biener <rguenther@suse.de>
7235 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7236 by using bitmap_and_compl instead of bitmap_and_compl_into.
7237 (df_rd_transfer_function): Likewise.
7239 2014-03-18 Richard Biener <rguenther@suse.de>
7241 * doc/lto.texi (fresolution): Fix typo.
7243 2014-03-18 Richard Biener <rguenther@suse.de>
7245 * doc/invoke.texi (flto): Update for changes in 4.9.
7247 2014-03-18 Richard Biener <rguenther@suse.de>
7249 * doc/loop.texi: Remove section on the removed lambda framework.
7250 Update loop docs with recent changes in preserving loop structure.
7252 2014-03-18 Richard Biener <rguenther@suse.de>
7254 * doc/lto.texi (-fresolution): Document.
7256 2014-03-18 Richard Biener <rguenther@suse.de>
7258 * doc/contrib.texi: Adjust my name.
7260 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7263 * internal-fn.c: Include diagnostic-core.h.
7264 (expand_BUILTIN_EXPECT): New function.
7265 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7266 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7267 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7268 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7270 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7271 Revert 3 argument __builtin_expect code.
7272 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7273 * gimple-fold.c (gimple_fold_call): Likewise.
7274 * tree.h (fold_builtin_expect): New prototype.
7275 * builtins.c (build_builtin_expect_predicate): Add predictor
7276 argument, if non-NULL, create 3 argument __builtin_expect.
7277 (fold_builtin_expect): No longer static. Add ARG2 argument,
7278 pass it through to build_builtin_expect_predicate.
7279 (fold_builtin_2): Adjust caller.
7280 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7281 * internal-fn.def (BUILTIN_EXPECT): New.
7283 2014-03-18 Tobias Burnus <burnus@net-b.de>
7286 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7287 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7288 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7290 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7293 * predict.c (combine_predictions_for_bb): Fix up formatting.
7294 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7295 fill what it points to if non-NULL.
7296 (tree_predict_by_opcode): Adjust caller, use the predictor.
7297 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7299 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7301 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7302 proper constant for the store mode.
7304 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7306 * symtab.c (change_decl_assembler_name): Fix transparent alias
7309 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7311 * config/aarch64/aarch64.c: Correct the comments about the
7312 aarch64 stack layout.
7314 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7316 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7317 check for GF_OMP_FOR_KIND_FOR.
7319 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7321 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7322 ymm and zmm register names.
7324 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7327 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7328 note creation for the 2010-08-31 changes.
7330 2014-03-17 Marek Polacek <polacek@redhat.com>
7333 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7334 as -fno-tree-loop-vectorize.
7335 (expand_omp_simd): Likewise.
7337 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7339 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7340 (eligible_for_call_delay): New prototype.
7341 * config/sparc/sparc.c (tls_call_delay): Rename into...
7342 (eligible_for_call_delay): ...this. Return false if the instruction
7343 cannot be put in the delay slot of a branch.
7344 (eligible_for_restore_insn): Simplify.
7345 (eligible_for_return_delay): Return false if the instruction cannot be
7346 put in the delay slot of a branch and simplify.
7347 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7348 put in the delay slot of a branch.
7349 * config/sparc/sparc.md (fix_ut699): New attribute.
7350 (tls_call_delay): Delete.
7351 (in_call_delay): Reimplement.
7352 (eligible_for_sibcall_delay): Rename into...
7353 (in_sibcall_delay): ...this.
7354 (eligible_for_return_delay): Rename into...
7355 (in_return_delay): ...this.
7356 (in_branch_delay): Reimplement.
7357 (in_uncond_branch_delay): Delete.
7358 (in_annul_branch_delay): Delete.
7360 2014-03-14 Richard Henderson <rth@redhat.com>
7363 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7364 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7365 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7366 (floathi splitters): Remove.
7367 (float<SWI48x>xf2): New pattern.
7368 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7369 code that tried to handle DImode for 32-bit, but which was excluded
7370 by the pattern's condition. Drop allocation of stack temporary.
7371 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7372 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7373 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7374 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7375 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7376 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7377 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7378 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7379 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7380 (*float<SWI48x><X87MODEF>2_i387): Remove.
7381 (all float _with_temp splitters): Remove.
7382 (*float<SWI48x><MODEF>2_i387): New pattern.
7383 (*float<SWI48><MODEF>2_sse): New pattern.
7384 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7385 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7387 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7388 Marek Polacek <polacek@redhat.com>
7391 * common.opt (dump_base_name_prefixed): New Variable.
7392 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7393 if x_dump_base_name_prefixed is already set, set it at the end.
7395 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7397 PR rtl-optimization/60508
7398 * lra-constraints.c (get_reload_reg): Add new parameter
7400 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7401 Pass the new parameter values.
7403 2014-03-14 Richard Biener <rguenther@suse.de>
7405 * common.opt: Revert unintented changes from r205065.
7408 2014-03-14 Richard Biener <rguenther@suse.de>
7411 * cfghooks.c (split_block): Properly adjust all loops the
7412 block was a latch of.
7414 2014-03-14 Martin Jambor <mjambor@suse.cz>
7417 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7420 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7423 * config/avr/avr.c (avr_set_current_function): Pass function name
7424 through default_strip_name_encoding before sanity checking instead
7425 of skipping the first char of the assembler name.
7427 2014-03-13 Richard Henderson <rth@redhat.com>
7430 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7431 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7432 * config/i386/i386-protos.h: Likewise.
7433 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7434 in the expander instead of a splitter.
7435 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7436 any possibility of requiring a memory.
7437 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7438 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7439 (fp branch splitters): Update for ix86_split_fp_branch.
7440 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7441 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7442 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7443 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7444 (*fop_<MODEF>_3_i387): Likewise.
7445 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7446 (splitters for the fop_* register patterns): Remove.
7447 (fscalexf4_i387): Rename from *fscalexf4_i387.
7448 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7450 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7452 PR tree-optimization/59779
7453 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7454 type for bitsize and maxsize instead of HOST_WIDE_INT.
7456 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7458 PR rtl-optimization/57320
7459 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7460 the CFG after thread_prologue_and_epilogue_insns.
7462 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7464 PR rtl-optimization/57189
7465 * lra-constraints.c (process_alt_operands): Disfavor spilling
7468 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7470 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7472 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7474 PR tree-optimization/59025
7476 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7477 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7479 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7482 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7483 calls of avr_out_plus_1.
7485 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7487 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7488 BB's single pred and update the father loop's latch info later.
7490 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7492 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7496 (VEC_base): Likewise.
7497 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7498 registers, we need to swap double words in little endian mode.
7500 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7501 to be a container mode for 128-bit integer operations added in ISA
7502 2.07. Unlike TImode and PTImode, the preferred register set is
7503 the Altivec/VMX registers for the 128-bit operations.
7505 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7507 (rs6000_split_128bit_ok_p): Likewise.
7509 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7510 macros for creating ISA 2.07 normal and overloaded builtin
7511 functions with 3 arguments.
7512 (BU_P8V_OVERLOAD_3): Likewise.
7513 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7514 for use as overloaded functions.
7515 (VPERM_1TI_UNS): Likewise.
7516 (VSEL_1TI): Likewise.
7517 (VSEL_1TI_UNS): Likewise.
7518 (ST_INTERNAL_1ti): Likewise.
7519 (LD_INTERNAL_1ti): Likewise.
7520 (XXSEL_1TI): Likewise.
7521 (XXSEL_1TI_UNS): Likewise.
7522 (VPERM_1TI): Likewise.
7523 (VPERM_1TI_UNS): Likewise.
7524 (XXPERMDI_1TI): Likewise.
7525 (SET_1TI): Likewise.
7526 (LXVD2X_V1TI): Likewise.
7527 (STXVD2X_V1TI): Likewise.
7528 (VEC_INIT_V1TI): Likewise.
7529 (VEC_SET_V1TI): Likewise.
7530 (VEC_EXT_V1TI): Likewise.
7531 (EQV_V1TI): Likewise.
7532 (NAND_V1TI): Likewise.
7533 (ORC_V1TI): Likewise.
7534 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7535 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7537 (VADDUQM): Likewise.
7538 (VSUBCUQ): Likewise.
7539 (VADDEUQM): Likewise.
7540 (VADDECUQ): Likewise.
7541 (VSUBEUQM): Likewise.
7542 (VSUBECUQ): Likewise.
7544 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7545 __int128_t and __uint128_t types.
7546 (__uint128_type): Likewise.
7547 (altivec_categorize_keyword): Add support for vector __int128_t,
7548 vector __uint128_t, vector __int128, and vector unsigned __int128
7549 as a container type for TImode operations that need to be done in
7550 VSX/Altivec registers.
7551 (rs6000_macro_to_expand): Likewise.
7552 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7553 to support 128-bit integer instructions vaddcuq, vadduqm,
7554 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7555 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7557 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7558 for V1TImode, and set up preferences to use VSX/Altivec registers.
7559 Setup VSX reload handlers.
7560 (rs6000_debug_reg_global): Likewise.
7561 (rs6000_init_hard_regno_mode_ok): Likewise.
7562 (rs6000_preferred_simd_mode): Likewise.
7563 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7564 (easy_altivec_constant): Likewise.
7565 (output_vec_const_move): Likewise.
7566 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7568 (rs6000_expand_vector_extract): Likewise.
7569 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7571 (rs6000_const_vec): Add support for V1TImode.
7572 (rs6000_emit_le_vsx_load): Swap double words when loading or
7573 storing TImode/V1TImode.
7574 (rs6000_emit_le_vsx_store): Likewise.
7575 (rs6000_emit_le_vsx_move): Likewise.
7576 (rs6000_emit_move): Add support for V1TImode.
7577 (altivec_expand_ld_builtin): Likewise.
7578 (altivec_expand_st_builtin): Likewise.
7579 (altivec_expand_vec_init_builtin): Likewise.
7580 (altivec_expand_builtin): Likewise.
7581 (rs6000_init_builtins): Add support for V1TImode type. Add
7582 support for ISA 2.07 128-bit integer builtins. Define type names
7583 for the VSX/Altivec vector types.
7584 (altivec_init_builtins): Add support for overloaded vector
7585 functions with V1TImode type.
7586 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7587 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7589 (rs6000_split_128bit_ok_p): Likewise.
7590 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7591 __int128_t and vector __uint128_t.
7593 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7594 and mode attributes.
7601 (VS_scalar): Likewise.
7602 (VS_double): Likewise.
7603 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7605 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7606 we support the ISA 2.07 128-bit integer arithmetic instructions.
7607 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7608 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7609 and TImode types for use with the builtin functions.
7610 (V1TI_type_node): Likewise.
7611 (unsigned_V1TI_type_node): Likewise.
7612 (intTI_type_internal_node): Likewise.
7613 (uintTI_type_internal_node): Likewise.
7615 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7616 128-bit builtin functions.
7617 (UNSPEC_VADDEUQM): Likewise.
7618 (UNSPEC_VADDECUQ): Likewise.
7619 (UNSPEC_VSUBCUQ): Likewise.
7620 (UNSPEC_VSUBEUQM): Likewise.
7621 (UNSPEC_VSUBECUQ): Likewise.
7622 (VM): Add V1TImode to vector mode iterators.
7624 (VI_unit): Likewise.
7625 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7626 (altivec_vaddcuq): Likewise.
7627 (altivec_vsubuqm): Likewise.
7628 (altivec_vsubcuq): Likewise.
7629 (altivec_vaddeuqm): Likewise.
7630 (altivec_vaddecuq): Likewise.
7631 (altivec_vsubeuqm): Likewise.
7632 (altivec_vsubecuq): Likewise.
7634 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7636 (BOOL_128): Likewise.
7637 (BOOL_REGS_OUTPUT): Likewise.
7638 (BOOL_REGS_OP1): Likewise.
7639 (BOOL_REGS_OP2): Likewise.
7640 (BOOL_REGS_UNARY): Likewise.
7641 (BOOL_REGS_AND_CR0): Likewise.
7643 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7644 128-bit integer builtin support.
7645 (vec_vadduqm): Likewise.
7646 (vec_vaddecuq): Likewise.
7647 (vec_vaddeuqm): Likewise.
7648 (vec_vsubecuq): Likewise.
7649 (vec_vsubeuqm): Likewise.
7650 (vec_vsubcuq): Likewise.
7651 (vec_vsubuqm): Likewise.
7653 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7654 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7655 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7656 128-bit integer add/subtract to ISA 2.07.
7658 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7660 * config/arc/arc.c (arc_predicate_delay_insns):
7661 Fix third argument passed to conditionalize_nonjump.
7663 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7665 * config/aarch64/aarch64-builtins.c
7666 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7667 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7668 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7669 instead of __builtin_lfloor.
7670 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7672 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7674 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7675 (tree_ssa_ifcombine_bb_1): New function.
7676 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7677 is an empty forwarder block to then_bb or vice versa and then_bb
7678 and else_bb are effectively swapped.
7680 2014-03-12 Christian Bruel <christian.bruel@st.com>
7683 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7684 REG_CFA_DEF_CFA note.
7685 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7686 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7688 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7690 PR tree-optimization/60454
7691 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7693 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7695 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7696 Do not define target_cpu_default2 to generic.
7697 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7698 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7699 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7701 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7702 Marc Glisse <marc.glisse@inria.fr>
7704 PR tree-optimization/60502
7705 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7706 instead of build_low_bits_mask.
7708 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7711 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7712 if there are multiple uses, but op doesn't live on E edge.
7713 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7714 clobber stmts before __builtin_unreachable.
7716 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7718 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7719 hard_frame_pointer_rtx.
7720 * cse.c (cse_insn): Remove volatile check.
7721 * cselib.c (cselib_process_insn): Likewise.
7722 * dse.c (scan_insn): Likewise.
7724 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7726 * config/arc/arc.c (conditionalize_nonjump): New function,
7728 (arc_ifcvt): ... this.
7729 (arc_predicate_delay_insns): Use it.
7731 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7733 * config/arc/predicates.md (extend_operand): During/after reload,
7734 allow const_int_operand.
7735 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7736 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7737 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7739 (umulsi3_highpart_i): Likewise.
7741 2014-03-11 Richard Biener <rguenther@suse.de>
7743 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7744 Add asserts to guard possible wrong-code bugs.
7746 2014-03-11 Richard Biener <rguenther@suse.de>
7748 PR tree-optimization/60429
7749 PR tree-optimization/60485
7750 * tree-ssa-structalias.c (set_union_with_increment): Properly
7751 take into account all fields that overlap the shifted vars.
7752 (do_sd_constraint): Likewise.
7753 (do_ds_constraint): Likewise.
7754 (get_constraint_for_ptr_offset): Likewise.
7756 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7758 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7759 (nios2_compute_frame_layout):
7760 Add calculation of cfun->machine->fp_save_offset.
7761 (nios2_expand_prologue): Correct setting of frame pointer register
7763 (nios2_expand_epilogue): Update recovery of stack pointer from
7764 frame pointer accordingly.
7765 (nios2_initial_elimination_offset): Update calculation of offset
7766 for eliminating to HARD_FRAME_POINTER_REGNUM.
7768 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7771 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7772 cgraph_get_create_node on VAR_DECLs.
7774 2014-03-10 Richard Biener <rguenther@suse.de>
7777 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7779 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7781 * config/vms/vms.opt (vms_float_format): New variable.
7783 2014-03-08 Tobias Burnus <burnus@net-b.de>
7785 * doc/invoke.texi (-fcilkplus): Update implementation status.
7787 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7788 Richard Biener <rguenther@suse.de>
7790 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7791 consistently accross all TUs.
7792 (run_gcc): Enable -fshort-double automatically at link at link-time
7793 and disallow override.
7795 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7798 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7799 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7800 if they can't be used.
7802 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7804 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7805 for Solaris 11/x86 ld.
7806 * configure: Regenerate.
7808 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7810 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7811 (LIB_TLS_SPEC): Save as ld_tls_libs.
7812 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7813 (HAVE_AS_IX86_TLSLDM): New test.
7814 * configure, config.in: Regenerate.
7815 * config/i386/i386.c (legitimize_tls_address): Fall back to
7816 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7817 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7818 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7819 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7821 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7823 * common.opt (fira-loop-pressure): Mark as optimization.
7825 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7827 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7828 an OpenMP mappable type.
7830 2014-03-06 Matthias Klose <doko@ubuntu.com>
7832 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7833 MULTILIB_OSDIRNAMES is not defined.
7835 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7836 Meador Inge <meadori@codesourcery.com>
7839 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7840 (arm_legitimize_address): Call legitimize_tls_address for any
7841 arm_tls_referenced_p expression, handle constant addend. Call it
7842 before testing for !TARGET_ARM.
7843 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7845 2014-03-06 Richard Biener <rguenther@suse.de>
7851 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7853 * tree-streamer.c (record_common_node): Assert we don't record
7854 nodes with type double.
7855 (preload_common_node): Skip type double, complex double and double
7856 pointer since it is now frontend dependent due to fshort-double option.
7858 2014-03-06 Richard Biener <rguenther@suse.de>
7860 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7861 or -fno-lto is specified and the linker has full plugin support.
7862 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7863 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7864 * lto-wrapper.c (merge_and_complain): Merge compile-time
7865 optimization levels.
7866 (run_gcc): And pass it through to the link options.
7868 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7872 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7874 * cselib.c (remove_useless_values): Skip to avoid quadratic
7875 behavior if the condition moved from...
7876 (cselib_process_insn): ... here holds.
7878 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7881 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7882 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7885 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7886 (TM_H): Add x86-tune.def.
7888 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7890 * config/aarch64/aarch64.c (generic_tunings):
7891 Use cortexa57_extra_costs.
7893 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7896 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7897 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7898 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7901 2014-03-04 Heiher <r@hev.cc>
7903 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7904 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7906 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7908 * config/i386/predicates.md (const2356_operand): Change to ...
7909 (const2367_operand): ... this.
7910 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7912 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7913 (*avx512pf_scatterpf<mode>sf): Ditto.
7914 (avx512pf_scatterpf<mode>df): Ditto.
7915 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7916 (*avx512pf_scatterpf<mode>df): Ditto.
7917 * config/i386/i386.c (ix86_expand_builtin): Update
7918 incorrect hint operand error message.
7920 2014-03-04 Richard Biener <rguenther@suse.de>
7922 * lto-section-in.c (lto_get_section_data): Fix const cast.
7924 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7926 * tree-streamer.c (record_common_node): Assert we don't record
7927 nodes with type double.
7928 (preload_common_node): Skip type double, complex double and double
7929 pointer since it is now frontend dependent due to fshort-double option.
7931 2014-03-04 Richard Biener <rguenther@suse.de>
7934 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7935 (lto_input_toplevel_asms): Likewise.
7936 * lto-section-in.c (lto_get_section_data): Instead do it here
7939 2014-03-04 Richard Biener <rguenther@suse.de>
7941 PR tree-optimization/60382
7942 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7943 dead PHIs a reduction.
7945 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7947 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7949 (_mm_prefetch): Move out of GCC target("sse") pragma.
7950 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7951 GCC target("prfchw") pragma.
7952 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7954 * config/i386/i386.c (ix86_option_override_internal): Enable
7955 -mprfchw with -mprefetchwt1.
7957 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7959 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7962 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7964 * opts.h (CL_PCH_IGNORE): Define.
7965 * targhooks.c (option_affects_pch_p):
7966 Return false for options that have CL_PCH_IGNORE set.
7967 * opt-functions.awk: Process PchIgnore.
7968 * doc/options.texi: Document PchIgnore.
7970 * config/arc/arc.opt (misize): Add PchIgnore property.
7972 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7974 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7975 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7976 constraint on constants to permit them being loaded into
7977 GENERAL_REGS or BASE_REGS.
7979 2014-03-03 Nick Clifton <nickc@redhat.com>
7981 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7982 anti-cacnonical alternatives.
7983 (negandhi3_real): New pattern.
7984 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7986 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7988 * config/avr/avr-mcus.def: Remove atxmega16x1.
7989 * config/avr/avr-tables.opt: Regenerate.
7990 * config/avr/t-multilib: Regenerate.
7991 * doc/avr-mmcu.texi: Regenerate.
7993 2014-03-03 Tobias Grosser <tobias@grosser.es>
7994 Mircea Namolaru <mircea.namolaru@inria.fr>
7996 PR tree-optimization/58028
7997 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8000 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8002 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8003 not handled by recognizers.
8005 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8008 * function.c (expand_function_end): Don't emit
8009 clobber_return_register sequence if clobber_after is a BARRIER.
8010 * cfgexpand.c (construct_exit_block): Append instructions before
8011 return_label to prev_bb.
8013 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8015 * config/rs6000/constraints.md: Document reserved use of "wc".
8017 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8020 * ipa.c (function_and_variable_visibility): When dissolving comdat
8021 group, also set all symbols to local.
8023 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8028 2013-12-14 Jan Hubicka <jh@suse.cz>
8030 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8032 2014-03-02 Jon Beniston <jon@beniston.com>
8038 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8039 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8040 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8041 (simple_return, *simple_return): New patterns
8042 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8043 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8045 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8047 * dwarf2out.c (gen_subprogram_die): Tidy.
8049 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8052 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8053 (*mov_t_msb_neg_negc): ... this new insn.
8055 2014-02-28 Jason Merrill <jason@redhat.com>
8058 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8061 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8064 * dwarf2out.c (decltype_auto_die): New static.
8065 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8066 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8067 (is_cxx_auto): Likewise.
8069 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8071 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8072 we are not using general regs only.
8074 2014-02-28 Richard Biener <rguenther@suse.de>
8077 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8078 previous fix and only allow to remove trivial pre-headers
8079 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8080 (remove_forwarder_block): Properly update the latch of a loop.
8082 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8085 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8086 (cselib_preserved_hash_table): New.
8087 (preserve_constants_and_equivs): Move preserved vals to it.
8088 (cselib_find_slot): Look it up first.
8089 (cselib_init): Initialize it.
8090 (cselib_finish): Release it.
8091 (dump_cselib_table): Dump it.
8093 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8096 * cselib.c (remove_useless_values): Skip to avoid quadratic
8097 behavior if the condition moved from...
8098 (cselib_process_insn): ... here holds.
8100 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8103 * var-tracking.c (vt_initialize): Apply the same condition to
8104 preserve the CFA base value.
8106 2014-02-28 Joey Ye <joey.ye@arm.com>
8109 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8110 if reload in progress or completed.
8112 2014-02-28 Tobias Burnus <burnus@net-b.de>
8115 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8118 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8120 * doc/tm.texi.in (Condition Code Status): Update documention for
8121 relative locations of cc0-setter and cc0-user.
8123 2014-02-27 Jeff Law <law@redhat.com>
8125 PR rtl-optimization/52714
8126 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8127 into two independent simple sets, if I3 is a jump, ensure the
8128 pattern we place into I3 is a (set (pc) ...).
8130 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8131 Jeff Law <law@redhat.com>
8133 PR rtl-optimization/49847
8134 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8135 are in different blocks.
8136 * doc/tm.texi (Condition Code Status): Update documention for
8137 relative locations of cc0-setter and cc0-user.
8139 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8142 * lra.c (lra_emit_add): Check SUBREG too.
8144 2014-02-27 Andreas Schwab <schwab@suse.de>
8146 * config/m68k/m68k.c (m68k_option_override): Disable
8147 -flive-range-shrinkage for classic m68k.
8148 (m68k_override_options_after_change): Likewise.
8150 2014-02-27 Marek Polacek <polacek@redhat.com>
8153 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8154 -Wmaybe-uninitialized.
8156 2014-02-27 Alan Modra <amodra@gmail.com>
8159 * reload1.c (emit_input_reload_insns): When reload_override_in,
8160 set old to rl->in_reg when rl->in_reg is a subreg.
8162 2014-02-26 Richard Biener <rguenther@suse.de>
8165 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8167 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8169 * common/config/i386/predicates.md (const1256_operand): Remove.
8170 (const2356_operand): New.
8171 (const_1_to_2_operand): Remove.
8172 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8173 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8174 (*avx512pf_gatherpf<mode>sf): Ditto.
8175 (avx512pf_gatherpf<mode>df): Ditto.
8176 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8177 (*avx512pf_gatherpf<mode>df): Ditto.
8178 (avx512pf_scatterpf<mode>sf): Ditto.
8179 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8180 (*avx512pf_scatterpf<mode>sf): Ditto.
8181 (avx512pf_scatterpf<mode>df): Ditto.
8182 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8183 (*avx512pf_scatterpf<mode>df): Ditto.
8184 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8186 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8188 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8189 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8190 (_mm512_mask_testn_epi64_mask): Move to ...
8191 * config/i386/avx512cdintrin.h: Here.
8192 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8193 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8194 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8195 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8196 TARGET_AVX512F from TARGET_AVX512CD.
8198 2014-02-26 Richard Biener <rguenther@suse.de>
8201 * ipa.c (walk_polymorphic_call_targets): Properly guard
8202 call to inline_update_overall_summary.
8204 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8207 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8208 and latches only if requested. Fix latch if it is removed.
8209 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8210 LOOPS_HAVE_PREHEADERS.
8212 2014-02-25 Andrew Pinski <apinski@cavium.com>
8214 * builtins.c (expand_builtin_thread_pointer): Create a new target
8215 when the target is NULL.
8217 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8219 PR rtl-optimization/60317
8220 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8221 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8222 * lra-assigns.c: Include params.h.
8223 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8224 other reload pseudos considerations.
8226 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8228 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8229 to use canonical form for nor<mode>3.
8231 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8234 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8237 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8239 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8240 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8241 (ix86_handle_option): Handle OPT_mprefetchwt1.
8242 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8243 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8245 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8246 OPTION_MASK_ISA_PREFETCHWT1.
8247 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8248 (PTA_PREFETCHWT1): New.
8249 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8250 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8251 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8252 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8253 (*prefetch_avx512pf_<mode>_: Change into ...
8254 (*prefetch_prefetchwt1_<mode>: This.
8255 * config/i386/i386.opt (mprefetchwt1): New.
8256 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8257 (_mm_prefetch): Handle intent to write.
8258 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8260 2014-02-25 Richard Biener <rguenther@suse.de>
8263 * emit-rtl.c (mem_attrs_htab): Remove.
8264 (mem_attrs_htab_hash): Likewise.
8265 (mem_attrs_htab_eq): Likewise.
8266 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8267 (init_emit_once): Do not allocate mem_attrs_htab.
8269 2014-02-25 Richard Biener <rguenther@suse.de>
8272 * lto-opts.c (lto_write_options): Output non-explicit conservative
8273 -fwrapv, -fno-trapv and -fno-strict-overflow.
8274 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8275 (run_gcc): And pass them through.
8277 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8279 * sel-sched.c (calculate_new_fences): New parameter ptime.
8280 Calculate it as a maximum over all fence cycles.
8281 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8282 Print the final schedule timing when sched_verbose.
8284 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8286 PR rtl-optimization/60292
8287 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8288 bit fot the fence instruction.
8290 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8292 * calls.h: Fix typo in comment.
8294 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8296 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8297 adjusting offsetable addresses.
8299 2014-02-24 Guozhi Wei <carrot@google.com>
8301 * sparseset.h (sparseset_pop): Fix the wrong index.
8303 2014-02-24 Walter Lee <walt@tilera.com>
8305 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8306 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8308 * common/config/tilegx/tilegx-common.c
8309 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8310 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8312 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8313 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8314 (tilegx_gimplify_va_arg_expr): Handle big endian.
8315 (tilegx_expand_unaligned_load): Ditto.
8316 (tilegx_expand_unaligned_store): Ditto.
8317 (TARGET_RETURN_IN_MSB): New.
8318 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8319 (TARGET_ENDIAN_DEFAULT): New.
8320 (TARGET_BIG_ENDIAN): Handle big endian.
8321 (BYTES_BIG_ENDIAN): Ditto.
8322 (WORDS_BIG_ENDIAN): Ditto.
8323 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8326 * config/tilegx/tilegx.md (extv): Handle big endian.
8328 (insn_st<n>): Ditto.
8329 (insn_st<n>_add<bitsuffix>): Ditto.
8330 (insn_stnt<n>): Ditto.
8331 (insn_stnt<n>_add<bitsuffix>):Ditto.
8332 (vec_interleave_highv8qi): Handle big endian.
8333 (vec_interleave_highv8qi_be): New.
8334 (vec_interleave_highv8qi_le): New.
8335 (insn_v1int_h): Handle big endian.
8336 (vec_interleave_lowv8qi): Handle big endian.
8337 (vec_interleave_lowv8qi_be): New.
8338 (vec_interleave_lowv8qi_le): New.
8339 (insn_v1int_l): Handle big endian.
8340 (vec_interleave_highv4hi): Handle big endian.
8341 (vec_interleave_highv4hi_be): New.
8342 (vec_interleave_highv4hi_le): New.
8343 (insn_v2int_h): Handle big endian.
8344 (vec_interleave_lowv4hi): Handle big endian.
8345 (vec_interleave_lowv4hi_be): New.
8346 (vec_interleave_lowv4hi_le): New.
8347 (insn_v2int_l): Handle big endian.
8348 (vec_interleave_highv2si): Handle big endian.
8349 (vec_interleave_highv2si_be): New.
8350 (vec_interleave_highv2si_le): New.
8351 (insn_v4int_h): Handle big endian.
8352 (vec_interleave_lowv2si): Handle big endian.
8353 (vec_interleave_lowv2si_be): New.
8354 (vec_interleave_lowv2si_le): New.
8355 (insn_v4int_l): Handle big endian.
8356 * config/tilegx/tilegx.opt (mbig-endian): New option.
8357 (mlittle-endian): New option.
8358 * doc/install.texi: Document tilegxbe-linux.
8359 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8361 2014-02-24 Martin Jambor <mjambor@suse.cz>
8364 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8365 there are no parameter descriptors.
8367 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8369 PR rtl-optimization/60268
8370 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8371 initialization to ...
8372 (sched_rgn_init): ... here.
8373 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8375 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8377 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8380 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8382 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8385 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8387 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8388 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8390 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8392 * config/microblaze/predicates.md: Add cmp_op predicate.
8393 * config/microblaze/microblaze.md: Add branch_compare instruction
8394 which uses cmp_op predicate and emits cmp insn before branch.
8395 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8396 to microblaze_expand_conditional_branch and consolidate logic.
8397 (microblaze_expand_conditional_branch): emit branch_compare
8398 insn instead of handling cmp op separate from branch insn.
8400 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8402 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8405 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8407 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8408 define_insn with define_expand and new define_insn
8409 *altivec_lve<VI_char>x_internal.
8410 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8411 and new define_insn *altivec_stve<VI_char>x_internal.
8412 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8414 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8416 (altivec_expand_stvex_be): New function.
8418 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8420 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8421 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8422 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8423 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8425 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8428 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8429 instead of emit_move_insn.
8431 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8433 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8435 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8436 gen_altivec_vsumsws.
8438 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8440 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8441 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8442 (altivec_lvxl_<mode>): New define_expand incorporating
8443 -maltivec=be semantics where needed.
8444 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8445 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8446 semantics where needed.
8447 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8448 (altivec_stvx_<mode>): New define_expand incorporating
8449 -maltivec=be semantics where needed.
8450 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8451 VM2 iterator instead of V4SI.
8452 (altivec_stvxl_<mode>): New define_expand incorporating
8453 -maltivec=be semantics where needed.
8454 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8455 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8456 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8457 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8458 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8459 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8460 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8461 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8462 ALTIVEC_BUILTIN_STVXL.
8463 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8464 (altivec_expand_stvx_be): Likewise.
8465 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8466 (altivec_expand_lvx_be): Likewise.
8467 (altivec_expand_stvx_be): Likewise.
8468 (altivec_expand_builtin): Add cases for
8469 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8470 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8471 (altivec_init_builtins): Add definitions for
8472 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8473 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8475 2014-02-21 Catherine Moore <clm@codesourcery.com>
8477 * doc/invoke.texi (mvirt, mno-virt): Document.
8478 * config/mips/mips.opt (mvirt): New option.
8479 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8481 2014-02-21 Richard Biener <rguenther@suse.de>
8483 PR tree-optimization/60276
8484 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8485 (STMT_VINFO_MIN_NEG_DIST): New macro.
8486 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8487 STMT_VINFO_MIN_NEG_DIST.
8488 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8489 made for negative dependence distances still hold.
8491 2014-02-21 Richard Biener <rguenther@suse.de>
8494 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8495 DECL_INITIAL for globals not in the current function context.
8497 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8499 PR tree-optimization/56490
8500 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8501 * tree-ssa-uninit.c: Include params.h.
8502 (compute_control_dep_chain): Add num_calls argument, return false
8503 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8504 num_calls to recursive call.
8505 (find_predicates): Change dep_chain into normal array,
8506 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8507 variable and adjust compute_control_dep_chain caller.
8508 (find_def_preds): Likewise.
8510 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8512 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8513 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8515 2014-02-21 Nick Clifton <nickc@redhat.com>
8517 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8518 (pushhi1): Likewise.
8519 (popqi1): Add mode to pre_dec.
8522 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8524 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8525 mode for mask of V8SFmode permutation.
8527 2014-02-20 Richard Henderson <rth@redhat.com>
8530 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8531 a new pseudo for OLDVAL.
8533 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8536 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8537 gen_reg_rtx if d->testing_p.
8538 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8539 if d->testing_p and we will certainly return true.
8540 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8543 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8545 * emit-rtl.c (gen_reg_rtx): Assert that
8546 crtl->emit.regno_pointer_align_length is non-zero.
8548 2014-02-20 Richard Henderson <rth@redhat.com>
8551 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8552 on failure the store back into EXPECT.
8554 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8555 Sandra Loosemore <sandra@codesourcery.com>
8557 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8558 * config/nios2/nios2.c (nios2_function_profiler): Add
8559 -fPIC (flag_pic == 2) support.
8560 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8561 (nios2_large_offset_p): New function.
8562 (nios2_unspec_reloc_p): Move up position, update to use
8563 nios2_large_offset_p.
8564 (nios2_unspec_address): Remove function.
8565 (nios2_unspec_offset): New function.
8566 (nios2_large_got_address): New function.
8567 (nios2_got_address): Add large offset support.
8568 (nios2_legitimize_tls_address): Update usage of removed and new
8570 (nios2_symbol_binds_local_p): New function.
8571 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8572 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8573 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8574 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8575 processing for (const (unspec ...)).
8576 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8578 2014-02-20 Richard Biener <rguenther@suse.de>
8580 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8581 doing the substitution.
8582 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8584 2014-02-20 Martin Jambor <mjambor@suse.cz>
8587 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8588 info when checking whether lattices are bottom.
8590 2014-02-20 Richard Biener <rguenther@suse.de>
8593 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8596 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8599 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8600 parameter specifying the scaling.
8601 (inline_call): Update.
8602 (want_inline_recursively): Guard division by zero.
8603 (recursive_inlining): Update.
8604 * ipa-inline.h (clone_inlined_nodes): Update.
8606 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8609 * config/i386/i386.c (classify_argument): Pass structures of size
8610 64 bytes or less in register.
8612 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8613 Kirill Yukhin <kirill.yukhin@intel.com>
8615 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8616 (_mm_rcp28_round_ss): Ditto.
8617 (_mm_rsqrt28_round_sd): Ditto.
8618 (_mm_rsqrt28_round_ss): Ditto.
8619 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8620 (_mm_rcp14_round_ss): Ditto.
8621 (_mm_rsqrt14_round_sd): Ditto.
8622 (_mm_rsqrt14_round_ss): Ditto.
8623 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8624 the first input operand, get rid of match_dup.
8625 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8627 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8629 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8630 operand as the first input operand, set type attribute.
8631 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8633 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8634 operand as the first input operand, set type attribute.
8636 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8638 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8641 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8644 * config/i386/i386.c (construct_container): Remove TFmode check
8645 for X86_64_INTEGER_CLASS.
8647 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8650 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8651 only when -Wpsabi is enabled.
8653 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8656 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8657 passing arrays in registers are the same as for structs, so remove the
8658 special case for them.
8660 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8662 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8663 destination type, extract only the valid bits if the source type is not
8664 integral and has a different mode.
8666 2014-02-19 Richard Biener <rguenther@suse.de>
8669 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8672 2014-02-19 Richard Biener <rguenther@suse.de>
8675 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8676 (ipa_modify_call_arguments): Emit an argument load explicitely and
8677 preserve virtual SSA form there and for the replacement call.
8678 Do not update SSA form nor free dominance info.
8680 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8682 * ipa.c (function_and_variable_visibility): Also clear WEAK
8683 flag when disolving COMDAT_GROUP.
8685 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8687 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8688 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8690 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8691 do more sanity checks.
8692 (detect_type_change): Return true when giving up early.
8693 (compute_complex_assign_jump_func): Fix type parameter of
8694 ipa_set_ancestor_jf.
8695 (compute_complex_ancestor_jump_func): Likewise.
8696 (update_jump_functions_after_inlining): Fix updating of
8698 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8700 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8702 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8703 inline clones when edge disappears.
8705 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8708 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8709 Split 64-bit moves into 2 patterns. Do not allow the use of
8710 direct move for TDmode in little endian, since the decimal value
8711 has little endian bytes within a word, but the 64-bit pieces are
8712 ordered in a big endian fashion, and normal subreg's of TDmode are
8714 (mov<mode>_64bit_dm): Likewise.
8715 (movtd_64bit_nodm): Likewise.
8717 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8719 PR tree-optimization/60174
8720 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8721 statement of an SSA_NAME that occurs in an abnormal PHI node.
8723 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8726 * final.c (SEEN_BB): Remove.
8727 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8728 (final_scan_insn): Don't force_source_line on second
8729 NOTE_INSN_BASIC_BLOCK.
8731 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8734 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8735 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8736 (type_natural_mode): Warn ABI change when %zmm register is not
8737 available for AVX512F vector value passing.
8739 2014-02-18 Kai Tietz <ktietz@redhat.com>
8742 * config/i386/i386.c (ix86_expand_prologue): Use value in
8743 rax register as displacement when restoring %r10 or %rax.
8744 Fix wrong offset when restoring both registers.
8746 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8748 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8749 assertion with conditional return.
8751 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8752 Uros Bizjak <ubizjak@gmail.com>
8755 * config/i386/driver-i386.c (host_detect_local_cpu): If
8756 YMM state is not saved by the OS, also clear has_f16c. Move
8757 CPUID 0x80000001 handling before YMM state saving checking.
8759 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8761 PR rtl-optimization/58960
8762 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8763 factored out from ...
8764 (sched_init): ... here.
8765 (free_global_sched_pressure_data): New, factored out from ...
8766 (sched_finish): ... here.
8767 * sched-int.h (free_global_sched_pressure_data): Declare.
8768 * sched-rgn.c (nr_regions_initial): New static global.
8769 (haifa_find_rgns): Initialize it.
8770 (schedule_region): Disable sched-pressure for the newly
8773 2014-02-17 Richard Biener <rguenther@suse.de>
8775 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8776 release SSA defs of pattern stmts.
8778 2014-02-17 Richard Biener <rguenther@suse.de>
8780 * tree-inline.c (expand_call_inline): Release the virtual
8781 operand defined by the call we are about to inline.
8783 2014-02-17 Richard Biener <rguenther@suse.de>
8785 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8787 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8788 Ilya Tocar <ilya.tocar@intel.com>
8790 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8791 arguments order in builtin.
8792 (_mm512_permutexvar_epi64): Ditto.
8793 (_mm512_mask_permutexvar_epi64): Ditto
8794 (_mm512_maskz_permutexvar_epi32): Ditto
8795 (_mm512_permutexvar_epi32): Ditto
8796 (_mm512_mask_permutexvar_epi32): Ditto
8798 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8800 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8801 (p8_vmrgow): Likewise.
8803 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8805 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8808 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8811 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8812 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8813 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8814 using direct move instructions on ISA 2.07. Also adjust
8815 instruction length for 64-bit.
8816 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8817 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8819 2014-02-15 Alan Modra <amodra@gmail.com>
8823 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8824 find_replacement on parts of insn rtl that might be reloaded.
8826 2014-02-15 Richard Biener <rguenther@suse.de>
8828 PR tree-optimization/60183
8829 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8830 (tree_ssa_phiprop): Calculate and free post-dominators.
8832 2014-02-14 Jeff Law <law@redhat.com>
8834 PR rtl-optimization/60131
8835 * ree.c (get_extended_src_reg): New function.
8836 (combine_reaching_defs): Use it rather than assuming location of REG.
8837 (find_and_remove_re): Verify first operand of extension is
8838 a REG before adding the insns to the copy list.
8840 2014-02-14 Roland McGrath <mcgrathr@google.com>
8842 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8843 * configure: Regenerated.
8844 * config.in: Regenerated.
8845 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8846 instead of ASM_SHORT.
8848 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8849 Richard Earnshaw <rearnsha@arm.com>
8851 PR rtl-optimization/59535
8852 * lra-constraints.c (process_alt_operands): Encourage alternative
8853 when unassigned pseudo class is superset of the alternative class.
8854 (inherit_reload_reg): Don't inherit when optimizing for code size.
8855 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8856 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8857 modes not less than 4 for Thumb1.
8859 2014-02-14 Kyle McMartin <kyle@redhat.com>
8862 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8864 2014-02-14 Richard Biener <rguenther@suse.de>
8866 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8867 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8868 Do not strip INDIRECT_REFs.
8870 2014-02-14 Richard Biener <rguenther@suse.de>
8873 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8874 DECL_FUNCTION_SPECIFIC_TARGET.
8875 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8876 * tree-streamer-out.c (pack_ts_target_option): Remove.
8877 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8878 (write_ts_function_decl_tree_pointers): Do not stream
8879 DECL_FUNCTION_SPECIFIC_TARGET.
8880 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8881 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8882 (lto_input_ts_function_decl_tree_pointers): Do not stream
8883 DECL_FUNCTION_SPECIFIC_TARGET.
8885 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8887 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8888 (get_initial_def_for_induction, vectorizable_induction): Ignore
8889 debug stmts when looking for exit_phi.
8890 (vectorizable_live_operation): Fix up condition.
8892 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8894 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8895 nreverse() because it changes the content of original tree list.
8897 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8899 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8900 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8902 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8904 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8905 GNU coding standards.
8907 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8910 * dwarf2out.c (gen_subprogram_die): Don't call
8911 add_calling_convention_attribute if subr_die is old_die.
8913 2014-02-13 Sharad Singhai <singhai@google.com>
8915 * doc/optinfo.texi: Fix order of nodes.
8917 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8919 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8920 operands[2], not operands[3].
8922 2014-02-13 Richard Biener <rguenther@suse.de>
8925 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8926 mention the possibility of an in-tree build.
8927 (CLooG): Update recommended version to 0.18.1, mention the
8928 possibility of an in-tree build and clarify that the ISL
8929 bundled with CLooG does not work.
8931 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8934 * expr.c (compress_float_constant): If x is a hard register,
8935 extend into a pseudo and then move to x.
8937 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8939 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8940 caused by bad second argument to warning_at() with -mhotpatch and
8941 nested functions (e.g. with gfortran).
8943 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8945 * opts.c (option_name): Remove "enabled by default" rider.
8947 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8949 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8951 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8952 Uros Bizjak <ubizjak@gmail.com>
8955 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8956 * configure: Regenerated.
8958 2014-02-12 Richard Biener <rguenther@suse.de>
8960 * vec.c (vec_prefix::calculate_allocation): Move as
8961 inline variant to vec.h.
8962 (vec_prefix::calculate_allocation_1): New out-of-line version.
8963 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8964 (vec_prefix::m_has_auto_buf): Rename to ...
8965 (vec_prefix::m_using_auto_storage): ... this.
8966 (vec_prefix::calculate_allocation): Inline the easy cases
8967 and dispatch to calculate_allocation_1 which doesn't need the
8969 (va_heap::reserve): Use gcc_checking_assert.
8970 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8971 m_using_auto_storage.
8972 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8974 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8975 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8976 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8978 2014-02-12 Richard Biener <rguenther@suse.de>
8980 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8981 when we found a dependence.
8983 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8985 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8987 (maybe_fold_stmt): ... into this new function.
8988 * omp-low.c (lower_omp): Update comment.
8990 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8993 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8996 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8998 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8999 identifiers in comments.
9000 (cortexa53_extra_costs): Likewise.
9001 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9002 (cortexa7_extra_costs): Likewise.
9003 (cortexa12_extra_costs): Likewise.
9004 (cortexa15_extra_costs): Likewise.
9005 (v7m_extra_costs): Likewise.
9007 2014-02-12 Richard Biener <rguenther@suse.de>
9010 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9011 of posix_memalign being successful.
9012 (lower_stmt): Restrict lowering of posix_memalign to when
9013 -ftree-bit-ccp is enabled.
9015 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9017 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9019 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9021 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9023 PR rtl-optimization/60116
9024 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9025 other_insn once the combination has been validated.
9027 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9030 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9032 * ipa-devirt.c: Include demangle.h
9033 (odr_violation_reported): New static variable.
9034 (add_type_duplicate): Update odr_violations.
9035 (maybe_record_node): Add completep parameter; update it.
9036 (record_target_from_binfo): Add COMPLETEP parameter;
9037 update it as needed.
9038 (possible_polymorphic_call_targets_1): Likewise.
9039 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9040 rename FINAL to COMPLETE.
9041 (record_targets_from_bases): Sanity check we found the binfo;
9042 fix COMPLETEP updating.
9043 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9044 parameter, fix computing of COMPLETEP.
9045 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9046 at LTO time do demangling.
9047 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9048 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9050 (gimple_get_virt_method_for_binfo): Likewise.
9051 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9052 gimple_get_virt_method_for_vtable): Update prototypes.
9054 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9057 * genautomata.c (add_presence_absence): Fix typo with
9058 {final_}presence_list.
9060 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9063 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9064 for VSX/Altivec vectors that land in GPR registers.
9066 2014-02-11 Richard Henderson <rth@redhat.com>
9067 Jakub Jelinek <jakub@redhat.com>
9070 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9071 around drhs if type conversion to lacc->type is not useless.
9073 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9075 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9077 (cortex-a57.cortex-a53): Likewise.
9078 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9080 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9082 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9085 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9087 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9088 add_options_for_arm_vfp3.
9090 2014-02-11 Jeff Law <law@redhat.com>
9093 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9094 object with an undesirable mode.
9096 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9099 * config/i386/sol2-9.h: New file.
9100 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9101 *-*-solaris2.9*): Use it.
9103 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9105 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9106 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9108 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9110 * config/microblaze/microblaze.c: Extend mcpu version format
9112 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9114 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9116 2014-02-10 Richard Henderson <rth@redhat.com>
9119 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9120 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9121 ms-abi vs -mno-accumulate-outgoing-args.
9122 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9123 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9126 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9129 * cfgexpand.c (expand_asm_operands): Attach source location to
9130 ASM_INPUT rtx objects.
9131 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9133 2014-02-10 Nick Clifton <nickc@redhat.com>
9135 * config/mn10300/mn10300.c (popcount): New function.
9136 (mn10300_expand_prologue): Include saved registers in stack usage
9139 2014-02-10 Jeff Law <law@redhat.com>
9142 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9143 when changing the SET_DEST of a prior insn to avoid an input reload.
9145 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9147 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9148 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9149 -mcall-openbsd, or -mcall-linux.
9150 (CC1_ENDIAN_BIG_SPEC): Remove.
9151 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9152 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9153 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9154 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9155 and %cc1_endian_default.
9156 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9158 2014-02-10 Richard Biener <rguenther@suse.de>
9160 PR tree-optimization/60115
9161 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9162 MEM_REF handling. Properly verify that the accesses are not
9163 out of the objects bound.
9165 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9167 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9170 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9172 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9173 proper constants and fix formatting.
9174 (possible_polymorphic_call_targets): Fix formatting.
9176 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9177 Ilya Tocar <ilya.tocar@intel.com>
9179 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9180 (_mm512_loadu_epi32): Renamed into...
9181 (_mm512_loadu_si512): This.
9182 (_mm512_storeu_epi32): Renamed into...
9183 (_mm512_storeu_si512): This.
9184 (_mm512_maskz_ceil_ps): Removed.
9185 (_mm512_maskz_ceil_pd): Ditto.
9186 (_mm512_maskz_floor_ps): Ditto.
9187 (_mm512_maskz_floor_pd): Ditto.
9188 (_mm512_floor_round_ps): Ditto.
9189 (_mm512_floor_round_pd): Ditto.
9190 (_mm512_ceil_round_ps): Ditto.
9191 (_mm512_ceil_round_pd): Ditto.
9192 (_mm512_mask_floor_round_ps): Ditto.
9193 (_mm512_mask_floor_round_pd): Ditto.
9194 (_mm512_mask_ceil_round_ps): Ditto.
9195 (_mm512_mask_ceil_round_pd): Ditto.
9196 (_mm512_maskz_floor_round_ps): Ditto.
9197 (_mm512_maskz_floor_round_pd): Ditto.
9198 (_mm512_maskz_ceil_round_ps): Ditto.
9199 (_mm512_maskz_ceil_round_pd): Ditto.
9200 (_mm512_expand_pd): Ditto.
9201 (_mm512_expand_ps): Ditto.
9202 * config/i386/i386.c (ix86_builtins): Remove
9203 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9204 (bdesc_args): Ditto.
9205 * config/i386/predicates.md (const1256_operand): New.
9206 (const_1_to_2_operand): Ditto.
9207 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9208 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9209 (*avx512pf_gatherpf<mode>sf): Ditto.
9210 (avx512pf_gatherpf<mode>df): Ditto.
9211 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9212 (*avx512pf_gatherpf<mode>df): Ditto.
9213 (avx512pf_scatterpf<mode>sf): Ditto.
9214 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9215 (*avx512pf_scatterpf<mode>sf): Ditto.
9216 (avx512pf_scatterpf<mode>df): Ditto.
9217 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9218 (*avx512pf_scatterpf<mode>df): Ditto.
9219 (avx512f_expand<mode>): Removed.
9220 (<shift_insn><mode>3<mask_name>): Change predicate type.
9222 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9224 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9225 not at the end of datarefs vector use ordered_remove to avoid
9226 reordering datarefs vector.
9229 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9230 mark local addressable non-static vars as GOVD_PRIVATE
9231 instead of GOVD_LOCAL.
9232 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9233 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9237 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9238 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9239 assume_aligned or alloc_align attributes.
9240 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9241 arguments. Handle also assume_aligned and alloc_align attributes.
9242 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9243 calls to functions with assume_aligned or alloc_align attributes.
9244 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9246 2014-02-08 Terry Guo <terry.guo@arm.com>
9248 * doc/invoke.texi: Document ARM -march=armv7e-m.
9250 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9252 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9253 flag on __cilkrts_rethrow builtin.
9256 * ipa-cp.c (determine_versionability): Fail at -O0
9257 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9258 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9261 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9264 * tree-inline.c (copy_forbidden): Fail for
9265 __attribute__((optimize (0))) functions.
9267 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9269 * varpool.c: Include pointer-set.h.
9270 (varpool_remove_unreferenced_decls): Variables in other partitions
9271 will not be output; be however careful to not lose information
9274 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9276 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9277 lookup in the vtable constructor.
9279 2014-02-07 Jeff Law <law@redhat.com>
9282 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9283 define_insn_and_split.
9285 * ipa-inline.c (inline_small_functions): Fix typos.
9287 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9289 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9290 (s390_can_use_return_insn): Declare.
9291 * config/s390/s390.h (EPILOGUE_USES): Define.
9292 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9294 (s390_chunkify_start): Handle return JUMP_LABELs.
9295 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9296 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9297 (s390_can_use_return_insn): New functions.
9298 (s390_fix_long_loop_prediction): Handle conditional returns.
9299 (TARGET_SET_UP_BY_PROLOGUE): Define.
9300 * config/s390/s390.md (ANY_RETURN): New code iterator.
9301 (*creturn, *csimple_return, return, simple_return): New patterns.
9303 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9305 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9306 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9307 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9308 REG_CFA_RESTORE list when deciding not to restore a register.
9310 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9312 * config/s390/s390.c: Include tree-pass.h and context.h.
9313 (s390_early_mach): New function, split out from...
9314 (s390_emit_prologue): ...here.
9315 (pass_data_s390_early_mach): New pass structure.
9316 (pass_s390_early_mach): New class.
9317 (s390_option_override): Create and register early_mach pass.
9318 Move to end of file.
9320 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9322 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9323 to match for the exit block.
9325 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9327 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9328 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9329 Reject misaligned operands.
9331 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9333 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9335 2014-02-07 Richard Biener <rguenther@suse.de>
9338 * gimple-low.c (lower_builtin_posix_memalign): New function.
9339 (lower_stmt): Call it to lower posix_memalign in a way
9340 to make alignment info accessible.
9342 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9345 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9346 __builtin_setjmp_receiver.
9348 2014-02-07 Richard Biener <rguenther@suse.de>
9351 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9352 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9353 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9354 Handle BUILT_IN_POSIX_MEMALIGN.
9355 (find_func_clobbers): Likewise.
9356 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9357 (call_may_clobber_ref_p_1): Likewise.
9359 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9362 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9365 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9368 * lto-cgraph.c (lto_output_node): Use
9369 symtab_get_symbol_partitioning_class.
9370 (lto_output_varpool_node): likewise.
9371 (symtab_get_symbol_partitioning_class): Move here from
9373 * cgraph.h (symbol_partitioning_class): Likewise.
9374 (symtab_get_symbol_partitioning_class): Declare.
9376 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9378 * ggc.h (ggc_internal_cleared_alloc): New macro.
9379 * vec.h (vec_safe_copy): Handle memory stats.
9380 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9381 * target-globals.c (save_target_globals): Likewise.
9383 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9386 * expr.c (emit_move_resolve_push): Export; be bit more selective
9387 on when to clear alias set.
9388 * expr.h (emit_move_resolve_push): Declare.
9389 * function.h (struct function): Add tail_call_marked.
9390 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9391 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9392 * config/i386/i386.md (TImode move expander): De not call
9394 (FP push expanders): Preserve memory attributes.
9395 * config/i386/sse.md (push<mode>1): Remove.
9396 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9397 (ix86_expand_push): Remove.
9398 * config/i386/mmx.md (push<mode>1): Remove.
9400 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9402 PR rtl-optimization/60030
9403 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9404 lopart with paradoxical subreg before shifting it up by hprec.
9406 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9408 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9409 Remove extra newline at end of file.
9410 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9411 (arm_issue_rate): Handle cortexa57.
9412 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9413 (cortex-a57.cortex-a53): Likewise.
9415 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9418 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9419 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9421 (arm_expand_prologue): Adjust all callers.
9422 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9423 info, registers also at the lowest numbered registers side. Use
9424 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9428 * var-tracking.c (adjust_mems): Before adding a SET to
9429 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9431 2014-02-06 Alan Modra <amodra@gmail.com>
9434 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9435 change SDmode to DDmode when lra_in_progress.
9437 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9440 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9441 free_data_ref on the dr first, and before goto again also set dr
9442 to the next dr. For simd_lane_access, free old datarefs[i] before
9443 overwriting it. For get_vectype_for_scalar_type failure, don't
9444 free_data_ref if simd_lane_access.
9446 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9449 * tree.h (opts_for_fn): New inline function.
9450 (opt_for_fn): Define.
9451 * config/i386/i386.c (ix86_function_regparm): Use
9452 opt_for_fn (decl, optimize) instead of optimize.
9454 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9456 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9457 for SYMBOL_REF in large memory model.
9459 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9461 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9463 (cortex-a57): Likewise.
9464 (cortex-a57.cortex-a53): Likewise.
9466 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9467 Kugan Vivekanandarajah <kuganv@linaro.org>
9469 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9471 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9473 2014-02-06 Richard Biener <rguenther@suse.de>
9475 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9476 set_loop_copy and initialize_original_copy_tables.
9478 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9480 * config/aarch64/aarch64-simd.md
9481 (aarch64_ashr_simddi): Change QI to SI.
9483 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9484 Jakub Jelinek <jakub@redhat.com>
9487 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9490 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9492 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9493 CODE_FOR_altivec_vpku[hw]um to
9494 CODE_FOR_altivec_vpku[hw]um_direct.
9495 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9496 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9497 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9498 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9500 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9502 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9503 generation for -maltivec=be.
9504 (altivec_vsumsws): Simplify redundant test.
9506 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9508 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9509 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9510 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9511 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9512 gen_altivec_vpkuwum.
9513 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9515 (altivec_vpks<VI_char>ss): Likewise.
9516 (altivec_vpks<VI_char>us): Likewise.
9517 (altivec_vpku<VI_char>us): Likewise.
9518 (altivec_vpku<VI_char>um): Likewise.
9519 (altivec_vpku<VI_char>um_direct): New (copy of
9520 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9522 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9523 target is little endian and -maltivec=be is not specified.
9524 (*altivec_vupkhs<VU_char>_direct): New (copy of
9525 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9526 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9527 target is little endian and -maltivec=be is not specified.
9528 (*altivec_vupkls<VU_char>_direct): New (copy of
9529 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9530 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9531 little endian and -maltivec=be is not specified.
9532 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9533 little endian and -maltivec=be is not specified.
9535 2014-02-05 Richard Henderson <rth@redhat.com>
9538 * combine-stack-adj.c: Revert r206943.
9539 * sched-int.h (struct deps_desc): Add last_args_size.
9540 * sched-deps.c (init_deps): Initialize it.
9541 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9542 contain REG_ARGS_SIZE notes.
9544 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9546 * lto-cgraph.c (asm_nodes_output): Make global.
9547 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9548 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9549 (driver_handle_option): Handle OPT_fwpa.
9551 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9554 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9555 a comment typo and formatting issue. If odr_hash hasn't been
9556 created, return vNULL and set *completep to false.
9559 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9560 bb with no successors.
9562 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9565 * doc/invoke.texi (-march): Clarify documentation for ARM.
9569 2014-02-05 Richard Biener <rguenther@suse.de>
9571 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9572 when not vectorizing because of too many alias checks.
9573 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9574 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9576 2014-02-05 Nick Clifton <nickc@redhat.com>
9578 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9579 accept extended registers in any mode when compiling for the MN10300.
9581 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9583 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9584 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9585 sanitization attributes.
9586 (can_inline_edge_p): Likewise.
9587 (sanitize_attrs_match_for_inline_p): New function.
9589 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9591 * ipa-prop.c (detect_type_change): Shor circuit testing of
9592 type changes on THIS pointer.
9594 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9597 * config/pa/pa.c (legitimize_tls_address): Return original address
9598 if not passed a SYMBOL_REF rtx.
9599 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9601 (pa_emit_move_sequence): Simplify TLS source operands.
9602 (pa_legitimate_constant_p): Reject all TLS constants.
9603 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9604 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9606 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9608 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9609 groups when we know they are controlled by LTO.
9610 * varasm.c (default_binds_local_p_1): If object is in other partition,
9611 it will be resolved locally.
9613 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9615 * config/host-linux.c (linux_gt_pch_use_address): Don't
9616 use SSIZE_MAX because it is not always defined.
9618 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9621 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9622 threshold for pseudo splitting.
9623 (update_ebb_live_info): Process call argument hard registers and
9624 hard registers from insn definition too.
9625 (max_small_class_regs_num): New constant.
9626 (inherit_in_ebb): Update live hard regs through EBBs. Update
9627 reloads_num only for small register classes. Don't split for
9630 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9633 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9636 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9638 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9641 2014-02-04 Marek Polacek <polacek@redhat.com>
9643 * gdbinit.in (pel): Define.
9645 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9647 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9650 2014-02-04 Richard Biener <rguenther@suse.de>
9653 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9654 in function context local.
9655 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9656 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9657 similar to LTO_imported_decl_ref.
9659 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9661 PR tree-optimization/60002
9662 * cgraphclones.c (build_function_decl_skip_args): Clear
9665 PR tree-optimization/60023
9666 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9667 false to gsi_replace.
9668 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9669 has been in some EH region and vec_stmt could throw, add
9670 vec_stmt into the same EH region.
9671 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9672 has no lhs, ignore it.
9673 * internal-fn.c (expand_MASK_LOAD): Likewise.
9676 * tree-inline.c (copy_forbidden): Fail for
9677 __attribute__((optimize (0))) functions.
9680 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9681 is set, copy one less argument.
9682 (expand_simd_clones): Don't subtract clone_info->inbranch
9683 from simd_clone_struct_alloc argument.
9685 PR rtl-optimization/57915
9686 * recog.c (simplify_while_replacing): If all unary/binary/relational
9687 operation arguments are constant, attempt to simplify those.
9690 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9691 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9693 2014-02-04 Richard Biener <rguenther@suse.de>
9695 PR tree-optimization/60012
9696 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9697 TBAA disambiguation to all DDRs.
9699 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9702 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9703 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9705 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9708 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9710 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9712 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9713 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9715 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9718 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9719 to figure out targets of polymorphic calls with known decl.
9720 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9721 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9722 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9723 (get_polymorphic_call_info): ... here.
9724 (get_polymorphic_call_info_from_invariant): New function.
9726 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9728 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9729 lookup via vtable pointer; check for type consistency
9730 and turn inconsitent facts into UNREACHABLE.
9731 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9732 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9733 type inconsistent querries; return UNREACHABLE instead.
9735 2014-02-03 Richard Henderson <rth@twiddle.net>
9738 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9739 already processed this node.
9740 (normalize_one_pred_1): Pass along mark_set.
9741 (normalize_one_pred): Create and destroy a pointer_set_t.
9742 (normalize_one_pred_chain): Likewise.
9744 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9746 PR gcov-profile/58602
9747 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9749 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9752 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9753 -fno-devirtualize; try to devirtualize by the knowledge of
9754 virtual table pointer given by aggregate propagation.
9755 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9756 (ipa_print_node_jump_functions): Dump also offset that
9757 is relevant for polymorphic calls.
9758 (determine_known_aggregate_parts): Add arg_type parameter; use it
9759 instead of determining the type from pointer type.
9760 (ipa_compute_jump_functions_for_edge): Update call of
9761 determine_known_aggregate_parts.
9762 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9763 (gimple_get_virt_method_for_binfo): ... here; simplify using
9764 vtable_pointer_value_to_vtable.
9765 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9766 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9767 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9768 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9770 (vtable_pointer_value_to_binfo): ... here.
9771 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9773 2014-02-03 Teresa Johnson <tejohnson@google.com>
9775 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9776 redef of outer loop index variable.
9778 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9782 * doc/extend.texi (Function Attributes): Typo.
9784 2014-02-03 Cong Hou <congh@google.com>
9786 PR tree-optimization/60000
9787 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9788 if the vectorized statement is a store. A store statement can only
9789 appear at the end of pattern statements.
9791 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9793 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9794 (ix86_option_override_internal): Default long double to 64-bit for
9795 32-bit Bionic and to 128-bit for 64-bit Bionic.
9797 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9798 TARGET_LONG_DOUBLE_128 is true.
9799 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9801 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9802 (mlong-double-64): Negate -mlong-double-128.
9803 (mlong-double-128): New option.
9805 * config/i386/i386-c.c (ix86_target_macros): Define
9806 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9808 * doc/invoke.texi: Document -mlong-double-128.
9810 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9812 PR rtl-optimization/60024
9813 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9815 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9817 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9819 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9821 PR rtl-optimization/57662
9822 * sel-sched.c (code_motion_path_driver): Do not mark already not
9823 existing blocks in the visiting bitmap.
9825 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9827 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9828 on the insn being emitted.
9830 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9831 Will Deacon <will.deacon@arm.com>
9833 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9835 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9837 * config/arm/arm-tables.opt: Regenerate.
9839 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9841 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9842 for vector types other than V16QImode.
9843 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9844 define_expand, and call altivec_expand_vec_perm_le when producing
9845 code with little endian element order.
9846 (*altivec_vperm_<mode>_internal): New insn having previous
9847 behavior of altivec_vperm_<mode>.
9848 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9849 altivec_expand_vec_perm_le when producing code with little endian
9851 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9852 behavior of altivec_vperm_<mode>_uns.
9854 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9856 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9857 (altivec_vsumsws): Add handling for -maltivec=be with a little
9859 (altivec_vsumsws_direct): New.
9860 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9861 gen_altivec_vsumsws.
9863 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9865 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9866 vtable_pointer_value_to_binfo): New functions.
9867 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9868 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9870 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9872 * config/nios2/nios2.md (load_got_register): Initialize GOT
9873 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9874 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9876 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9878 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9879 preserverd by passthrough, do not propagate the type.
9881 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9883 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9884 (mips_atomic_assign_expand_fenv): New function.
9885 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9887 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9889 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9890 (__builtin_mips_set_fcsr): Likewise.
9891 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9892 MIPS_USI_FTYPE_VOID.
9893 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9894 (mips16_expand_set_fcsr): Likewise.
9895 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9896 (mips16_set_fcsr_stub): Likewise.
9897 (mips16_get_fcsr_one_only_stub): New class.
9898 (mips16_set_fcsr_one_only_stub): Likewise.
9899 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9900 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9901 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9902 (hard_float): New availability predicate.
9903 (mips_builtins): Add get_fcsr and set_fcsr.
9904 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9905 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9906 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9907 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9908 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9911 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9913 * config/mips/mips.c (mips_one_only_stub): New class.
9914 (mips_need_mips16_rdhwr_p): Replace with...
9915 (mips16_rdhwr_stub): ...this new variable.
9916 (mips16_stub_call_address): New function.
9917 (mips16_rdhwr_one_only_stub): New class.
9918 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9919 (mips_output_mips16_rdhwr): Delete.
9920 (mips_finish_stub): New function.
9921 (mips_code_end): Use it to handle rdhwr stubs.
9923 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9926 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9927 when calculating size of integer atomic types.
9929 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9931 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9933 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9935 PR tree-optimization/60003
9936 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9937 * profile.c (branch_prob): Use gimple_call_builtin_p
9938 to check for BUILT_IN_SETJMP_RECEIVER.
9939 * tree-inline.c (copy_bb): Call notice_special_calls.
9941 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9944 * lra-constraints.c (process_alt_operands): Update reload_sum only
9947 2014-01-31 Richard Henderson <rth@redhat.com>
9950 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9951 until after else_eh is processed.
9953 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9955 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9956 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9957 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9958 in smmintrin.h, remove them.
9959 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9960 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9961 * config/i386/i386.md (ROUND_SAE): Fix value.
9962 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9963 (const48_operand): New.
9964 * config/i386/subst.md (round), (round_expand): Use
9965 const_4_or_8_to_11_operand.
9966 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9968 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9970 * config/i386/constraints.md (Yk): Swap meaning with k.
9971 * config/i386/i386.md (movhi_internal): Change Yk to k.
9972 (movqi_internal): Ditto.
9973 (*k<logic><mode>): Ditto.
9976 (kandn<mode>): Ditto.
9977 (*<code>hi_1): Ditto.
9978 (*<code>qi_1): Ditto.
9979 (kxnor<mode>): Ditto.
9980 (kortestzhi): Ditto.
9981 (kortestchi): Ditto.
9983 (*one_cmplhi2_1): Ditto.
9984 (*one_cmplqi2_1): Ditto.
9985 * config/i386/sse.md (): Change k to Yk.
9986 (avx512f_load<mode>_mask): Ditto.
9987 (avx512f_blendm<mode>): Ditto.
9988 (avx512f_store<mode>_mask): Ditto.
9989 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9990 (avx512f_storedqu<mode>_mask): Ditto.
9991 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9993 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9994 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9995 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9996 (avx512f_maskcmp<mode>3): Ditto.
9997 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9998 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9999 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10000 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10001 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10002 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10003 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10004 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10005 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10006 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10007 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10008 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10009 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10010 (vec_extract_lo_<mode>_maskm): Ditto.
10011 (vec_extract_hi_<mode>_maskm): Ditto.
10012 (avx512f_vternlog<mode>_mask): Ditto.
10013 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10014 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10015 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10016 (avx512f_<code>v8div16qi2_mask): Ditto.
10017 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10018 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10019 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10020 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10021 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10022 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10023 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10024 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10025 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10026 (avx512cd_maskb_vec_dupv8di): Ditto.
10027 (avx512cd_maskw_vec_dupv16si): Ditto.
10028 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10029 (avx512f_vpermi2var<mode>3_mask): Ditto.
10030 (avx512f_vpermi2var<mode>3_mask): Ditto.
10031 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10032 (*avx512f_gathersi<mode>): Ditto.
10033 (*avx512f_gathersi<mode>_2): Ditto.
10034 (*avx512f_gatherdi<mode>): Ditto.
10035 (*avx512f_gatherdi<mode>_2): Ditto.
10036 (*avx512f_scattersi<mode>): Ditto.
10037 (*avx512f_scatterdi<mode>): Ditto.
10038 (avx512f_compress<mode>_mask): Ditto.
10039 (avx512f_compressstore<mode>_mask): Ditto.
10040 (avx512f_expand<mode>_mask): Ditto.
10041 * config/i386/subst.md (mask): Change k to Yk.
10042 (mask_scalar_merge): Ditto.
10045 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10047 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10049 2014-01-31 Richard Biener <rguenther@suse.de>
10051 PR middle-end/59990
10052 * builtins.c (fold_builtin_memory_op): Make sure to not
10053 use a floating-point mode or a boolean or enumeral type for
10054 the copy operation.
10056 2014-01-30 DJ Delorie <dj@redhat.com>
10058 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10059 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10060 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10061 whenever main() has an epilogue.
10063 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10065 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10066 unused variable "field".
10067 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10068 (vsx_mergeh_<mode>): Likewise.
10069 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10070 (altivec_vmrghh): Likewise.
10071 (altivec_vmrghw): Likewise.
10072 (altivec_vmrglb): Likewise.
10073 (altivec_vmrglh): Likewise.
10074 (altivec_vmrglw): Likewise.
10075 (altivec_vspltb): Add missing uses.
10076 (altivec_vsplth): Likewise.
10077 (altivec_vspltw): Likewise.
10078 (altivec_vspltsf): Likewise.
10080 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10083 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10084 frame related instructions.
10086 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10088 PR rtl-optimization/59959
10089 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10090 any reload of register whose subreg is invalid.
10092 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10094 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10095 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10096 Add missing return type - void.
10098 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10100 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10101 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10102 remove element index adjustment for endian (now handled in vsx.md
10104 (altivec_expand_vec_perm_const): Use
10105 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10106 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10107 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10108 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10109 define_expand and a new define_insn *altivec_vspltb_internal;
10110 adjust for -maltivec=be on a little endian target.
10111 (altivec_vspltb_direct): New.
10112 (altivec_vsplth): Divide into a define_expand and a new
10113 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10114 little endian target.
10115 (altivec_vsplth_direct): New.
10116 (altivec_vspltw): Divide into a define_expand and a new
10117 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10118 little endian target.
10119 (altivec_vspltw_direct): New.
10120 (altivec_vspltsf): Divide into a define_expand and a new
10121 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10122 a little endian target.
10124 2014-01-30 Richard Biener <rguenther@suse.de>
10126 PR tree-optimization/59993
10127 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10128 can propagate form the earlier stmt and avoid the transform
10129 when the intermediate result is needed.
10131 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10133 * README.Portability: Fix typo.
10135 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10137 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10138 comparison_operator with ordered_comparison_operator.
10140 2014-01-30 Nick Clifton <nickc@redhat.com>
10142 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10143 Rename to mn10300_store_multiple_regs.
10144 * config/mn10300/mn10300.c: Likewise.
10145 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10146 store_multiple_regs.
10147 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10148 Call mn10300_store_multiple_regs.
10150 2014-01-30 Nick Clifton <nickc@redhat.com>
10151 DJ Delorie <dj@redhat.com>
10153 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10154 %fp 2 to keep registers after it properly word-aligned.
10155 (rl78_alloc_physical_registers_umul): Handle the case where both
10156 input operands are the same.
10158 2014-01-30 Richard Biener <rguenther@suse.de>
10160 PR tree-optimization/59903
10161 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10164 2014-01-30 Jason Merrill <jason@redhat.com>
10167 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10170 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10172 2014-01-30 Richard Biener <rguenther@suse.de>
10174 PR tree-optimization/59951
10175 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10177 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10180 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10181 SFmode to DFmode case.
10183 2014-01-29 DJ Delorie <dj@redhat.com>
10185 * config/msp430/msp430.opt (-minrt): New.
10186 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10188 (ENDFILE_SPEC): Likewise.
10190 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10192 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10193 (estimate_function_body_sizes): Use it.
10195 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10198 * dwarf2out.c (is_cxx_auto): New.
10199 (is_base_type): Use it.
10200 (gen_type_die_with_usage): Likewise.
10202 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10204 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10205 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10206 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10207 -maltivec=be with LE targets.
10208 (vsx_mergeh_<mode>): Likewise.
10209 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10210 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10211 (altivec_vmrghb): Replace with define_expand and new
10212 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10213 (altivec_vmrghb_direct): New define_insn.
10214 (altivec_vmrghh): Replace with define_expand and new
10215 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10216 (altivec_vmrghh_direct): New define_insn.
10217 (altivec_vmrghw): Replace with define_expand and new
10218 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10219 (altivec_vmrghw_direct): New define_insn.
10220 (*altivec_vmrghsf): Adjust for endianness.
10221 (altivec_vmrglb): Replace with define_expand and new
10222 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10223 (altivec_vmrglb_direct): New define_insn.
10224 (altivec_vmrglh): Replace with define_expand and new
10225 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10226 (altivec_vmrglh_direct): New define_insn.
10227 (altivec_vmrglw): Replace with define_expand and new
10228 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10229 (altivec_vmrglw_direct): New define_insn.
10230 (*altivec_vmrglsf): Adjust for endianness.
10231 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10232 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10233 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10234 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10235 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10236 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10237 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10238 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10240 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10242 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10243 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10246 2014-01-29 Richard Biener <rguenther@suse.de>
10248 PR tree-optimization/58742
10249 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10250 associate_pointerplus_align.
10251 (associate_pointerplus_diff): New function.
10252 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10253 and associate_pointerplus_diff.
10255 2014-01-29 Richard Biener <rguenther@suse.de>
10257 * lto-streamer.h (LTO_major_version): Bump to 3.
10258 (LTO_minor_version): Reset to 0.
10260 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10262 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10263 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10264 (arm_file_start): Generate correct asm header for armv7ve.
10265 * config/arm/bpabi.h: Add multilib support for armv7ve.
10266 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10267 and cortex-a15 to armv7ve.
10268 * config/arm/t-aprofile: Add multilib support for armv7ve.
10269 * doc/invoke.texi: Document -march=armv7ve.
10271 2014-01-29 Richard Biener <rguenther@suse.de>
10273 PR tree-optimization/58742
10274 * tree-ssa-forwprop.c (associate_plusminus): Return true
10275 if we changed sth, defer EH cleanup to ...
10276 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10277 (simplify_mult): New function.
10279 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10281 PR middle-end/59917
10282 PR tree-optimization/59920
10283 * tree.c (build_common_builtin_nodes): Remove
10284 __builtin_setjmp_dispatcher initialization.
10285 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10286 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10287 instead of gsi_after_labels + manually skipping debug stmts.
10288 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10289 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10290 * tree-inline.c (copy_edges_for_bb): Remove
10291 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10292 argument. Ignore computed_goto_p stmts. Don't call
10293 make_abnormal_goto_edges. If a call might need abnormal edges
10294 for non-local gotos, see if it already has an edge to
10295 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10296 with true argument, don't do anything then, otherwise add
10297 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10298 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10300 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10301 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10302 (lower_stmt): Don't set data->calls_builtin_setjmp.
10303 (lower_builtin_setjmp): Adjust comment.
10304 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10305 * tree-cfg.c (found_computed_goto): Remove.
10306 (factor_computed_gotos): Remove.
10307 (make_goto_expr_edges): Return bool, true for computed gotos.
10308 Don't call make_abnormal_goto_edges.
10309 (build_gimple_cfg): Don't set found_computed_goto, don't call
10310 factor_computed_gotos.
10311 (computed_goto_p): No longer static.
10312 (make_blocks): Don't set found_computed_goto.
10313 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10314 (make_edges): If make_goto_expr_edges returns true, push bb
10315 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10316 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10317 vector. Record mapping between bbs and OpenMP regions if there
10318 are any, adjust make_gimple_omp_edges caller. Call
10319 handle_abnormal_edges.
10320 (make_abnormal_goto_edges): Remove.
10321 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10322 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10323 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10324 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10325 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10326 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10327 filling *region also set *region_idx to (*region)->entry->index.
10330 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10331 For REGs set ORIGINAL_REGNO.
10333 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10335 * doc/md.texi: Mention that a target shouldn't implement
10336 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10339 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10341 PR tree-optimization/59594
10342 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10343 a copy of the datarefs vector rather than the vector itself.
10345 2014-01-28 Jason Merrill <jason@redhat.com>
10348 * dwarf2out.c (auto_die): New static.
10349 (gen_type_die_with_usage): Handle C++1y 'auto'.
10350 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10353 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10356 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10357 (SPEC_X32): Likewise.
10358 (SPEC_64): Likewise.
10359 * config/i386/i386.c (ix86_option_override_internal): Turn off
10360 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10362 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10363 * config/i386/i386.h (TARGET_16BIT): New macro.
10364 (TARGET_16BIT_P): Likewise.
10365 * config/i386/i386.opt: Add m16.
10366 * doc/invoke.texi: Document -m16.
10368 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10370 PR preprocessor/59935
10371 * input.c (location_get_source_line): Bail out on when line number
10372 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10374 2014-01-28 Richard Biener <rguenther@suse.de>
10376 PR tree-optimization/58742
10377 * tree-ssa-forwprop.c (associate_plusminus): Handle
10378 pointer subtraction of the form (T)(P + A) - (T)P.
10380 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10382 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10385 2014-01-28 Richard Biener <rguenther@suse.de>
10388 2014-01-28 Richard Biener <rguenther@suse.de>
10390 PR rtl-optimization/45364
10391 PR rtl-optimization/59890
10392 * var-tracking.c (local_get_addr_clear_given_value): Handle
10393 already cleared slot.
10394 (val_reset): Handle not allocated local_get_addr_cache.
10395 (vt_find_locations): Use post-order on the inverted CFG.
10397 2014-01-28 Richard Biener <rguenther@suse.de>
10399 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10401 2014-01-28 Richard Biener <rguenther@suse.de>
10403 PR rtl-optimization/45364
10404 PR rtl-optimization/59890
10405 * var-tracking.c (local_get_addr_clear_given_value): Handle
10406 already cleared slot.
10407 (val_reset): Handle not allocated local_get_addr_cache.
10408 (vt_find_locations): Use post-order on the inverted CFG.
10410 2014-01-28 Alan Modra <amodra@gmail.com>
10412 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10413 * configure.ac <recursive call for build != host>: Define
10414 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10415 and LD_FOR_BUILD too.
10416 * configure: Regenerate.
10418 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10420 * config/i386/i386.c (get_builtin_code_for_version): Separate
10421 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10422 Broadwell from Haswell.
10424 2014-01-27 Steve Ellcey <sellcey@mips.com>
10426 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10427 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10428 * config/mips/mips.c (mips_option_override): Change setting
10430 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10431 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10432 Change from Mask to Var.
10434 2014-01-27 Jeff Law <law@redhat.com>
10436 * ipa-inline.c (inline_small_functions): Fix typo.
10438 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10440 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10441 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10442 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10443 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10444 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10445 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10446 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10447 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10448 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10449 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10450 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10451 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10452 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10453 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10454 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10455 (_mm512_storeu_epi64): Ditto.
10456 (_mm512_cmpge_epi32_mask): Ditto.
10457 (_mm512_cmpge_epu32_mask): Ditto.
10458 (_mm512_cmpge_epi64_mask): Ditto.
10459 (_mm512_cmpge_epu64_mask): Ditto.
10460 (_mm512_cmple_epi32_mask): Ditto.
10461 (_mm512_cmple_epu32_mask): Ditto.
10462 (_mm512_cmple_epi64_mask): Ditto.
10463 (_mm512_cmple_epu64_mask): Ditto.
10464 (_mm512_cmplt_epi32_mask): Ditto.
10465 (_mm512_cmplt_epu32_mask): Ditto.
10466 (_mm512_cmplt_epi64_mask): Ditto.
10467 (_mm512_cmplt_epu64_mask): Ditto.
10468 (_mm512_cmpneq_epi32_mask): Ditto.
10469 (_mm512_cmpneq_epu32_mask): Ditto.
10470 (_mm512_cmpneq_epi64_mask): Ditto.
10471 (_mm512_cmpneq_epu64_mask): Ditto.
10472 (_mm512_expand_pd): Ditto.
10473 (_mm512_expand_ps): Ditto.
10474 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10475 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10476 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10477 * config/i386/i386.c (ix86_builtins): Add
10478 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10479 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10480 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10481 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10482 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10483 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10484 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10485 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10486 IX86_BUILTIN_PMOVUSQW512_MEM.
10487 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10488 __builtin_ia32_pmovsqd512mem_mask,
10489 __builtin_ia32_pmovqd512mem_mask,
10490 __builtin_ia32_pmovusqw512mem_mask,
10491 __builtin_ia32_pmovsqw512mem_mask,
10492 __builtin_ia32_pmovqw512mem_mask,
10493 __builtin_ia32_pmovusdw512mem_mask,
10494 __builtin_ia32_pmovsdw512mem_mask,
10495 __builtin_ia32_pmovdw512mem_mask,
10496 __builtin_ia32_pmovqb512mem_mask,
10497 __builtin_ia32_pmovusqb512mem_mask,
10498 __builtin_ia32_pmovsqb512mem_mask,
10499 __builtin_ia32_pmovusdb512mem_mask,
10500 __builtin_ia32_pmovsdb512mem_mask,
10501 __builtin_ia32_pmovdb512mem_mask.
10502 (bdesc_args): Add __builtin_ia32_expanddf512,
10503 __builtin_ia32_expandsf512.
10504 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10505 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10506 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10507 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10508 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10509 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10510 (avx512f_<code>v8div16qi2_mask_store): This.
10511 (avx512f_expand<mode>): New.
10513 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10515 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10517 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10518 (_mm512_prefetch_i32scatter_pd): Ditto.
10519 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10520 (_mm512_prefetch_i64scatter_pd): Ditto.
10521 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10522 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10523 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10524 (_mm512_prefetch_i32scatter_ps): Ditto.
10525 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10526 (_mm512_prefetch_i64scatter_ps): Ditto.
10527 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10528 * config/i386/i386-builtin-types.def: Define
10529 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10530 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10531 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10532 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10533 IX86_BUILTIN_SCATTERPFQPD.
10534 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10535 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10536 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10537 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10538 __builtin_ia32_scatterpfqps.
10539 (ix86_expand_builtin): Expand new built-ins.
10540 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10541 fix memory access data type.
10542 (*avx512pf_gatherpf<mode>_mask): Ditto.
10543 (*avx512pf_gatherpf<mode>): Ditto.
10544 (avx512pf_scatterpf<mode>): Ditto.
10545 (*avx512pf_scatterpf<mode>_mask): Ditto.
10546 (*avx512pf_scatterpf<mode>): Ditto.
10547 (GATHER_SCATTER_SF_MEM_MODE): New.
10548 (avx512pf_gatherpf<mode>df): Ditto.
10549 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10550 (*avx512pf_scatterpf<mode>df): Ditto.
10552 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10555 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10556 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10559 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10561 * common/config/arm/arm-common.c
10562 (arm_rewrite_mcpu): Handle multiple names.
10564 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10566 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10568 * gimple-builder.h (create_gimple_tmp): Delete.
10570 2014-01-27 Christian Bruel <christian.bruel@st.com>
10572 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10575 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10577 * config/pa/pa.md (call): Generate indirect long calls to non-local
10578 functions when outputing 32-bit code.
10579 (call_value): Likewise except for special call to buggy powf function.
10581 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10582 portable runtime and PIC indirect calls.
10583 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10584 and PIC call sequences. Use ldo instead of blr to set return register
10585 in PIC call sequence.
10587 2014-01-25 Walter Lee <walt@tilera.com>
10589 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10590 avoid clobbering a live register.
10592 2014-01-25 Walter Lee <walt@tilera.com>
10594 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10595 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10596 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10597 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10599 2014-01-25 Walter Lee <walt@tilera.com>
10601 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10602 arguments on even registers.
10603 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10605 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10606 (BIGGEST_ALIGNMENT): Ditto.
10607 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10609 2014-01-25 Walter Lee <walt@tilera.com>
10611 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10612 insns before bundling.
10613 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10615 2014-01-25 Walter Lee <walt@tilera.com>
10617 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10618 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10619 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10621 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10623 * config/mips/constraints.md (kl): Delete.
10624 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10625 define expands, using...
10626 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10627 instructions for MIPS16.
10628 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10629 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10631 2014-01-25 Walter Lee <walt@tilera.com>
10633 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10637 2014-01-25 Walter Lee <walt@tilera.com>
10639 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10640 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10642 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10644 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10647 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10649 * print-rtl.c (in_call_function_usage): New var.
10650 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10651 EXPR_LIST mode as mode and not as reg note name.
10653 PR middle-end/59561
10654 * cfgloopmanip.c (copy_loop_info): If
10655 loop->warned_aggressive_loop_optimizations, make sure
10656 the flag is set in target loop too.
10658 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10660 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10662 * builtins.def: Likewise.
10663 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10664 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10665 * ira.c (ira_setup_eliminable_regset): Likewise.
10666 * omp-low.c (gate_expand_omp): Likewise.
10667 (execute_lower_omp): Likewise.
10668 (diagnose_sb_0): Likewise.
10669 (gate_diagnose_omp_blocks): Likewise.
10670 (simd_clone_clauses_extract): Likewise.
10673 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10675 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10676 correction for little endian...
10677 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10680 2014-01-24 Jeff Law <law@redhat.com>
10682 PR tree-optimization/59919
10683 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10684 for non-returning calls.
10686 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10688 * common/config/aarch64/aarch64-common.c
10689 (aarch64_rewrite_mcpu): Handle multiple names.
10690 * config/aarch64/aarch64.h
10691 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10693 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10695 * input.c (add_file_to_cache_tab): Handle the case where fopen
10698 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10701 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10702 from push operand if code of push isn't PRE_DEC.
10704 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10707 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10708 -mquad-memory-atomic. Update -mquad-memory documentation to say
10709 it is only used for non-atomic loads/stores.
10711 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10712 -mquad-memory or -mquad-memory-atomic switches.
10714 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10715 -mquad-memory-atomic to ISA 2.07 support.
10717 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10718 to separate support of normal quad word memory operations (ldq, stq)
10719 from the atomic quad word memory operations.
10721 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10722 support to separate non-atomic quad word operations from atomic
10723 quad word operations. Disable non-atomic quad word operations in
10724 little endian mode so that we don't have to swap words after the
10725 load and before the store.
10726 (quad_load_store_p): Add comment about atomic quad word support.
10727 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10728 options printed with -mdebug=reg.
10730 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10731 -mquad-memory-atomic as the test for whether we have quad word
10732 atomic instructions.
10733 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10734 or -mp8-vector are used, allow byte/half-word atomic operations.
10736 * config/rs6000/sync.md (load_lockedti): Insure that the address
10737 is a proper indexed or indirect address for the lqarx instruction.
10738 On little endian systems, swap the hi/lo registers after the lqarx
10740 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10741 insure the address is valid for the lqarx instruction.
10742 (store_conditionalti): Insure that the address is a proper indexed
10743 or indirect address for the stqcrx. instruction. On little endian
10744 systems, swap the hi/lo registers before doing the stqcrx.
10746 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10747 insure the address is valid for the stqcrx. instruction.
10749 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10750 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10751 type of quad memory support is available.
10753 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10755 PR regression/59915
10756 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10757 there is a danger of looping.
10759 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10761 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10762 force flag_ira_loop_pressure if set via command line.
10764 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10766 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10767 (ashr_simd): New builtin handling DI mode.
10768 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10769 (aarch64_sshr_simddi): New match pattern.
10770 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10771 (vshrd_n_s64): Likewise.
10772 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10774 2014-01-23 Nick Clifton <nickc@redhat.com>
10776 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10777 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10778 favour of mcu specific scripts.
10779 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10782 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10783 Alex Velenko <Alex.Velenko@arm.com>
10785 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10786 (vaddv_s16): Likewise.
10787 (vaddv_s32): Likewise.
10788 (vaddv_u8): Likewise.
10789 (vaddv_u16): Likewise.
10790 (vaddv_u32): Likewise.
10791 (vaddvq_s8): Likewise.
10792 (vaddvq_s16): Likewise.
10793 (vaddvq_s32): Likewise.
10794 (vaddvq_s64): Likewise.
10795 (vaddvq_u8): Likewise.
10796 (vaddvq_u16): Likewise.
10797 (vaddvq_u32): Likewise.
10798 (vaddvq_u64): Likewise.
10799 (vaddv_f32): Likewise.
10800 (vaddvq_f32): Likewise.
10801 (vaddvq_f64): Likewise.
10802 (vmaxv_f32): Likewise.
10803 (vmaxv_s8): Likewise.
10804 (vmaxv_s16): Likewise.
10805 (vmaxv_s32): Likewise.
10806 (vmaxv_u8): Likewise.
10807 (vmaxv_u16): Likewise.
10808 (vmaxv_u32): Likewise.
10809 (vmaxvq_f32): Likewise.
10810 (vmaxvq_f64): Likewise.
10811 (vmaxvq_s8): Likewise.
10812 (vmaxvq_s16): Likewise.
10813 (vmaxvq_s32): Likewise.
10814 (vmaxvq_u8): Likewise.
10815 (vmaxvq_u16): Likewise.
10816 (vmaxvq_u32): Likewise.
10817 (vmaxnmv_f32): Likewise.
10818 (vmaxnmvq_f32): Likewise.
10819 (vmaxnmvq_f64): Likewise.
10820 (vminv_f32): Likewise.
10821 (vminv_s8): Likewise.
10822 (vminv_s16): Likewise.
10823 (vminv_s32): Likewise.
10824 (vminv_u8): Likewise.
10825 (vminv_u16): Likewise.
10826 (vminv_u32): Likewise.
10827 (vminvq_f32): Likewise.
10828 (vminvq_f64): Likewise.
10829 (vminvq_s8): Likewise.
10830 (vminvq_s16): Likewise.
10831 (vminvq_s32): Likewise.
10832 (vminvq_u8): Likewise.
10833 (vminvq_u16): Likewise.
10834 (vminvq_u32): Likewise.
10835 (vminnmv_f32): Likewise.
10836 (vminnmvq_f32): Likewise.
10837 (vminnmvq_f64): Likewise.
10839 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10841 * config/aarch64/aarch64-simd.md
10842 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10843 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10844 (*aarch64_mul3_elt<mode>): Likewise.
10845 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10846 (*aarch64_mul3_elt_to_64v2df): Likewise.
10847 (*aarch64_mla_elt<mode>): Likewise.
10848 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10849 (*aarch64_mls_elt<mode>): Likewise.
10850 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10851 (*aarch64_fma4_elt<mode>): Likewise.
10852 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10853 (*aarch64_fma4_elt_to_64v2df): Likewise.
10854 (*aarch64_fnma4_elt<mode>): Likewise.
10855 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10856 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10857 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10858 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10859 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10860 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10861 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10862 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10863 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10865 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10867 * config/aarch64/aarch64-simd.md
10868 (aarch64_be_checked_get_lane<mode>): New define_expand.
10869 * config/aarch64/aarch64-simd-builtins.def
10870 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10871 New builtin definition.
10872 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10873 Use new safe be builtin.
10875 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10877 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10879 (aarch64_be_st1<mode>): Likewise.
10880 (aarch_ld1<VALL:mode>): Define_expand modified.
10881 (aarch_st1<VALL:mode>): Likewise.
10882 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10883 (UNSPEC_ST1): Likewise.
10885 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10887 * config/microblaze/microblaze.md: Add trap insn and attribute
10889 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10891 PR preprocessor/58580
10892 * input.h (location_get_source_line): Take an additional line_size
10894 (void diagnostics_file_cache_fini): Declare new function.
10895 * input.c (struct fcache): New type.
10896 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10897 New static constants.
10898 (diagnostic_file_cache_init, total_lines_num)
10899 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10900 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10901 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10902 (get_next_line, read_next_line, goto_next_line, read_line_num):
10903 New static function definitions.
10904 (diagnostic_file_cache_fini): New function.
10905 (location_get_source_line): Take an additional output line_len
10906 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10908 * diagnostic.c (diagnostic_finish): Call
10909 diagnostic_file_cache_fini.
10910 (adjust_line): Take an additional input parameter for the length
10911 of the line, rather than calculating it with strlen.
10912 (diagnostic_show_locus): Adjust the use of
10913 location_get_source_line and adjust_line with respect to their new
10914 signature. While displaying a line now, do not stop at the first
10915 null byte. Rather, display the zero byte as a space and keep
10916 going until we reach the size of the line.
10917 * Makefile.in: Add vec.o to OBJS-libcommon
10919 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10920 Ilya Tocar <ilya.tocar@intel.com>
10922 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10923 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10924 (__builtin_ia32_kmov16): Ditto.
10925 * config/i386/i386.md (UNSPEC_KMOV): New.
10928 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10930 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10931 (_mm512_storeu_si512): Ditto.
10933 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10936 * rtl.h (get_referenced_operands): Declare.
10937 * recog.c (get_referenced_operands): New function.
10938 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10939 operands have been referenced when recording LO_SUM references.
10941 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10943 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10945 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10947 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10948 Enable for generic and recent AMD targets.
10950 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10952 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10953 ARG_SIZE note when adjustment was eliminated.
10955 2014-01-22 Jeff Law <law@redhat.com>
10957 PR tree-optimization/59597
10958 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10959 in file. Accept new argument REGISTERING and use it to modify
10960 dump output appropriately.
10961 (register_jump_thread): Corresponding changes.
10962 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10963 thread paths involving joiner blocks. Add code to dump cancelled
10964 jump threading paths.
10966 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10968 PR rtl-optimization/59477
10969 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10970 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10972 2014-01-22 Tom Tromey <tromey@redhat.com>
10974 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10976 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10978 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10980 PR rtl-optimization/59896
10981 * lra-constraints.c (process_alt_operands): Check unused note for
10982 matched operands of insn with no output reloads.
10984 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10986 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10987 (mips_move_from_gpr_cost): Likewise.
10989 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10991 PR rtl-optimization/59858
10992 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10993 ira_class_hard_regs_num.
10994 (process_alt_operands): Increase reject for dying matched operand.
10996 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10999 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11000 smaller than size, perform several stores or loads and stores
11001 at dst + count - size to store or copy all of size bytes, rather
11002 than just last modesize bytes.
11004 2014-01-20 DJ Delorie <dj@redhat.com>
11006 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11007 that CLOBBERs are REGs before propogating their values.
11009 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11011 PR middle-end/59789
11012 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11013 (cgraph_inline_failed_type): New function.
11014 * cgraph.h (DEFCIFCODE): Add type.
11015 (cgraph_inline_failed_type_t): New enum.
11016 (cgraph_inline_failed_type): New prototype.
11017 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11018 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11019 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11020 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11021 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11022 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11023 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11024 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11025 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11026 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11027 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11028 OPTIMIZATION_MISMATCH.
11029 * tree-inline.c (expand_call_inline): Emit errors during
11030 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11032 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11035 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11036 mode attribute in insn output.
11038 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11040 * output.h (output_constant): Delete.
11041 * varasm.c (output_constant): Make private.
11043 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11045 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11047 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11049 PR middle-end/59860
11050 * tree.h (fold_builtin_strcat): New prototype.
11051 * builtins.c (fold_builtin_strcat): No longer static. Add len
11052 argument, if non-NULL, don't call c_strlen. Optimize
11053 directly into __builtin_memcpy instead of __builtin_strcpy.
11054 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11055 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11057 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11059 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11060 for SImode_address_operand operands, having only a REG argument.
11062 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11064 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11065 loader name using mbig-endian.
11066 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11068 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11070 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11071 (-mtune): Likewise.
11074 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11076 * config/aarch64/aarch64-protos.h
11077 (aarch64_cannot_change_mode_class_ptr): Declare.
11078 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11079 aarch64_cannot_change_mode_class_ptr): New.
11080 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11081 backend hook aarch64_cannot_change_mode_class.
11083 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11085 * common/config/aarch64/aarch64-common.c
11086 (aarch64_handle_option): Don't handle any option order logic here.
11087 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11088 selected_cpu, warn on architecture version mismatch.
11089 (aarch64_override_options): Fix parsing order for option strings.
11091 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11092 Iain Sandoe <iain@codesourcery.com>
11095 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11096 warning. Amend comment to reflect current functionality.
11098 2014-01-20 Richard Biener <rguenther@suse.de>
11100 PR middle-end/59860
11101 * builtins.c (fold_builtin_strcat): Remove case better handled
11102 by tree-ssa-strlen.c.
11104 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11106 * config/aarch64/aarch64.opt
11107 (mcpu, march, mtune): Make case-insensitive.
11109 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11112 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11113 if operands[1] is a REG or ZERO_EXTEND of a REG.
11115 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11117 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11119 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11121 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11122 long non-pic millicode calls.
11124 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11126 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11128 2014-01-19 Kito Cheng <kito@0xlab.org>
11130 * builtins.c (expand_movstr): Check movstr expand done or fail.
11132 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11133 H.J. Lu <hongjiu.lu@intel.com>
11136 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11137 to DImode for zero-extended addresses.
11139 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11141 PR rtl-optimization/57763
11142 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11143 on the new indirect jump_insn and increment LABEL_NUSES (label).
11145 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11149 * config.gcc (x86_archs): New variable.
11150 (x86_64_archs): Likewise.
11151 (x86_cpus): Likewise.
11152 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11153 --with-arch/--with-cpu= options.
11154 Support --with-arch=/--with-cpu={nehalem,westmere,
11155 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11157 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11159 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11160 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11162 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11164 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11166 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11169 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11170 clear cpp_get_options (parse_in)->warn_unused_macros for
11171 ix86_target_macros_internal with cpp_define.
11173 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11175 * jump.c (delete_related_insns): Keep (use (insn))s.
11176 * reorg.c (redundant_insn): Check for barriers too.
11178 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11180 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11182 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11184 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11185 call to $$dyncall when TARGET_LONG_CALLS is true.
11187 2014-01-17 Jeff Law <law@redhat.com>
11189 * ree.c (combine_set_extension): Temporarily disable test for
11190 changing number of hard registers.
11192 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11194 PR middle-end/58125
11195 * ipa-inline-analysis.c (inline_free_summary):
11196 Do not free summary of aliases.
11198 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11200 PR middle-end/59706
11201 * gimplify.c (gimplify_expr): Use create_tmp_var
11202 instead of create_tmp_var_raw. If cond doesn't have
11203 integral type, don't add the IFN_ANNOTATE builtin at all.
11205 2014-01-17 Martin Jambor <mjambor@suse.cz>
11208 * ipa-cp.c (prev_edge_clone): New variable.
11209 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11210 Also resize prev_edge_clone vector.
11211 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11212 (ipcp_edge_removal_hook): New function.
11213 (ipcp_driver): Register ipcp_edge_removal_hook.
11215 2014-01-17 Andrew Pinski <apinski@cavium.com>
11216 Steve Ellcey <sellcey@mips.com>
11219 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11222 2014-01-17 Jeff Law <law@redhat.com>
11224 PR middle-end/57904
11225 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11226 so that pass_ccp runs first.
11228 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11230 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11231 (ix86_adjust_cost): Use !TARGET_XXX.
11232 (do_reorder_for_imul): Likewise.
11233 (swap_top_of_ready_list): Likewise.
11234 (ix86_sched_reorder): Likewise.
11236 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11238 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11239 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11240 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11241 (intel_memset): New. Duplicate slm_memset.
11242 (intel_cost): New. Duplicate slm_cost.
11243 (m_INTEL): New macro.
11244 (processor_target_table): Add "intel".
11245 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11246 with PROCESSOR_INTEL for "intel".
11247 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11248 PROCESSOR_SILVERMONT.
11249 (ix86_issue_rate): Likewise.
11250 (ix86_adjust_cost): Likewise.
11251 (ia32_multipass_dfa_lookahead): Likewise.
11252 (swap_top_of_ready_list): Likewise.
11253 (ix86_sched_reorder): Likewise.
11254 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11255 instead of TARGET_OPT_AGU.
11256 * config/i386/i386.h (TARGET_INTEL): New.
11257 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11258 (processor_type): Add PROCESSOR_INTEL.
11259 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11260 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11262 2014-01-17 Marek Polacek <polacek@redhat.com>
11265 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11268 2014-01-17 Richard Biener <rguenther@suse.de>
11270 PR tree-optimization/46590
11271 * opts.c (default_options_table): Add entries for
11272 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11273 all enabled at -O1 but not for -Og.
11274 * common.opt (fbranch-count-reg): Remove Init(1).
11275 (fmove-loop-invariants): Likewise.
11276 (ftree-pta): Likewise.
11278 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11280 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11281 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11282 decls to at least the GCC 4.8 used alignments.
11285 * tree-nested.c (convert_nonlocal_reference_stmt,
11286 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11287 of GIMPLE_BIND stmts, adjust associated decls.
11289 2014-01-17 Richard Biener <rguenther@suse.de>
11291 PR tree-optimization/46590
11292 * vec.h (vec<>::bseach): New member function implementing
11293 binary search according to C89 bsearch.
11294 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11295 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11296 bitmap pointer again. Make accesses_in_loop a flat array.
11297 (mem_ref_obstack): New global.
11298 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11299 (mark_ref_stored): Likewise.
11300 (ref_indep_loop_p_2): Likewise.
11301 (set_ref_stored_in_loop): New helper function.
11302 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11303 (memref_free): Adjust.
11304 (record_mem_ref_loc): Simplify.
11305 (gather_mem_refs_stmt): Adjust.
11306 (sort_locs_in_loop_postorder_cmp): New function.
11307 (analyze_memory_references): Sort accesses_in_loop after
11308 loop postorder number.
11309 (find_ref_loc_in_loop_cmp): New function.
11310 (for_all_locs_in_loop): Find relevant cluster of locs in
11311 accesses_in_loop and iterate without recursion.
11312 (execute_sm): Avoid uninit warning.
11313 (struct ref_always_accessed): Simplify.
11314 (ref_always_accessed::operator ()): Likewise.
11315 (ref_always_accessed_p): Likewise.
11316 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11317 loop postorder numbers here.
11318 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11321 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11324 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11325 on decls for which assemble_alias has been called.
11327 2014-01-17 Nick Clifton <nickc@redhat.com>
11329 * config/msp430/msp430.opt: (mcpu): New option.
11330 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11331 (msp430_option_override): Parse target_cpu. If the MCU name
11332 matches a generic string, clear target_mcu.
11333 (msp430_attr): Allow numeric interrupt values up to 63.
11334 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11335 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11337 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11339 * config/msp430/msp430.md (popm): Use %J rather than %I.
11340 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11341 (addhi_cy_i): Use immediate_operand for operand 2.
11342 * doc/invoke.texi: Document -mcpu option.
11344 2014-01-17 Richard Biener <rguenther@suse.de>
11346 PR rtl-optimization/38518
11347 * df.h (df_analyze_loop): Declare.
11348 * df-core.c: Include cfgloop.h.
11349 (df_analyze_1): Split out main part of df_analyze.
11350 (df_analyze): Adjust.
11351 (loop_inverted_post_order_compute): New function.
11352 (loop_post_order_compute): Likewise.
11353 (df_analyze_loop): New function avoiding whole-function
11354 postorder computes.
11355 * loop-invariant.c (find_defs): Use df_analyze_loop.
11356 (find_invariants): Adjust.
11357 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11359 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11361 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11362 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11364 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11366 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11367 traversal when removing references.
11369 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11372 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11374 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11376 PR middle-end/56791
11377 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11378 pushing a reload for an autoinc when we had previously reloaded an
11379 inner part of the address.
11381 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11383 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11385 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11386 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11387 when not giving up or versioning for alias only because of
11389 (vect_analyze_data_ref_dependences): Set to true.
11390 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11392 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11393 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11396 PR middle-end/58344
11397 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11400 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11401 operand 0 predicate for gathers, use a new pseudo as subtarget.
11403 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11405 PR middle-end/59609
11406 * lra-constraints.c (process_alt_operands): Add printing debug info.
11407 Check absence of input/output reloads for matched operands too.
11409 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11411 PR rtl-optimization/59835
11412 * ira.c (ira_init_register_move_cost): Increase cost for
11415 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11417 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11419 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11422 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11423 non-register objects. Use gen_(high/low)part more consistently.
11426 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11429 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11430 endian support, remove tests for WORDS_BIG_ENDIAN.
11431 (p8_mfvsrd_3_<mode>): Likewise.
11432 (reload_gpr_from_vsx<mode>): Likewise.
11433 (reload_gpr_from_vsxsf): Likewise.
11434 (p8_mfvsrd_4_disf): Likewise.
11436 2014-01-16 Richard Biener <rguenther@suse.de>
11438 PR rtl-optimization/46590
11439 * lcm.c (compute_antinout_edge): Use postorder iteration.
11440 (compute_laterin): Use inverted postorder iteration.
11442 2014-01-16 Nick Clifton <nickc@redhat.com>
11444 PR middle-end/28865
11445 * varasm.c (output_constant): Return the number of bytes actually
11447 (output_constructor_array_range): Update the field size with the
11448 number of bytes emitted by output_constant.
11449 (output_constructor_regular_field): Likewise. Also do not
11450 complain if the total number of bytes emitted is now greater
11451 than the expected fieldpos.
11452 * output.h (output_constant): Update prototype and descriptive comment.
11454 2014-01-16 Marek Polacek <polacek@redhat.com>
11456 PR middle-end/59827
11457 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11458 it is error_mark_node.
11460 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11462 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11463 VALID_AVX256_REG_OR_OI_MODE.
11465 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11467 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11468 current procedure should be profiled.
11470 2014-01-15 Andrew Pinski <apinski@cavium.com>
11472 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11473 of moving from/to the STACK_REG register class.
11475 2014-01-15 Richard Henderson <rth@redhat.com>
11478 * reginfo.c (global_regs_decl): Globalize.
11479 * rtl.h (global_regs_decl): Declare.
11480 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11481 reserved via global_regs.
11483 2014-01-15 Teresa Johnson <tejohnson@google.com>
11485 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11487 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11489 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11490 and vmulosh rather than call gen_vec_widen_smult_*.
11491 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11492 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11493 (vec_widen_smult_even_v16qi): Likewise.
11494 (vec_widen_umult_even_v8hi): Likewise.
11495 (vec_widen_smult_even_v8hi): Likewise.
11496 (vec_widen_umult_odd_v16qi): Likewise.
11497 (vec_widen_smult_odd_v16qi): Likewise.
11498 (vec_widen_umult_odd_v8hi): Likewise.
11499 (vec_widen_smult_odd_v8hi): Likewise.
11500 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11501 vmuloub rather than call gen_vec_widen_umult_*.
11502 (vec_widen_umult_lo_v16qi): Likewise.
11503 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11504 vmulosb rather than call gen_vec_widen_smult_*.
11505 (vec_widen_smult_lo_v16qi): Likewise.
11506 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11507 rather than call gen_vec_widen_umult_*.
11508 (vec_widen_umult_lo_v8hi): Likewise.
11509 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11510 rather than call gen_vec_widen_smult_*.
11511 (vec_widen_smult_lo_v8hi): Likewise.
11513 2014-01-15 Jeff Law <law@redhat.com>
11515 PR tree-optimization/59747
11516 * ree.c (find_and_remove_re): Properly handle case where a second
11517 eliminated extension requires widening a copy created for elimination
11518 of a prior extension.
11519 (combine_set_extension): Ensure that the number of hard regs needed
11520 for a destination register does not change when we widen it.
11522 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11524 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11525 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11526 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11527 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11528 (avr-*-rtems*): Likewise.
11529 (bfin*-rtems*): Likewise.
11530 (moxie-*-rtems*): Likewise.
11531 (h8300-*-rtems*): Likewise.
11532 (i[34567]86-*-rtems*): Likewise.
11533 (lm32-*-rtems*): Likewise.
11534 (m32r-*-rtems*): Likewise.
11535 (m68k-*-rtems*): Likewise.
11536 (microblaze*-*-rtems*): Likewise.
11537 (mips*-*-rtems*): Likewise.
11538 (powerpc-*-rtems*): Likewise.
11539 (sh-*-rtems*): Likewise.
11540 (sparc-*-rtems*): Likewise.
11541 (sparc64-*-rtems*): Likewise.
11542 (v850-*-rtems*): Likewise.
11543 (m32c-*-rtems*): Likewise.
11545 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11547 PR rtl-optimization/59511
11548 * ira.c (ira_init_register_move_cost): Use memory costs for some
11549 cases of register move cost calculations.
11550 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11551 instead of BB frequency.
11552 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11553 * lra-assigns.c (find_hard_regno_for): Ditto.
11555 2014-01-15 Richard Biener <rguenther@suse.de>
11557 PR tree-optimization/59822
11558 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11559 (vectorizable_load): Use it to hoist defs of uses of invariant
11560 loads out of the loop.
11562 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11563 Kugan Vivekanandarajah <kuganv@linaro.org>
11566 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11569 2014-01-15 Richard Biener <rguenther@suse.de>
11571 PR rtl-optimization/59802
11572 * lcm.c (compute_available): Use inverted postorder to seed
11573 the initial worklist.
11575 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11578 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11579 ADDR_REGS for invalid symrefs in non-PIC code.
11581 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11584 * builtins.c (determine_block_size): Initialize *probable_max_size
11585 even if len_rtx is CONST_INT.
11587 2014-01-14 Andrew Pinski <apinski@cavium.com>
11589 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11590 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11591 (cortexa53_tunings): Likewise.
11592 (aarch64_sched_issue_rate): New function.
11593 (TARGET_SCHED_ISSUE_RATE): Define.
11595 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11597 * ira-costs.c (find_costs_and_classes): Add missed
11598 ira_init_register_move_cost_if_necessary.
11600 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11603 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11605 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11608 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11609 to indicate if type is used for function return value. Warn ABI
11610 change if the vector mode isn't available for function return value.
11611 (ix86_function_arg_advance): Pass false to type_natural_mode.
11612 (ix86_function_arg): Likewise.
11613 (ix86_gimplify_va_arg): Likewise.
11614 (function_arg_32): Don't warn ABI change.
11615 (ix86_function_value): Pass true to type_natural_mode.
11616 (ix86_return_in_memory): Likewise.
11617 (ix86_struct_value_rtx): Removed.
11618 (TARGET_STRUCT_VALUE_RTX): Likewise.
11620 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11622 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11623 converting a conditional jump into a conditional return.
11625 2014-01-14 Richard Biener <rguenther@suse.de>
11627 PR tree-optimization/58921
11628 PR tree-optimization/59006
11629 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11630 hoisting invariant stmts.
11631 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11632 invariant loads on the preheader edge if possible.
11634 2014-01-14 Joey Ye <joey.ye@arm.com>
11636 * doc/plugin.texi (Building GCC plugins): Update to C++.
11638 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11640 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11641 (_mm_rcp28_round_ss): Ditto.
11642 (_mm_rsqrt28_round_sd): Ditto.
11643 (_mm_rsqrt28_round_ss): Ditto.
11644 (_mm_rcp28_sd): Ditto.
11645 (_mm_rcp28_ss): Ditto.
11646 (_mm_rsqrt28_sd): Ditto.
11647 (_mm_rsqrt28_ss): Ditto.
11648 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11649 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11650 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11651 (IX86_BUILTIN_RCP28SD): Ditto.
11652 (IX86_BUILTIN_RCP28SS): Ditto.
11653 (IX86_BUILTIN_RSQRT28SD): Ditto.
11654 (IX86_BUILTIN_RSQRT28SS): Ditto.
11655 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11656 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11657 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11658 (ix86_expand_special_args_builtin): Expand new FTYPE.
11659 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11660 (srcp14<mode>): Make insn unary.
11661 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11662 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11663 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11664 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11665 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11666 Fix rounding: make it SAE only.
11667 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11669 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11671 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11672 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11673 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11674 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11675 (round_saeonly_mask_scalar_operand4): Ditto.
11676 (round_saeonly_mask_scalar_op3): Ditto.
11677 (round_saeonly_mask_scalar_op4): Ditto.
11679 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11681 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11682 Implement -maltivec=be for vec_insert and vec_extract.
11684 2014-01-10 DJ Delorie <dj@redhat.com>
11686 * config/msp430/msp430.md (call_internal): Don't allow memory
11687 references with SP as the base register.
11688 (call_value_internal): Likewise.
11689 * config/msp430/constraints.md (Yc): New. For memory references
11690 that don't use SP as a base register.
11692 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11693 "an integer without a # prefix"
11694 * config/msp430/msp430.md (epilogue_helper): Use it.
11696 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11699 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11700 AVX512F gather builtins.
11701 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11702 on gather decls with INTEGER_TYPE masktype.
11703 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11704 directly into the builtin rather than hoisting it before loop.
11706 PR tree-optimization/59387
11707 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11708 (scev_const_prop): If folded_casts and type has undefined overflow,
11709 use force_gimple_operand instead of force_gimple_operand_gsi and
11710 for each added stmt if it is assign with
11711 arith_code_with_undefined_signed_overflow, call
11712 rewrite_to_defined_overflow.
11713 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11714 gimple-fold.h instead.
11715 (arith_code_with_undefined_signed_overflow,
11716 rewrite_to_defined_overflow): Moved to ...
11717 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11718 rewrite_to_defined_overflow): ... here. No longer static.
11719 Include gimplify-me.h.
11720 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11721 rewrite_to_defined_overflow): New prototypes.
11723 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11725 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11727 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11729 * builtins.c (get_object_alignment_2): Minor tweak.
11730 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11732 2014-01-13 Christian Bruel <christian.bruel@st.com>
11734 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11735 optimized non constant lengths.
11737 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11740 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11741 load as __atomic_load_N if possible.
11743 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11745 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11747 (rs6000_expand_builtin): Adjust call.
11749 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11752 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11753 * config/rs6000/rs6000.c: Include target-globals.h.
11754 (rs6000_set_current_function): Instead of doing target_reinit
11755 unconditionally, use save_target_globals_default_opts and
11756 restore_target_globals.
11758 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11760 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11761 (rs6000_expand_builtin): Handle mffs and mtfsf.
11762 (rs6000_init_builtins): Define mffs and mtfsf.
11763 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11764 (rs6000_mffs): New pattern.
11765 (rs6000_mtfsf): New pattern.
11767 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11769 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11770 Start narrowing with START. Apply candidate-use pair
11771 and check overall cost in narrowing.
11772 (iv_ca_prune): Pass new argument.
11774 2014-01-10 Jeff Law <law@redhat.com>
11776 PR middle-end/59743
11777 * ree.c (combine_reaching_defs): Ensure the defining statement
11778 occurs before the extension when optimizing extensions with
11779 different source and destination hard registers.
11781 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11784 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11785 vtables into the type inheritance graph.
11787 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11789 PR rtl-optimization/59754
11790 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11791 modes in the REGNO != REGNO case.
11793 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11795 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11797 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11799 PR tree-optimization/59745
11800 * tree-predcom.c (tree_predictive_commoning_loop): Call
11801 free_affine_expand_cache if giving up because components is NULL.
11803 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11804 GC in payload of target_globals struct instead of allocating them on
11805 the heap and the larger structs separately using GC.
11806 * target-globals.h (struct target_globals): Make regs, hard_regs,
11807 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11808 of GTY((skip)) and change type to void *.
11809 (reset_target_globals): Cast loads from those fields to corresponding
11812 2014-01-10 Steve Ellcey <sellcey@mips.com>
11815 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11816 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11817 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11819 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11822 * aarch64-modes.def (CC_Zmode): New flags mode.
11823 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11824 represents an equality.
11825 (aarch64_get_condition_code): Handle CC_Zmode.
11826 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11828 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11830 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11831 extraction in good case.
11833 2014-01-10 Richard Biener <rguenther@suse.de>
11835 PR tree-optimization/59374
11836 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11837 checking after SLP discovery. Mark stmts not participating
11838 in any SLP instance properly.
11840 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11842 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11843 when handling a SET rtx.
11845 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11847 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11848 (cortex-a57): Likewise.
11849 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11851 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11853 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11854 non-iwmmxt builtins.
11856 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11860 * ipa-devirt.c record_target_from_binfo): Take as argument
11861 stack of binfos and lookup matching one for virtual inheritance.
11862 (possible_polymorphic_call_targets_1): Update.
11864 2014-01-10 Huacai Chen <chenhc@lemote.com>
11866 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11867 kernel strings for Loongson-2E/2F/3A.
11869 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11871 PR middle-end/59670
11872 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11873 is_gimple_call before calling gimple_call_internal_p.
11875 2014-01-09 Steve Ellcey <sellcey@mips.com>
11877 * Makefile.in (TREE_FLOW_H): Remove.
11878 (TREE_SSA_H): Add file names from tree-flow.h.
11879 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11880 * tree.h: Remove tree-flow.h reference.
11881 * hash-table.h: Remove tree-flow.h reference.
11882 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11883 reference with tree-ssa-loop.h.
11885 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11887 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11888 default element-order behavior for -maltivec.
11889 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11890 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11891 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11892 when targeting big endian, at least for now.
11893 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11895 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11897 PR middle-end/47735
11898 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11899 var satisfies use_register_for_decl, just take into account type
11900 alignment, rather than decl alignment.
11902 PR tree-optimization/59622
11903 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11904 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11905 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11906 Don't devirtualize for inplace at all. For targets.length () == 1,
11907 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11909 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11911 * config/i386/i386.md (cpu): Remove the unused btver1.
11913 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11915 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11917 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11920 * tree-core.h (struct target_globals): New forward declaration.
11921 (struct tree_target_option): Add globals field.
11922 * tree.h (TREE_TARGET_GLOBALS): Define.
11923 (prepare_target_option_nodes_for_pch): New prototype.
11924 * target-globals.h (struct target_globals): Define even if
11925 !SWITCHABLE_TARGET.
11926 * tree.c (prepare_target_option_node_for_pch,
11927 prepare_target_option_nodes_for_pch): New functions.
11928 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11929 * config/i386/i386.c: Include target-globals.h.
11930 (ix86_set_current_function): Instead of doing target_reinit
11931 unconditionally, use save_target_globals_default_opts and
11932 restore_target_globals.
11934 2014-01-09 Richard Biener <rguenther@suse.de>
11936 PR tree-optimization/59715
11937 * tree-cfg.h (split_critical_edges): Declare.
11938 * tree-cfg.c (split_critical_edges): Export.
11939 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11941 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11943 * cfgexpand.c (expand_stack_vars): Optionally disable
11944 asan stack protection.
11945 (expand_used_vars): Likewise.
11946 (partition_stack_vars): Likewise.
11947 * asan.c (asan_emit_stack_protection): Optionally disable
11948 after return stack usage.
11949 (instrument_derefs): Optionally disable memory access instrumentation.
11950 (instrument_builtin_call): Likewise.
11951 (instrument_strlen_call): Likewise.
11952 (asan_protect_global): Optionally disable global variables protection.
11953 * doc/invoke.texi: Added doc for new options.
11954 * params.def: Added new options.
11955 * params.h: Likewise.
11957 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11959 PR rtl-optimization/59724
11960 * ifcvt.c (cond_exec_process_if_block): Don't call
11961 flow_find_head_matching_sequence with 0 longest_match.
11962 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11963 non-active insns if !stop_after.
11964 (try_head_merge_bb): Revert 2014-01-07 changes.
11966 2014-01-08 Jeff Law <law@redhat.com>
11968 * ree.c (get_sub_rtx): New function, extracted from...
11969 (merge_def_and_ext): Here.
11970 (combine_reaching_defs): Use get_sub_rtx.
11972 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11974 * cgraph.h (varpool_variable_node): Do not choke on null node.
11976 2014-01-08 Catherine Moore <clm@codesourcery.com>
11978 * config/mips/mips.md (simple_return): Attempt to use JRC
11980 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11982 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11984 PR rtl-optimization/59137
11985 * reorg.c (steal_delay_list_from_target): Call update_block for
11987 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11989 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11991 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11992 two duplicate entries.
11994 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11997 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11999 * config/mips/mips.c (mips_truncated_op_cost): New function.
12000 (mips_rtx_costs): Adjust test for BADDU.
12001 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12003 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12005 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12006 (*baddu_si): ...this new pattern.
12008 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12011 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12013 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12015 PR middle-end/57748
12016 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12018 (expand_expr, expand_normal): Adjust.
12019 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12020 inner_reference_p. Use inner_reference_p to expand inner references.
12021 (store_expr): Adjust.
12022 * cfgexpand.c (expand_call_stmt): Adjust.
12024 2014-01-08 Rong Xu <xur@google.com>
12026 * gcov-io.c (gcov_var): Move from gcov-io.h.
12027 (gcov_position): Ditto.
12028 (gcov_is_error): Ditto.
12029 (gcov_rewrite): Ditto.
12030 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12031 only part to libgcc/libgcov.h.
12033 2014-01-08 Marek Polacek <polacek@redhat.com>
12035 PR middle-end/59669
12036 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12038 2014-01-08 Marek Polacek <polacek@redhat.com>
12041 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12043 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12045 PR rtl-optimization/59649
12046 * stor-layout.c (get_mode_bounds): For BImode return
12047 0 and STORE_FLAG_VALUE.
12049 2014-01-08 Richard Biener <rguenther@suse.de>
12051 PR middle-end/59630
12052 * gimple.h (is_gimple_builtin_call): Remove.
12053 (gimple_builtin_call_types_compatible_p): New.
12054 (gimple_call_builtin_p): New overload.
12055 * gimple.c (is_gimple_builtin_call): Remove.
12056 (validate_call): Rename to ...
12057 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12058 check return types.
12059 (validate_type): New static function.
12060 (gimple_call_builtin_p): New overload and adjust.
12061 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12062 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12063 (gimple_fold_stmt_to_constant_1): Likewise.
12064 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12066 2014-01-08 Richard Biener <rguenther@suse.de>
12068 PR middle-end/59471
12069 * gimplify.c (gimplify_expr): Gimplify register-register type
12070 VIEW_CONVERT_EXPRs to separate stmts.
12072 2014-01-07 Jeff Law <law@redhat.com>
12074 PR middle-end/53623
12075 * ree.c (combine_set_extension): Handle case where source
12076 and destination registers in an extension insn are different.
12077 (combine_reaching_defs): Allow source and destination registers
12078 in extension to be different under limited circumstances.
12079 (add_removable_extension): Remove restriction that the
12080 source and destination registers in the extension are the same.
12081 (find_and_remove_re): Emit a copy from the extension's
12082 destination to its source after the defining insn if
12083 the source and destination registers are different.
12085 PR middle-end/59285
12086 * ifcvt.c (merge_if_block): If we are merging a block with more than
12087 one successor with a block with no successors, remove any BARRIER
12088 after the second block.
12090 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12092 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12094 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12097 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12098 for 14-bit register offsets when INT14_OK_STRICT is false.
12100 2014-01-07 Roland Stigge <stigge@antcom.de>
12101 Michael Meissner <meissner@linux.vnet.ibm.com>
12104 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12105 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12107 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12109 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12112 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12114 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12115 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12116 rtx is const0_rtx or not.
12118 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12121 * target-globals.c (save_target_globals): Remove this_fn_optab
12123 * toplev.c: Include optabs.h.
12124 (target_reinit): Temporarily restore the global options if another
12125 set of options are in force.
12127 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12129 PR rtl-optimization/58668
12130 * cfgcleanup.c (flow_find_cross_jump): Don't count
12131 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12132 to determine what is counted.
12133 (flow_find_head_matching_sequence): Use active_insn_p to determine
12135 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12137 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12138 determine what is counted.
12140 PR tree-optimization/59643
12141 * tree-predcom.c (split_data_refs_to_components): If one dr is
12142 read and one write, determine_offset fails and the write isn't
12143 in the bad component, just put the read into the bad component.
12145 2014-01-07 Mike Stump <mikestump@comcast.net>
12146 Jakub Jelinek <jakub@redhat.com>
12149 * tree-core.h (struct tree_optimization_option): Change optabs
12150 type from unsigned char * to void *.
12151 * optabs.c (init_tree_optimization_optabs): Adjust
12152 TREE_OPTIMIZATION_OPTABS initialization.
12154 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12157 * config/i386/i386.h (struct machine_function): Add
12158 no_drap_save_restore field.
12159 * config/i386/i386.c (ix86_save_reg): Use
12160 !cfun->machine->no_drap_save_restore instead of
12161 crtl->stack_realign_needed.
12162 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12163 this function clears frame_pointer_needed. Set
12164 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12165 and DRAP reg is needed.
12167 2014-01-06 Marek Polacek <polacek@redhat.com>
12170 * doc/implement-c.texi: Mention that other integer types are
12171 permitted as bit-field types in strictly conforming mode.
12173 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12175 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12176 is newly allocated.
12178 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12180 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12182 2014-01-06 Martin Jambor <mjambor@suse.cz>
12185 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12187 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12189 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12193 * var-tracking.c (add_stores): Preserve the value of the source even if
12194 we don't record the store.
12196 2014-01-06 Terry Guo <terry.guo@arm.com>
12198 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12200 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12203 * config/darwin.c (darwin_function_section): Adjust return values to
12204 correspond to optimisation changes made in r206070.
12206 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12208 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12209 from prefetch_block tune setting.
12210 (nocona_cost): Correct size of prefetch block to 64.
12212 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12214 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12215 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12216 used to save the static chain register in the computation of the offset
12217 from which the FP registers need to be restored.
12219 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12221 PR tree-optimization/59519
12222 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12223 ICE if get_current_def (current_new_name) is already non-NULL, as long
12224 as it is a phi result of some other phi in *new_exit_bb that has
12227 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12228 or vmovdqu* for misaligned_operand.
12229 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12230 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12231 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12232 aligned_mem for AVX512F masked aligned load and store builtins and for
12233 non-temporal moves.
12235 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12237 PR tree-optimization/59651
12238 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12239 Address range for negative step should be added by TYPE_SIZE_UNIT.
12241 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12243 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12244 overlapping registers also for registers other than the stack pointer.
12246 2014-01-03 Marek Polacek <polacek@redhat.com>
12249 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12252 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12255 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12258 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12259 (push splitter): Use <P:MODE_SIZE> instead of
12260 GET_MODE_SIZE (<P:MODE>mode).
12261 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12262 (mov -1, reg peephole2): Likewise.
12263 * config/i386/sse.md (*mov<mode>_internal,
12264 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12265 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12266 *<code><mode>3, *andnot<mode>3<mask_name>,
12267 <mask_codefor><code><mode>3<mask_name>): Likewise.
12268 * config/i386/subst.md (mask_mode512bit_condition,
12269 sd_mask_mode512bit_condition): Likewise.
12271 2014-01-02 Xinliang David Li <davidxl@google.com>
12273 PR tree-optimization/59303
12274 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12275 (dump_predicates): Better output format.
12276 (pred_equal_p): New function.
12277 (is_neq_relop_p): Ditto.
12278 (is_neq_zero_form_p): Ditto.
12279 (pred_expr_equal_p): Ditto.
12280 (pred_neg_p): Ditto.
12281 (simplify_pred): Ditto.
12282 (simplify_preds_2): Ditto.
12283 (simplify_preds_3): Ditto.
12284 (simplify_preds_4): Ditto.
12285 (simplify_preds): Ditto.
12286 (push_pred): Ditto.
12287 (push_to_worklist): Ditto.
12288 (get_pred_info_from_cmp): Ditto.
12289 (is_degenerated_phi): Ditto.
12290 (normalize_one_pred_1): Ditto.
12291 (normalize_one_pred): Ditto.
12292 (normalize_one_pred_chain): Ditto.
12293 (normalize_preds): Ditto.
12294 (normalize_cond_1): Remove function.
12295 (normalize_cond): Ditto.
12296 (is_gcond_subset_of): Ditto.
12297 (is_subset_of_any): Ditto.
12298 (is_or_set_subset_of): Ditto.
12299 (is_and_set_subset_of): Ditto.
12300 (is_norm_cond_subset_of): Ditto.
12301 (pred_chain_length_cmp): Ditto.
12302 (convert_control_dep_chain_into_preds): Type change.
12303 (find_predicates): Ditto.
12304 (find_def_preds): Ditto.
12305 (destroy_predicates_vecs): Ditto.
12306 (find_matching_predicates_in_rest_chains): Ditto.
12307 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12308 (is_pred_expr_subset): Ditto.
12309 (is_pred_chain_subset_of): Ditto.
12310 (is_included_in): Ditto.
12311 (is_superset_of): Ditto.
12313 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12315 Update copyright years.
12317 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12319 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12320 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12321 config/arc/arc.md, config/arc/arc.opt,
12322 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12323 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12324 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12325 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12326 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12327 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12328 vtable-verify.c, vtable-verify.h: Use the standard form for the
12331 2014-01-02 Tobias Burnus <burnus@net-b.de>
12333 * gcc.c (process_command): Update copyright notice dates.
12334 * gcov-dump.c: Ditto.
12336 * doc/cpp.texi: Bump @copying's copyright year.
12337 * doc/cppinternals.texi: Ditto.
12338 * doc/gcc.texi: Ditto.
12339 * doc/gccint.texi: Ditto.
12340 * doc/gcov.texi: Ditto.
12341 * doc/install.texi: Ditto.
12342 * doc/invoke.texi: Ditto.
12344 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12346 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12348 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12350 * config/i386/sse.md (*mov<mode>_internal): Guard
12351 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12353 PR rtl-optimization/59647
12354 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12355 new_rtx into UNSIGNED_FLOAT rtxes.
12357 Copyright (C) 2014 Free Software Foundation, Inc.
12359 Copying and distribution of this file, with or without modification,
12360 are permitted in any medium without royalty provided the copyright
12361 notice and this notice are preserved.